Home

LOGIX-UM004A-EN-P RSLogix 5000 Fuzzy Designer User Manual

1. m mig a Paki Minister ie ia ONS We Ta i PITAT f i a ARPEI EN Position Fuzzy System Monitoring Feature Description Numerical and graphical display Monitoring of all internal variables Archiving Recording specified internal or external variables History graph Plotting history graph for on line or off line monitoring Fuzzy System Monitoring Through Numerical Displays Watch Egg ashController O x SEENTE PaL Input Ports egq_dirt 0 11739 conveyor ep 0 4258 water flowor O4122 oil flow orate 0 3511 Output Ports conveyor ep 0 0326 Water valve 0 01086 oll valve pos 0 01172 Intermediate Components water valve 0 01086 conveyor sp 0 0326 oll valve pos 0 01172 Publication LOGIX UMO04A EN P March 2007 wow inian eoi Get Started with FuzzyDesigner 21 Fuzzy System Monitoring Through Plotting Historical Recordings and On Line Update zH History Graph ball Of x FAAA B pos errar Absolute Graph velocity angle Base variable dm 15 16 21 24 27 30 33 36 39 1 1 Time 3 di FuzzyDesigner Project Formats File Format Description XML FSP complete project file generated by FuzzyDesigner XML user supplied fuzzy system or project file Publication LOGIX UMO04A EN P March 2007 22 Get Started with FuzzyDesigner eanwidiniinanncom Direct Support of Logix5000 controllers FuzzyDesig
2. 000 0 0 ee eee ee 47 Supported Format of Rules 0 0 00000000000 48 Connections 6ee5u eee be eae bee ewe eee Beka ws 52 PARAM lel ea nre ha oe eo oe ee ea ee a py PID Controller 0 00000 00000000 aa 52 COENS 4 oth can Ge eee ke eee eae Bee ees 55 Parameters noaa a aaa a 56 Output Port 0 eee 56 Connections oaaao aa ede ebae he ee eee ews 56 Parameters 0 aa aaa a ened 56 4 www klinkmann com eee Chapter 3 FuzzyDesigner Graphical User Introduction ou 244 ice encter owed ce etsesseavky to 57 Interface Seting OpPUONS essas eee eens ari ekaa mda aE ae ia 57 TOLD esera eee ee oo r 58 FuzzyDesigner Control Basics nanana aaaea 59 MAIO NICU EEEE E EE 60 Chapter 4 FuzzyDesigner Projects Introduction 6 ou aa a 67 Working with Projects 0 0 0 en 67 Creating a Project aw o dx wet keene ens tg mee eka 69 Opening an Existing Project 0 0 0 0 00005 69 Changing the Active Project 060s bew te vneaees 70 Project IOLOMNAUOM so ates hoa nee ene hte nee eS bo 70 paving ATO 2 eeiags Bou eh ae hee eer ees 71 Closing PIOQ ECE 64 6 nn ke Ed ey eee ea DoH DODD EES 71 Designing a Project 2450 b oe ehh eee BRASH ABS T2 PE UOO AP a eer eE e eee ae oe a 72 Designing a Fuzzy System 0 0 00 0 0 008 Ta Fuzzy System Project Window aaa 73 Working with Blocks 00 0 000002 73 Working with Text noaa aaaea ee 7 Fuzzy System Components 0 00000 eee eee 75
3. b Use the Number of Rules numeric control to decrease the number of rules you want to generate c Click the Remove Just Applied Rules check box to remove all already applied rules of the selected Rule Block d Click OK to generate the requested rules or click Cancel or Close the top right dialog corner to return to the appropriate Rule Editor dialog without any changes Shift Selected Rule Up Click this button to shift the selected rule one position up Shift Selected Rule Down Click this button to shift the selected rule one position down Shift Selected Rule to Required Position Click this button to shift the selected rule to the required position A dialog see Target Position Dialog of Rule Editor opens and you can select the required final position for the selected rule in the Position No box Click OK or click Cancel or Close to return to the Rule Editor dialog without making any changes Hide Column Bars Click this button to hide or show the column bars with a graphic representation of the rule DOFs Show Rules As Text Click this button to see the rule base rules as text A dialog see Rules As Text Dialog of Rule Editor with rules in the text format appears and you can perform operations such as copy rules to another application through the clipboard Help Click this button to show information about the Rule Editor Publication LOGIX UMO04A EN P March 2007 112 Fu
4. A x if x E X nin Anal DOF A A X min if x lt A inin A X max if x gt Xmax sna ldlnieiwantkeern FuzzyDesigner Component Library 35 This value is simply membership degree of value x to fuzzy set A and can be interpreted as a degree to which the proposition x IS A is true An example of fuzzification of the crisp input value x is shown in the figure Process of Crisp Input Fuzzification The component input value is 0 3191 The component consists of three linguistic terms negative zero and positive The output of the component is the vector 0 6383 0 3617 0 where 0 6383 is a degree of fulfillment of the term negative 0 3617 is a degree of fulfillment of the term zero and 0 is a degree of fulfillment of the term positive This value is simply membership degree of value x to fuzzy set A and can be interpreted as a degree to which the proposition x IS A is true Process of Crisp Input Fuzzification Term Editor ball pos_error a Ioj x E R NHEMANBD ee eRBls pos error negative postive 0 5 7 Current input value 1 000 0 000 Input Value 0 3 g Parameter Term DOFs negative Zer0 positive 0 3617 0 0000 Term DOF membership degree DOFs of all terms are provided to connect rule blocks to complete the fuzzy logic inference Publication LOGIX UMO04A EN P March 2007 36 FuzzyDesigner Component Library vankilana om Co
5. General Tab Dialog Gutput Port ball angle E IOl x General Description Port General Port Hame Jangle Input Link angle as oe Port Name Specify the OP name This name will be used as the output parameter name when the fuzzy algorithm is compiled to an Add On Instruction Input Link Set up the OP input link The link can realize the connection between the OP and the Input Port Output Linguistic Variable Output Takagi Sugeno Variable or the PID Controller e Description tab dialog Specify the description of a new or existing OP Output Port Properties Dialog Description Tab Dialog Gutput Port ball angle IOl x General Description Port Description angle sen Port Description Enter the description of the OP This description will be used as the output parameter description when the fuzzy algorithm is compiled to an Add On Instruction e OK button Accept the entered properties for the project Publication LOGIX UMO04A EN P March 2007 84 FuzzyDesigner Projects sioneidinkinanncom e Cancel button Click this button to leave the OP properties dialog Any changes made are not applied You can also click Close at the top right corner of the dialog Output Linguistic Variable Use the Edit New Variable New Output Linguistic Variable main menu command or the New Output Linguistic Variable tool bar button to add a new Output Lingu
6. Zoom in or out of the graph to best fit the window e Graph Line Change the thickness and color of the graph line Click the Default Setting option to restore the default settings e Graph Axis Change the bounds grid and values parameters for the graph axis If the bounds are not displayed no other axis parameter is accessible The grid step is automatic and is equal to 1 10 of the interval specified by bounds The value item of the Graph Axis includes both values and names Click the Default Setting option to restore the default settings Publication LOGIX UMO04A EN P March 2007 wow klinkmmannicors FuzzyDesigner Projects 125 e Path Change the visibility and color of the path trace see section 2D Graph You can also delete the complete path e Graph Properties Open the 2D Graph Properties Dialog see section 2D Graph 2D Graph Control Tool Bar The tool bar see 2D Graph Tool Bar has three buttons Zoom In Zoom Out and Zoom To Fit Window All of them have the same function as the menu items with the same name 2D Graph Tool Bar 3D Graph When you click the Tools 3D Graph main menu command then 3D Graph Properties dialog see 3D Graph Properties Dialog appears and you can set up the graph properties e General group box Specify general properties of the required 3D graph 3D Graph Properties Dialog 7 3D Graph Properties ball velocity port From E To fi angle
7. Urt inane Term Editor Cancel Predefined Click this radio button to select a predefined unit of the variable e Variable of Select one of the predefined quantities that has the same meaning as the OTSV e In Select one of the predefined units as the requested unit of the OTSV User Defined Click this radio button to insert a user defined unit of the variable e Unit Specify a unit of the OTSV The unit name can be up to 100 characters long Publication LOGIX UMO04A EN P March 2007 won iniiao ei FuzzyDesigner Projects 91 e Range tab dialog Specify the range of the variable including its default value Output Takagi Sugeno Variable Properties Dialog Range Tab Dialog Am Ea General Unit Range Description Variable Range Minimum 1 hM axinniurn i Default Value fo M Rescale Functions of the Applied Terms Term Editor Cancel Minimum Specify the lower limit of the variable range Maximum Specify the upper limit of the variable range Default Value Specify the default value of the variable The default value must be within the variable range Rescale Functions of the Applied Terms Click this check box to rescale functions of the all applied terms of the OTSV IMPORTANT The minimum value must be always lower then the maximum value When the dialog for a new OTSV is opened the default range Is preset
8. e Activity flag check box if checked the rule is enabled otherwise it is disabled e DOF An active rule is evaluated during the inference process and might influence the fuzzy system output Inactive rules remain in the database but are not evaluated Temporarily enabling and disabling rules can help you fine tune your fuzzy system The rule specification contains the premise variables and weighted consequent variables The rule base columns in the Rule Editor are arranged in the same order as the related variable links are applied If you want to create a rule then use a mouse Select the table cell you want to modify and choose the required parameter By clicking on the cell with the rule activity you can set the rule as active checked or inactive unchecked When you click the cell of the premise or consequent variable choose the required term from the visualized combo box All relevant terms of the respective variables are offered including their inverses The rule base is the only place where you can apply the fuzzy logic NOT operator Inversed terms are marked by NOT followed by the term name In the rule base table the OR operator can also be used 1 Select OR in the combo box then check the terms that you want to use in the OR expression see OR Expression Editor Publication LOGIX UMO04A EN P March 2007 110 FuzzyDesigner Projects Publication LOGIX UMO04A EN P March 2007 www klinkmann com 2 When
9. s LSX or Pe ere kuali shee pal bat arih oo oo he ie spent ptas pr A g iar rarai gumin ie a iei Gt S shot enttba it ee eS Tie waaneripriarec neS CDATA salasar tpr kam awi rst dom teh do Dy Marhel ine ajir ae oe E er Hime ri Plow ate Butalyo ameecr pears ry jactat erata arg lt pat ae G g Sipe aT p RSLinx Classic J ma n 4 Logix If you are unfamiliar with fuzzy logic the next section introduces fuzzy logic terms and principles you might use in your fuzzy system Publication LOGIX UM004A EN P March 2007 12 Get Started with FuzzyDesigner eanwidiniinanncom Fuzzy Logic and Fuzzy Control Essentials This section introduces basic concepts used in a Fuzzy Add On Instruction The designer should know how to deal with an instruction s inputs outputs and fuzzy If Then rules that will be used to define input output mapping There are quite a number of systems or processes that are highly nonlinear not well understood from the formal description point of view or for which a mathematical model is not readily available For these systems or processes there is often an expert that is capable of supervising or controlling the process in a satisfactory manner The figure Nonlinear System Example illustrates the difference between linear and nonlinear systems Nonlinear System Example 100 o Flow Linear System Linear output gives a linear response 0 0 Yo Command 100
10. Allen Bradley TT Le ee ee a Ff f T m 1 a a Oo ae ot oe mm w od rens m a 7 T RSLogix 5000 Fuzzy Designer User Manual Rockwell Automation Li YM www klinkmann com XN pl x www klinkmann com Important User Information Solid state equipment has operational characteristics differing from those of electromechanical equipment Safety Guidelines for the Application Installation and Maintenance of Solid State Controls publication SGI 1 1 available from your local Rockwell Automation sales office or online at http literature rockwellautomation com describes some important differences between solid state equipment and hard wired electromechanical devices Because of this difference and also because of the wide variety of uses for solid state equipment all persons responsible for applying this equipment must satisfy themselves that each intended application of this equipment is acceptable In no event will Rockwell Automation Inc be responsible or liable for indirect or consequential damages resulting from the use or application of this equipment The examples and diagrams in this manual are included solely for illustrative purposes Because of the many variables and requirements associated with any particular installation Rockwell Automation Inc cannot assume responsibility or liability for actual use based on the examples and diagrams No patent liability is assumed
11. FuzzyDesigner Projects 129 Graph Grid Changes the graph grid visibility and color Click the Default Setting option to restore the default settings Graph Texture Changes the graph texture There are four options e Single color texture the graph has the texture defined as color 1 e Gradient texture the graph gets texture from gradient color The lowest parts of the graph are assigned the color defined as color 2 and the highest parts of the graph color 1 e Shadow texture the graph gets a shadowed texture The base color for this is color 1 e No texture the graph is transparent Graph Axis Changes the parameters of the graph axis You can change the visibility and color of all three axes Axis parameters are bounds grid and values Bounds are essential If bounds are not visible then the other two parameters are not displayed The grid distributes the interval defined by bounds into ten identical subintervals The values item includes both values for all axis and their labels Click the Default Setting option to restore the default settings Path Change the visibility and color of the path trace see section 3D Graph You can change its visibility and color You can also delete the complete path Graph Properties Open the 3D Graph Properties Dialog see section 3D Graph Publication LOGIX UMO04A EN P March 2007 130 FuzzyDesigner Projects www klinkmann com 3D Graph Control
12. a yt l a y t n bou t buit 1 t b u t m H d y d H d u d Filter numerator parameters bo b1 6 filter denominator parameters Ai An There are two ways for designing the filter e user defined filter e butterworth low pass filter www klinkmann com FuzzyDesigner Component Library 33 User Defined Filter You set the numerator and denominator coefficients bo b1 b and a a directly the parameters are entered in the specified order separated by the space character Butterworth Low Pass Filter This filter can be created by specifying a normalized cutoff frequency q taken from the interval 0 01 1 and the order of the filter 1 2 3 Bode Plot of the Butterworth Low Pass Filter OdB Hgo C We Q This normalized frequency q corresponds to the absolute frequency O q where T T is the Nyquist frequency for the sampling period T WARNING All dynamical terms in a fuzzy system filters PID controllers have to share the common sampling period Is otherwise the system will not work correctly Connections The output link of the input port is connectable to all components expecting a crisp value at the input This includes the following components e Input Linguistic Variable e Output Port e Output Takagi Sugeno Variable accepts crisp values only e PID accepts crisp values only Publication LOGIX UMO04A EN P March 2007 34 FuzzyD
13. shows all tool bar buttons of the application Create New Project shows or hides the Create New Project tool bar button of the application Open Project shows or hides the Open Project tool bar button of the application Save Active Project shows or hides the Save Active Project tool bar button of the application Undo shows or hides the Undo tool bar button of the application Redo shows or hides the Redo tool bar button of the application Refresh Active Project shows or hides the Refresh Active Project tool bar button of the application Go to Design mode Go to Monitoring mode shows or hides the Go to Design mode Go to Monitoring mode tool bar button of the application Preview shows or hides the Preview tool bar button of the application Print shows or hides the Print tool bar button of the application Hide Tree View Show Tree View shows or hides the Hide Tree View Show Tree View tool bar button of the application New Input Port shows or hides the New Input Port tool bar button of the application New Input Linguistic Variable shows or hides the New Input Linguistic Variable tool bar button of the application New Output Port shows or hides the New Output Port tool bar button of the application New Output Linguistic Variable shows or hides the New Output Linguistic Variable tool bar button of the application
14. 100 oy aFlow Non Linear System Linear output does not give a linear response 0 o 0 10 50 80 100 Command The decision making the expert uses in control system supervision can be expressed as a set of Fuzzy Logic If Then rules Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com Get Started with FuzzyDesigner 13 An expert may be an operator a maintenance person or a control engineer who knows what adjustments are needed during process instability These adjustments may include defining setpoints for process variables defining control action in feedforward or feedback contro or setting gains of conventional controllers and may be as simple as turning a valve or knob Rockwell Automation is introducing a tool for building smart instructions that encode If Then rules and use fuzzy logic internally to describe vague and incomplete knowledge in a natural way Fuzzy Logic may serve in situations where e the process has not been automated and is running in Manual mode e a well tuned PID controller does not provide the desired response however the expert knowledge is available to define the rules for a fuzzy algorithm Let s look at an example where we will discuss building a Heat Ventilation and Air Conditioning HVAC system that manipulates the compressor speed based on room temperature and humidity In HVAC systems room comfort is often associated with vague fuzzy values of tempe
15. March 2007 www klinkmann com To close all open projects use the Project Close All main menu command FuzzyDesigner may display a dialog prompting you to save your project before closing it To save the project s changes click Yes To close the project without saving the project s changes click No To return to the project without saving or closing the project click Cancel Message Box of Closing an Unsaved Project P Do you want to save the changes el s La Designing a Project The fuzzy system is designed in the project window which is the main window for every project opened in FuzzyDesigner There is a graphical representation of the designed hierarchical fuzzy system with applied project components and an additional text comment A detailed description of how to work in the project window is available the section Designing a Fuzzy System Printing a Project You can print the active project using the Print main menu command in the Project menu You can also configure your printer from the Print dialog that appears Having a graphical representation of a Fuzzy System speeds up the design and also makes the internal architecture more transparent and interpretable It also serves as natural project documentation The graphics consist of three parts the workspace of the project window blocks and text These are explained in the following sections 8 2011 www klinkmann com FuzzyDesigner Projects 7
16. voltage current small medium large zero positive 0 0 0 09 0 0 0 0 _ 0 1 min premise part DOF 0 1 Fe 0 1 2 large ati gt small 08 voltage current small medium large zero positive 0 08 0 0 00 0 0 0 0 SG we 0 08 premise a DOF 0 1 3 gt small positive 1 0 voltage current small medium large zero positive premise ae ae 0 aan 0 5 N05 00 0 0 0 0 0 5 Va voltage current total DOF maximum small medium large zero positive 0 5 0 09 0 0 0 0 0 5 Example of a Block Diagram of the Rule Block Evaluation Rule Block voltage medium 0 9 0 09 postive 1 0 fer ALA rae temp erature pressure ono cra pressure current small 1 0 postive 1 0 Publication LOGIX UMO04A EN P March 2007 52 FuzzyDesigner Component Library vankilana om Connections The input logical link of the Rule Block can be connected to a component providing a DOF value as a result of fuzzification or defuzzification such as the e Input Linguistic Variable component e Output Takagi Sugeno Variable component e Output Linguistic Variable component e Intermediate Linguistic Variable component The output logical link of the Rule Block can be connected to components expecting a DOF value as result of fuzzy inference such as the e Output Takagi Sugeno Variable component e Output Linguistic Variable component e Int
17. Cal U 0000 negative positive zero Cal 0 0000 ZeD negative negative 1 Cal 1 0000 Zero zero Zero a Cal U 0000 zero positive positive 7 Cal U 0000 pozitive negative zera i Cal U 0000 pozitive ZeIC positive 1 Cal U 0000 positive positive postive 1 The Rule Editor shows the following columns e Index Shows the index of a rule of the selected Rule Block rule base e Active Shows rule activity e Rule DOF Shows current degree of fulfillment of every rule of the rule base e IF Shows the premise variables of the selected Rule Block and the addressed terms which define the antecedent part of rules stored in the Rule Block e THEN Shows the term of the consequent variables addressed by the rules of the selected Rule Block Rule weights are also shown The column with the weight of the consequent variable marked by the consequent variable name and RW Rule Weight 8 2011 www klinkmann com FuzzyDesigner Projects 109 Operations with Rules You can add create or delete rules Adding a Rule To add a rule click the first available blank row and a new rule with the correct index will be created Repeat this if you want to add more rules Creating a Rule With the Rule Editor you can make or modify the rule base of the selected Rule Block Each rule has the following assigned to it e Index the position of the rule in the rule base usually greater than zero
18. FuzzyDesigner FuzzyDesigner Component Library Table of Contents About This Publication 0 0 00 0000000000000 7 Who Should Use This Publication aoaaa aaa aaa 7 CONVENTIONS 428 6264 6408644444405 BAe ee oe A ee os 7 Chapter 1 Introduction 1 0 ee Q Understanding FuzzyDesigner nnna aaa 9 Fuzzy Logic and Fuzzy Control Essentials 12 Potential Use of Fuzzy Logic 0 0 0 0 000 cee 16 Specifications and Features 0 0 0 0 000 eee 18 Integrated Design Environment IDE screen captures 22 Chapter 2 Introduction 0 00 a 29 Component Interface nnana a 29 Library of Components oaaae ee 30 Supported Membership Functions 00045 30 MOU POEs sesten au 2 ee AS AE ROSS RESP AEA Se BERS OS 32 User Defined Filter nanahan 33 Butterworth Low Pass Filter 0 0 000 0 33 Connections sasaaa a ee 43 Parameters onnaa aaa ee ee ee 34 Input Linguistic Variable 0 0 ee 34 Connections oaaao aaa a a eee 36 Parameters noaua aaa a 36 Output Linguistic Variable ooann aaua 36 Defuzzification 0 ee 37 Connections a oa aaa ee ees 42 Parameters 000 cee ee eee eee 42 Output Takagi Sugeno Variable 0 00004 42 Connections 0 00 0 000 cc eee 45 Parameters noaua aaa a 46 Intermediate Linguistic Variable onnan aana aaa 46 Connections ooa aaa ce eee 47 Parameters 6 hk oo MEK G ERA a 47 Rule Block
19. FuzzyDesigner Graphical User Interface 59 FuzzyDesigner Tree View Tree View FE Projects EA Bi Input Forts o ef pos_error o C9 velocity E Input Linguistic Variables L pos_error The controls in FuzzyDesigner are very similar to other Microsoft Windows applications and are easy to use FuzzyDesigner s functions can be accessed from the main menu or by clicking the tool bar button commands By right clicking the tree view item or in the FuzzyDesigner project window a related pop up menu appears You can also control most of the functions using these pop up menus These menus are context sensitive they offer the most useful commands applicable to the current window Use the Project Recent Projects menu command to reopen the most recent projects in FuzzyDesigner Up to four projects are shown in the Recent Projects menu Detailed Help is accessible through the FuzzyDesigner main menu The Help button can be used in the Term and the Rule Editor tool bars as well Using this button information about the particular editor can be displayed You can apply the Edit Undo and Edit Redo menu commands to a few changes made in the active project The Undo menu command is always applied to the last change in the active project and the Redo menu command is always applied to the next change in the history of changes made in the active project Publication LOGIX UM004A EN P March 2007 60 FuzzyDesigner Graphical U
20. PULPO a 644 44 bs a Ea E 76 Input Linguistic Variable naaa aaa aaa 79 MTO ee ee egean eA 82 Output Linguistic Variable 4 siivsng eee bebeea des 84 Output Takagi Sugeno Variable 045 88 Intermediate Linguistic Variable 0 92 RUC DIO k 4 ae 446 44 oe bynes 4 oe ogee ee weg Bee es 94 PID CODO 20645554 ae eae os eee ee eee 97 Tent POUOr cae n4 wees oe Bee eae eee ne leer ee ae 102 Term Properties Dialog c25446 ds ae ashe ae ee Kee G ee oe 105 R le EOQUO o2e46 e4apeusee ew anced eee owas Hawes 108 Operations with Rules 2 0 2 0 0000000 eee 109 Rule Editor Tool Bar 2 0 00 00 0 00 eee 110 Porn Order Eio sea 08 wed ene ee eae Bee ae ee eS 113 WAU ato enone eee eee de ee beeen ees eee ee eee 113 ENSO GADE s046 546 e 85546450 aE ees ee ee 117 History Graph Control Context Menu 119 History Graph Control Tool Bar 121 History Graph Control Mouse Dragging 122 ADO eener ed heheh ee SaG GGG G4 sees oe 122 2D Graph Control Context Menu 124 2D Graph Control Tool Bar 0 00005 125 DOP a454504 8 564405465446 E Gs A 125 3D Graph Control Context Menu 128 www klinkmann com Fuzzy System Simulation RSLogix 5000 Add On Instruction XML Format of a Fuzzy Project Glossary 3D Graph Control Tool Bar 004 3D Graph Control Mouse Dragging Chap
21. Publication LOGIX UMO04A EN P March 2007 waia kiinkinaniicoi RSLogix 5000 Add On Instruction 137 5 Create an instance tag of the fuzzy Add On Instruction The type of the tag has to match the imported Add On Instruction Controller Tags ball controller Scope fin bal Show Showa Neme e as Bas Data Typd Style Description a ange REAL Float control variable S0 for angle actuator output Ea O posero REAL Float ball position error SF actual position input pee M velocity of the ball input meae oe pe t Nonlinear Fuzzy PO Controller of Ball amp Bearm System TI a A a g H 6 Create the tags if not yet created in which you want to store inputs and outputs of the fuzzy AOI The type of these tags has to be REAL 7 Add the fuzzy Add On Instruction to code cfd T 4 Favortes A Add On A Bit A Tim B MainProgram MainRoutine l Saele a E rect Nonlinear Fuzz J Ball pos error velocity angle O D O OO DO OD OM DO Oo A wo Publication LOGIX UMO04A EN P March 2007 138 RSLogix 5000 Add On Instruction saaiildinkmann com 8 Enter parameters operands for the AOI C MainProgram MainRoutine E Oj x Controller Tags ball controller a E ioj x lst iH IME EA Scope fa ball Show Show All Name amp Alias Bas Data Type Style Description fuzzy controller a g angle REAL Float
22. Shift Port Up button Click this button to shift any port up by one position Select a port in the Input Ports table or the Output Ports table e Shift Port Down button Click this button to shift any port down by one position Select a port in the Input Ports table or the Output Ports table e OK button Accept the entered properties for the project e Cancel button Click this button to leave the Port Order Editor dialog Any changes made are not applied You can also click Close at the top right corner of the dialog Watch Use the Tools Watch main menu command to open a dialog with all available component values of the active project see Watch Dialog The following buttons are available e Data History Click this button to set up a history depth see Watch Data History Dialog The history depth is specified by number of samples which are archived for user needs e History Data Click this button to select components or choose one or more variables for archiving see Watch Data History Dialog Publication LOGIX UMO04A EN P March 2007 114 FuzzyDesigner Projects sioneidinkinanncom e History Graph Click this button to show the history graph When a dialog see Watch History Graph Properties Dialog with a check box lists appears choose values for the history graph To view the history graph see History Graph click the Show button at the bottom of the dialog When the history is empty the g
23. Specify additional information for the project in the appropriate text boxes e Project Name e Description e Author e Company e Project ID free form project identification code wow iniiao eoi FuzzyDesigner Projects 71 OK button Accept the entered properties for the project Cancel button Click the button if you do not want to apply the changes but you want to close the dialog Project Information Dialog Project Information oO x Properties Project Name ball Description Author Company Project ID OF Cancel Saving a Project To save changes made in the active project use the Project Save main menu command To save newly created projects or to save to another project use the Project Save As main menu command The standard Save As dialog appears immediately so you may click the directory where the project will be stored and type a new project name in the dialog The FuzzyDesigner project can be saved to the file with fsp extension where all the information about the project is stored in the XML format This file can be opened in any XML editor Closing a Project You can close an active project using the Project Close main menu command or directly close the active project by clicking Close in the top right corner of the project window Publication LOGIX UMO04A EN P March 2007 72 FuzzyDesigner Projects Designing a Fuzzy System Publication LOGIX UMO04A EN P
24. Term Names in Graph Hide or show names of the applied variable terms seen in the membership function editor except for Intermediate Linguistic Variable and Output Takagi Sugeno Variable Variable Properties Show the appropriate variable properties dialog The Term Editor dialog remains open Publication LOGIX UMO04A EN P March 2007 104 FuzzyDesigner Projects sioneidinkinanncom Help Show information about the Term Editor When you want to specify input values of the Input Linguistic Variable you can use the membership function editor slider Position the mouse cursor above the slider click the right mouse button and move the slider to the desired position then release the right mouse button The actual slider position of the Input Linguistic Variable is seen in the Input Value text box below the membership function editor The input value is immediately stored in the appropriate project which is then Statically evaluated e Close button Click this button to close the Term Editor dialog Input Term Editor for Input Linguistic Variable Term Editor ball pos_error o O x EEEIEE E R pos error re negative Fera positive 1 5 0 0 1 000 0 500 0 000 0 500 1 000 Input Value 0 0000 Parameter Term DOFs negative zero positive 1 0000 0 0000 _ Publication LOGIX UMO04A EN P March 2007 www klinkmann com Term Properties Dialog FuzzyDesigner P
25. YANKS 0 2 pos error The 3D graph can be controlled either from the context menu or tool bar buttons You can rotate the graph by using the mouse Publication LOGIX UMO04A EN P March 2007 128 FuzzyDesigner Projects sioneidinkinanncom 3D Graph Control Context Menu Right click to display the context menu The menu structure see 3D Graph Context Menu is explained below 3D Graph Context Menu t Visible Color Default Setting Single Color e Gradient oor Ir Chadw oom ut n None oom To Fik window Color 1 Rotate Het Color 2 Rotate Aight Rotate Up Default Setting Rotate Down Rotate Cw w Bounds Visible Rotate CLM Bounds Color Graph Grid arid Visible Graph Texture b Brid Color Graph Axi Doe v Values Visible Path Values Color Graph Properties Default Setting Active Point Visible Active Point Color w Path Visible Path Color Clear Path Default Setting e Zoom In Zoom in on the graph e Zoom Out Zoom out of the graph e Zoom To Fit Window Zoom in or out of the graph to best fit the window e Rotate Left Rotate the 3D graph to the left e Rotate Right Rotate the 3D graph to the right e Rotate Up Rotate the 3D graph up e Rotate Down Rotate the 3D graph down e Rotate CW Rotate the 3D graph clockwise e Rotate CCW Rotate the 3D graph counter clockwise Publication LOGIX UMO04A EN P March 2007
26. all rules are computed from DOFs of the rule premise by using the selected t norm 2 DOFs of all consegent variables terms are computed for every rule These DOFs are obtained from DOFs computed in step 1 multiplied by weights of consequent variables 3 Total DOFs of all consequent variables are computed for the overall fuzzy system Total DOF of one consequent variable is computed as maximum value of DOFs computed in step 2 for the appropriate consequent variable Assume a simple fuzzy system with these two premise variables temperature pressure with terms e Temperature small large e Pressure negative zero positive This system also has two consequent variables voltage current with terms e Voltage small medium large e Current zero positive This system also has a minimum t norm The rule base can be formulated as the following Temperature is small Oressure is negative Voltage is medium 0 9 Current is positive 1 0 Temperature is large Voltage is small 0 8 Pressure is positive Voltage is small 1 0 Current is positive 1 0 Publication LOGIX UMO04A EN P March 2007 50 FuzzyDesigner Component Library sauwiildinkimann com You can also formulate the rule base schematically e small negative gt medium 0 9 positive 1 0 e large negative gt small 0 8 o positive gt small 1 0 positive 1 0 In the following example some premis
27. by Rockwell Automation Inc with respect to use of information circuits equipment or software described in this manual Reproduction of the contents of this manual in whole or in part without written permission of Rockwell Automation Inc is prohibited Throughout this manual when necessary we use notes to make you aware of safety considerations Identifies information about practices or circumstances that can cause an explosion in a hazardous environment which may lead to personal injury or death property damage or economic loss IMPORTANT Identifies information that is critical for successful application and understanding of the product Identifies information about practices or circumstances that can lead to personal injury or death property damage or economic loss Attentions help you identify a hazard avoid a hazard and recognize the consequence ATTENTION TNH ava Labels may be on or inside the equipment for example a drive or motor to alert people that dangerous voltage may be present TTT aha Labels may be on or inside the equipment for example a drive or motor to alert people that surfaces may reach dangerous temperatures ne A rora CN A A A Allen Bradley ControlLogix RSLogix 5000 Logix and RSLinx are trademarks of Rockwell Automation Inc Trademarks not belonging to Rockwell Automation are property of their respective companies www klinkmann com Preface Get Started with
28. cutoff frequency here Filter with Specific Transfer Function group box There are two text boxes for you to specify the filter parameters Numerator Coefficients If you selected the Filter with Specific Transfer Function radio button enter the numerator coefficients here see Input Port component description Separate coefficients with a space by b Db Denominator Coefficients If you selected the Filter with Specific Transfer Function radio button enter the denominator coefficients here see Input Port component description Separate coefficients with a space 4 ap m Publication LOGIX UM004A EN P March 2007 78 FuzzyDesigner Projects vaikina com e Description tab dialog Specify the description of a new or existing IP Input Port Properties Dialog Description Tab Dialog Input Port ball pos_ error jol x General Description Port Description Reset Filter State Cancel Port Description Enter the description of the IP This description will be used as the input parameter description when the fuzzy algorithm is compiled to an Add On Instruction e Reset Filter State button Click this button to reset the internal state of the implemented filter e OK button Accept the entered properties for the project e Cancel button Click this button to close the IP properties dialog Any changes made are not applied You can also click Close at the top ri
29. e i Term Editor EggWashControleses x O y TN OON a A a fa Input Ports PAA egg_dirt conveyor_speed_dif 3 A A na ative zero positive on Nan conveyor_speed_dif C water_flow_rate OON conveyor_speed_dif 0 0326 05 9 oil_flow_rate Mani i a0 0 0326 E A Input Linguistic Variat conveyor_speed conveyor_speed el al ry A 0 0 o f egg_dirt 0 426 nA oe 0 3000 01500 00000 01500 0 3000 RuleBlock1 ofl Dan at dif 7 conveyor_speed at i dif water_vave_pos_ 1 water_flow_rate PON water_va ve pos 0 0109 p aie Ce 7 4 0 0109 Output Value 0 03261 Parameter nsn oil_flow_rate z4 E E Output Ports water_flow_rae water_flow_rae Term DOFs 0 412 I Non ol Nan oil_valve_pos_dif m ED conveyor_speed_ negative zero positive lt default gt 0 6739 0 3261 0 0000 Bee D watervalve_pos i Non oil_valye_pos_dif 0 0117 gD oll_valve_pos_dif 0 0117 E F Output Linguistic Vari RuleBlock2 i watervalve_pos i errs EggWashCo tro le 10 49 J Rule Editor EggWashController RuleBlock1 C conveyor_speed_ 7 oil_valve_pos_dif NARNARA ae E z 3 3 Tf B Ee Rule Blocks gg_dirt IF THEN BB RuleBlock1 low acceptable bextra_high ff RuleBlock2 1 0 Index Active RuleDOF egg ditt conveyorspeed water_flow_rate conveyor_speed_dif 05 1 0 0000 high low zero 0 0 2 mz 00000 i high high high nega
30. help search dialog e Product Activation shows the Product Activation dialog with the Computer ID Send this together with the serial number to technical support to get the Activation Key to access the desired application features e About about the FuzzyDesigner Publication LOGIX UMO04A EN P March 2007 66 FuzzyDesigner Graphical User Interface T E con Tool Bar Menu This section lists all menu buttons in the tool bar menu see FuzzyDesigner Tool Bar of the FuzzyDesigner main window e Create new project e Open project e Save active project e Undo see section FuzzyDesigner Control Basics e Redo see section FuzzyDesigner Control Basics e Refresh evaluates static the just active project e Go to Design Mode Go to Monitoring Mode switches FuzzyDesigner between the Design and Monitoring Mode see section FuzzyDesigner Control Basics e Preview shows the preview of the current project e Print e Hide Tree View see section FuzzyDesigner Tree View e New Input Port e New Input Linguistic Variable e New Output Port e New Output Linguistic Variable e New Output Takagi Sugeno Variable e New Intermediate Linguistic Variable e New Rule Block e New PID Controller Help shows the FuzzyDesigner help Publication LOGIX UMO04A EN P March 2007 www klinkmann com Introduction Working with Projects Chapter 4 FuzzyDesigner Projects The basic working unit in th
31. of the crisp input value of the linguistic terms defined with each input linguistic variable fuzzy number A convex fuzzy set with maximum membership degree equaling 1 Hierarchical Fuzzy System HFS A fuzzy system with chained rules and hierarchical structure of the rule base Inference An application of linguistic rules on input values in order to generate output values Input Linguistic Variable ALV Stores linguistic terms and is used for classification of the actual component input represented by a crisp value into the fuzzy sets defined for the respective linguistic terms Input Port CIP Preprocesses and stores values of a fuzzy system s input variables Intermediate Linguistic Variable MLV Bridges logical chaining of rule blocks Publication LOGIX UMO04A EN P March 2007 Glossary 151 largest of maximum LOM A discontinuous defuzzification technique linguistic rule IF THEN rule with condition and conclusion one or both linguistic linguistic term In the context of fuzzy control linguistic terms are defined by fuzzy sets linguistic variable A variable that takes values in the range of linguistic terms link A type of connection between fuzzy system components specifying format of data transferred Maximum Center Average MCA A continuous defuzzification technique Mean of Maximum MOM A discontinuous defuzzification technique Membership Function MF Function which defines the degree of me
32. part Fror To fi Grid Density Density 2 ce coe Publication LOGIX UMO04A EN P March 2007 126 FuzzyDesigner Projects sioneidinkinanncom X Axis Link the selected Input Port or a disconnected Input Linguistic Variable to the x axis of the 3D mesh plot From Sets the lower limit of the x axis range To Sets the upper limit of the x axis range Y Axis Links the selected Input Port or a disconnected Input Linguistic Variable to the y axis of the 3D graph From Sets the lower limit of the y axis range To Sets the upper limit of the y axis range Z Axis Links the selected Output Port Output Linguistic Variable or the Output Takagi Sugeno Variable to the z axis of the 3D graph From Displays the lower limit of the z axis range To Displays the upper limit of the z axis range e Grid Density group box Specify the grid density of the 3D graph Density Displays the selected grid density of the graph Select the density by clicking a point on the slider e Create button Create the required 3D graph see 3D Graph e Cancel button Close the 3D Graph Properties dialog No graph will be created You can also click Close at the top right corner of the dialog Publication LOGIX UMO04A EN P March 2007 eal iterken FuzzyDesigner Projects 127 3D Graph 12 3D Graph No 0 ball a 3 loj x ab A
33. properties dialog appears and a default name is assigned to the component Click OK to add the component to the appropriate project All Intermediate Linguistic Variable names must be unique in the same project e General tab dialog Specify the name of the IMLV Intermediate Linguistic Variable Properties Dialog General Tab Dialog Intermediate Linguistic Yariable ball Intermediate ariabl a x General Tems Description Yariable General Varable Hame Intermediate ariabled ee Variable Name Enter the name of the variable e Terms tab dialog The IMLV properties dialog defines the variable through the following terms Intermediate Linguistic Variable Properties Dialog Terms Tab Dialog Intermediate Linguistic Yariable ball Intermediate e General Terms Description Mames flow medium high Term Editor Cancel Count Specify the number of terms defined for the intermediate linguistic variable When the IMLV properties dialog is open for an existing variable the current number of terms is shown Publication LOGIX UMO04A EN P March 2007 94 FuzzyDesigner Projects sioneidinkinanncom Names Change default names and specify names of terms for the newly created intermediate variable When the IMLV properties dialog is open for the existing variable the names of terms already applied are shown IMPORTANT When the IMLV p
34. right corner of the dialog Open this dialog by double clicking the right mouse button on the selected term in the Term Editor membership function editor or from the table of term DOFs or the Term Properties tool bar button of each Term Editor For the Intermediate Linguistic Variable and the Output Takagi Sugeno Variable the terms can be selected only from the table of term DOFs Input Term Properties Dialog SL Input Term Properties balli pos ermar IOl x General Parameters O00 E O00 oad Publication LOGIX UMO04A EN P March 2007 FuzzyDesigner Projects 107 Output Term Properties Dialog General Term Name a Output Term Properties Dialog for the Output Takagi Sugeno Variable Output Term Properties pendulum CYK 0 3 General Term Name ko Term Type LINEAR Parameters lo 4 4037 4 5046 4 4037 15 4325 3 7569 0 0 e Publication LOGIX UM004A EN P March 2007 108 FuzzyDesigner Projects Rule Editor Publication LOGIX UMO04A EN P March 2007 www klinkmann com Click Rule Editor in the Rule Block properties dialog to open the Rule Editor dialog see Rule Block Rule Editor with a Rule Base You can also open the Rule Editor by double clicking a Rule Block Rule Block Rule Editor with a Rule Base Rule Editor ball Rules l Ioj x EE ri Cal U 0000 ease negative negative 1 i Cal U 0000 negalve zero negative 1
35. shape membership functions can be used in input and output linguistic variable components Inverse Trapezoidal Membership Function with Parameters vertices a b c d l if x lt a x b a b if xe a b A x 0 if xe b c x c d c if xelc d 1 if x2d If a b then Ala 1 If c d then A c 1 Inverse trapezoidal membership functions can be used in an input linguistic variable component Publication LOGIX UMO04A EN P March 2007 32 FuzzyDesigner Component Library Input Port Publication LOGIX UMO04A EN P March 2007 www klinkmann com Inverse S shaped Membership Function cubic spline with Parameters a b c d if sP if xe a b if xe b c 2 s if xel c d if x2d If a b then Ala 1 If c d then A c 1 Inverse S shaped membership functions can be used in an input linguistic variable component Singleton Membership Function with Parameter position center c l if x c 0 otherwise A x l Singleton membership functions can be used in an output linguistic variable component The fuzzy system Input Port component stores an actual input value entering the HFS Optionally you can preprocess the input values by using the linear digital filter This filter is defined by its pulse transfer operator H expressed in terms of the backward shift operator d or equivalently in time domain as a difference equation as follows b d by bd b d a d l ad a d y t
36. that the project deals with crisp inputs and outputs which is the case in most control applications The inference realized by the designed fuzzy system generates static input output mapping 2D graphs and 3D mesh plots are graphical representations of the mapping IMPORTANT Dynamic components such as input ports with filters and PID controllers resulting in a dynamic fuzzy system give 2D and 3D plots different meaning and use When the 2D 3D Graph is opened and the related project is manually simulated the system behavior is traced in yellow To trace the system behavior for another simulated input the 2D 3D graph must be regenerated When you click the Tools 2D Graph main menu command the 2D Graph Properties dialog see 2D Graph Properties Dialog appears and you can specify the 2D graph properties e General group box Specify the general properties of the 2D graph 2D Graph Properties Dialog 7 2D Graph Properties ball cel wow inani eon FuzzyDesigner Projects 123 X Axis Select the Input Port or a disconnected Input Linguistic Variable as the x axis of the 2D graph From Specify the lower limit of the x axis range To Specify the upper limit of the x axis range Y Axis Select the Output Port the Output Linguistic Variable or the Output Takagi Sugeno Variable as the y axis of the 2D graph From Displays the lower limit of the y axis range To D
37. to 1 for minimum and 1 for maximum The default value is set up to the middle of this variable range Publication LOGIX UMO04A EN P March 2007 92 FuzzyDesigner Projects sionedinkimanncom e Description tab dialog Specify the description of a new or existing OTSV Output Takagi Sugeno Variable Properties Dialog Description Tab Dialog Output Takagi Sugeno ariable ball akagiSugeno ariab olx General Unit Range Description Variable Description Term Editor Cancel Variable Description Enter the description of the OTSV e Term Editor button Click this button to open the Term Editor see section Term Editor where you can chage the terms of the variable e OK button Accept the entered properties for the project e Cancel button Click this button to close the OTSV properties dialog Any changes made are not applied You can also click Close at the top right corner of the dialog Intermediate Linguistic Variable Use the Edit New Variable New Intermediate Linguistic Variable main menu command or the New Intermediate Linguistic Variable tool bar button to add a new Intermediate Linguistic Variable MLV to the active project Alternatively use the New Intermediate Linguistic Variable project window context menu command to add a variable Publication LOGIX UMO04A EN P March 2007 wow diniinan eer FuzzyDesigner Projects 93 The Intermediate Linguistic Variable
38. tracking maximum e Description tab dialog Specify the description of a new or existing PID controller PID Controller Properties Dialog Description Tab Dialog Controller Description a Controller Description Enter the description of the PID controller e OK button Accept the entered properties for the project Publication LOGIX UMO04A EN P March 2007 102 FuzzyDesigner Projects sioneidinkinanncom e Cancel button Click this button to close the PIDC properties dialog Any changes made are not applied You can also click Close at the top right corner of the dialog Term Editor Default terms are defined for a variable when you add it to the project The exception is the Output Takagi Sugeno Variable To change the default or to modify the existing terms setting you need to open the Term Editor dialog see Input Term Editor for Input Linguistic Variable or Output Term Editor for The Output Linguistic Variable Open the editor by clicking Term Editor or double clicking a variable in the project window or tree view The Term Editor consists of two main parts the membership function editor and the table of term DOFs DOF degree of fulfillment The membership function editor does not exist for the Output Takagi Sugeno Variable and for the Intermediate Linguistic Variable e Term Editor tool bar Use the buttons in the tool bar to create a new term delete an existing term or change variable t
39. you want to apply a NOT operator to the whole OR expression check Apply NOT for the expression 3 Click OK OR Expression Editor 2 OR Expression pendulum Rules O x Apply NOT for the expression Select Terms Connected by OR C neg zero cence _ When you set up a term in the consequent variable cell you can select the weight of the consequent variable All visible changes are stored in the appropriate Rule Block of the project immediately Deleting a Rule When you click the first column the appropriate rule will be marked as selected and the rule background will turn blue You can also selection multiple rules To delete the selection press the Delete key on the keyboard Rule Editor Tool Bar From this tool bar you can access editor commands by right clicking the tool bar buttons The following buttons are available e Generate Possible Rules Enables automatic generation of all rules resulting from the combination of terms and variables entering the rule premise Negated terms terms modified by NOT operator in the rule table are not considered a Click the tool bar button and a dialog see Generate Rules Dialog of Rule Editor appears where you can check the premise variables in the checked list box The maximum number of generated rules of the selected premise variables is shown in the Maximum Number of Rules textbox 8 2011 www klinkmann com FuzzyDesigner Projects 111
40. 2 327 3752 tel 7 495 641 1616 automation klinkmann fi klinkmann klinkmann spb ru moscow klinkmann spb ru www klinkmann com Yekaterinburg Samara Kiev tel 7 343 376 5393 tel 7 846 273 95 85 tel 38 044 495 33 40 yekaterinburg klinkmann spb ru samara klinkmann spb ru klinkmann klinkmann kiev ua Riga Vilnius Tallinn Minsk tel 371 6738 1617 tel 370 5 215 1646 tel 372 668 4500 tel 375 17 200 0876 klinkmann klinkmann v post klinkmann It klinkmann est klinkmann ee minsk klinkmann com
41. 3 Fuzzy System Project Window The fuzzy system project window enables the designer to create block diagrams of a fuzzy system by inserting and linking graphical objects library components and text The window size is user defined and accommodates any structure of a fuzzy system Avoid moving any object outside of this area If you right click the empty Design Sheet window the Design Sheet context menu see Fuzzy System Project Window Context Menu appears The following menu items are available Fuzzy System Project Window Context Menu Select All New Input Fort New Output Fort New Input Linguistic Yanable New Output Linguistic Variable New Output T akagi Sugeno Warable New Intermediate Linguistic Variable New Rule Block Hew PID Hew Text Project Information e Select all Selects all blocks and texts e New Input Port New Output Port Adds a new block to the sheet e Project Information Opens a window with project information Working with Blocks There are eight types of Fuzzy System Components see section Fuzzy System Components referred to as blocks which enable you to design the fuzzy system as a block diagram Blocks are graphical objects Adding a Block A new block can be added to the existing diagram in several ways The first is by using context menu This menu is described in the section Fuzzy System Project Window The other way to add a new block to the existing diagram is
42. AOI instance tag does not have the same parameters as the fuzzy project in FuzzyDesigner and only the project structures are the same the following dialog is shown This can happen if you have changed some parameters off line for example the membership function position Question 2 Only the project structures are the same Do you want to continue eT If communication is successfully established the following window appears On line Connection Panel Eqg ashController O x Ae b gt wh Connection Server Sling OPC Sever Mode DTT Topic EqgwashConolles Add On EggWashContoler i sS S OPC Path RSLins OPC Server EggwashContoller Online EggwashContoller Monitoring Periods Requested 3000 ms Real inane me 4 Start monitoring by clicking on the Run Monitoring gt icon The On line Connection Panel offers the following functionality e If you want to change the monitoring period click the Change Monitoring Period icon On line Monitoring Period EggWashController x On line Monitoring Period Monitoring Period OOD Jin millisecond m Publication LOGIX UM004A EN P March 2007 wwa hiinkinanicoi RSLogix 5000 Add On Instruction 143 e If you make some changes for example changing membership function position in the fuzzy project in FuzzyDesigner and you want to apply the changes in Logix click the Apply Changes to Logix icon on the On line Con
43. However vague classifications are more realistic as there is usually no sharp border between Low Medium or High temperatures In this situation however a single numerical value might fall into multiple categories For example it might be partially Medium and partially High as shown in the following figure A specification of how much the particular value of temperature fits into the meaning of the label of the category fuzzy set is described by the membership function which becomes a design parameter of the fuzzy controller www klinkmann com Get Started with FuzzyDesigner 15 Similar fuzzy terms are designed for the output variables that is Low Medium and High for compressor speed in our example degree of membership evel of da ssification Lowy Medium High 1 mene Class fication Result 0 75 Medium 0 High 2 Low oa SS Ba Crisp Value Fuzzy rules The way in which the classified inputs are treated when passing through rules is shown in the following figure for our compressor control example Temperature i5 Aigh hedium Tem perature Compressor B runninghledium or Compressor iS runningAigh Humidity amp Aigh very High Humidity Medkri High Wery_Hgh een Compressor Speer Center of rawit EERE Publication LOGIX UMO04A EN P March 2007 16 Get Started with FuzzyDesigner Publication LOGIX UMO04A EN P March 2007 www klinkmann com First the numerical values of Tempe
44. IX UMO04A EN P March 2007 98 FuzzyDesigner Projects sauwiildinkimann com e General tab control page Specify the main properties of a new or existing PIDC PID Controller Properties Dialog General Tab Dialog a Options Description FID Controller Hame PIDControllerd Input Links Process Yarable none Set Point Link none iat Set Point Value oO P Gain Link none 7 P Gain Value oO Gain Link none bi Gain Value OD l D Gain Link none D Gain Value 0 Bias Link none bi Bias Walue OD Manual Control none Mode Switch none ie PID Controller Name Specify the controller name Input Links group box Set up all available input links or values e Process Variable Set up required process variable link e Set Point Link Click the check box and select the set point link e Set Point Value When the Set Point Link check box is not checked enter the set point value in the text box e P Gain Link Click the check box and select the P gain link e P Gain Value When the P Gain Link check box is not checked enter the P gain constant value e I Gain Link Click the check box and select the I gain link Publication LOGIX UMO04A EN P March 2007 wow inani eon FuzzyDesigner Projects 99 I Gain Value When the I Gain Link check box is not checked enter the I gain constant value e D Gain Link Click the c
45. Main Menu Project Structure New Ctrl M Open Chrl 0 Close Close All Save Ctrl 5 Save AS Project Information Preview Princ Ctrl P Recent Projects d Exit e New creates a new project e Open opens an existing project e Close closes the active project e Close All closes all open projects e Save saves the active project e Save As save the active project with a different name e Project Information displays the properties for the current project e Preview shows the preview of the currently active project e Print prints the active project e Recent Projects shows the four most recent projects e Exit closes the FuzzyDesigner application Edit FuzzyDesigner Main Menu Edit Structure Undo Redo Refresh Ctrl 2 Ctrl Mew Input Port Mew Qurput Pork Go to Monitoring Mode i Mew Port Mew Input Linguistic Variable Mew Variable t Mew Output Linguistic Yariable Mew Rule Block Mew Output Takagi Sugeno Variable New PID Controller Mew Intermediate Linguistic Variable e Undo see section FuzzyDesigner Control Basics e Redo see section FuzzyDesigner Control Basics e Refresh updates the display of the active project Publication LOGIX UMO04A EN P March 2007 62 FuzzyDesigner Graphical User Interface T E con e Go to Design mode Go to Monitoring mode switches the project
46. N P March 2007 Weini kilnikinaniicoi RSLogix 5000 Add On Instruction 139 e An RSLinx Topic corresponding to the Logix controller is created see section Configuring RSLinx OPC Server Topic If an RSLogix 5000 project is saved then a default RSLinx Topic is created with a default name that corresponds to the Controller Name e The fuzzy project used to generate the fuzzy AOI is open in FuzzyDesigner Follow these steps to establish online communication with a Logix controller 1 In FuzzyDesigner click Tools gt Add On Instruction gt On line Connection Wizard 2 Specify the RSLinx OPC server If RSLinx software is installed on the same computer select the RSLinx OPC Server option and click Next On line Connection Wizard Egg ashController ol x OPC Server Specification ASLins OPC Server C ASLins Remote OPC Server Node computer name IP address domaini Eee ee ee ______ Publication LOGIX UMO04A EN P March 2007 140 RSLogix 5000 Add On Instruction navikn com If RSLinx software is installed on a remote computer select the RSLinx Remote OPC Server option and specify a Node then click Next Ff On line Connection Wizard EggWashController l If communication is not established contact your network administrator to configure DCOM settings Publication LOGIX UM004A EN P March 2007 Wadai kiinikinaniicoi RSLogix 5000 Add On Instruction 141 3 Select the Add On I
47. N P March 2007 www klinkmann com e Description tab dialog Specify the description of a new or existing ILV a okl Input Linguistic Variable Properties Dialog Description Ta Input Linguistic Yariable ball jpos error General Urut Range Terns Description Variable Description position error Term Editor Cancel Variable Description Enter the description of the ILV e Term Editor button Click this button to open the Term Editor see section Term Editor where the predetermined variable terms can be changed count names e OK button Accept the entered properties for the project e Cancel button Click this button to close the ILV properties dialog Any changes made are not applied You can also click Close at the top right corner of the dialog Output Port Use the Edit New Port New Output Port main menu command or the New Output Port tool bar button to add a new Output Port OP to the appropriate fuzzy system project Alternatively use the New Output Port project window context menu command to add a port The Output Port properties dialog appears and a default name is assigned to the component Click OK to add the component to the appropriate project All Output Port names must be unique in the same project wow klinkmnannicors FuzzyDesigner Projects 83 e General tab dialog Specify all parameters of a new or existing OP Output Port Properties Dialog
48. New Output Takagi Sugeno Variable shows or hides the New Output Takagi Sugeno Variable tool bar button of the application New Intermediate Linguistic Variable shows or hides the New Intermediate Linguistic Variable tool bar button of the application New Rule Block shows or hides the New Rule Block tool bar button of the application New PID Controller shows or hides the New PID Controller tool bar button of the application Help shows or hides the Help tool bar button of the application Publication LOGIX UMO04A EN P March 2007 64 FuzzyDesigner Graphical User Interface PATR e Status Bar shows or hides the status bar of the application e Tree View shows or hides the Tree View tab control page of the application Tools FuzzyDesigner Main Menu Tools Structure Options Show in Status Area Reset Internal Stakes Process Membership Functions Set Port Order Watch Simulation 2D Graph 3D Graph Instruction Generator Radda Tachnichian On line Connection Wizard Import e Options Show in Status Area check this menu item to set the FuzzyDesigner to the server mode The FuzzyDesigner icon shows in the status area and the application should be closed only through the icon context menu e Reset Internal States resets the internal states of all Input Port filters and PID Controllers in the just active project e Process M
49. P March 2007 avea iinkinaniicoi FuzzyDesigner Component Library 45 Different linear state feedback controllers are to be smoothly activated for different process states and setpoints scheduling controller gains There are three rules e IF x1 IS A1 AND x2 IS B1 THEN y KO K1 x1 K2 x2 y1 e IF x1 IS A2 AND x2 IS B2 THEN y GO G1 x1 G2 x2 y2 e IF x1 IS A3 AND x2 IS B3 THEN y CO y3 setpoits Defined functional terms K type LINEAR parameters KO K1 K2 G type LINEAR parameters G0 G1 G2 C type CONSTANT parameters CO Evaluation weighted average of functions yt DOF K y2 DOF G y3 DOF C X from a rule block a la y Default Value if all DOFs 0 Connections The input logical link of the output Takagi Sugeno variable can be connected to a component providing a DOF value as result of fuzzy inference that is the Rule Block component Publication LOGIX UM004A EN P March 2007 46 FuzzyDesigner Component Library sauwiildinkimann com The input value link of the output Takagi Sugeno variable can be connected to components providing a crisp value such as e Input Port component e Output Linguistic Variable component e PID component e Output Takagi Sugeno Variable component The output value link of the output Takagi Sugeno variable can be connected to components expecting a crisp value such as e Output Port component e Input Li
50. Terms Click this check box to rescale the membership functions of all terms of the ILV IMPORTANT The minimum value must be always lower then the maximum value When the dialog for a new ILV is opened the default range is preset to 1 for minimum and 1 for maximum e Terms tab dialog The ILV properties dialog defines the variable through the following terms Input Linguistic Variable Properties Dialog Terms Tab Dialog Input Linguistic ariable ball pos _error x General Unit Range Terme Description Variable Terms Count E Type TRAPEZOID Mames negative zero positive Term Editor Cancel Count Specify number of terms fuzzy sets related to the variable Type Select either the trapezoid or s function ILV term type When the ILV properties dialog is open for the existing variable the term type of the applied variable terms is displayed Other is displayed when term types for the selected variable differ Names Select default names of terms for the variable When the ILV properties dialog is open for the existing variable any applied terms are shown IMPORTANT When the ILV properties dialog is open for an existing variable then the terms count the type and the names are visible but you cannot change them in the properties dialog Publication LOGIX UMO04A EN P March 2007 82 FuzzyDesigner Projects Publication LOGIX UMO04A E
51. The component has two output links e Value link e Logical link The output logical link enables the connection of the component directly to other rule blocks If the component input link is connected to one rule block the output degrees of fulfillment are the same as the input degrees of fulfillment If the component is connected to several rule blocks the output degrees of fulfillment of reference membership functions are computed as a maximum of the corresponding input degrees of fulfillment Publication LOGIX UMO04A EN P March 2007 44 FuzzyDesigner Component Library sauwiildinkimann com The output Takagi Sugeno Variable component consists of functional terms Each functional term is defined by its parameters a a1 Ap and its name the parameters are entered in the specified order separated by the space character The type of every linguistic term can be different There are two supported functions e Linear function Rx4 X2 Xp do a aA X1 dz X2 Eent dn Xn e Constant function Rx1 X2 X n dp Where x1 X2 X are outputs from preceding components providing crisp values The component calculates an inference result as a crisp value y D dof f X Xn y lt i Za i where dof is DOF of i th term This value is finally limited to the range Wim Ymax If no term is activated DOF 0 the inference result is a user defined crisp default value Publication LOGIX UMO04A EN
52. Tool Bar The tool bar see 3D Graph Tool Bar has nine buttons Rotate Left Rotate Right Rotate Up Rotate Down Rotate CCW Rotate CW Zoom In Zoom Out and Zoom To Fit Window All of them have the function identical to the menu items described previously 3D Graph Tool Bar eA Y ANIA AA Publication LOGIX UMO04A EN P March 2007 3D Graph Control Mouse Dragging Drag the mouse to modify the 3D graph orientation The graph is rotated in the left right or up down direction if you hold the left mouse button and drag horizontally or vertically To rotate the graph CW and CCW press the SHIFT key the left mouse button and drag horizontally www klinkmann com Introduction Chapter 5 Fuzzy System Simulation FuzzyDesigner enables manual simulation of inputs and tracking of outputs generated by the fuzzy system This feature serves as a basis for off line tuning of a fuzzy system design parameters All internal variables are also monitored and displayed during the input simulation run Use Tools Simulation main menu command to start the monitoring FuzzyDesigner switches to Monitoring mode and a Simulation Watch dialog appears Values of all component variables used in the fuzzy system project are displayed in three tables in the Simulation Watch dialog see Simulation Watch Dialog The first table displays Input Ports and the third Output Ports All the remaining Intermediate Components are displayed in the second t
53. a user defined crisp default value Defuzzification SOM MOM LOM for singletons Output value is computed as a reference singleton with maximal term DOF If more terms have the same maximal DOF gt 0 then e SOM output smallest of the singletons with maximal DOF e LOM output largest of the singletons with maximal DOF e MOM output mean of the singletons with maximal DOF Waai kiinkinaniicoi FuzzyDesigner Component Library Term Editor ball angle O x SANOMA ee eae E A angle ie gative Zero positive Output term with maximal DOF zero 0 5 oo ATN ail Output Default Value if all DOFs 0 0 700 0 350 0 000 0 350 O 700 Output Yale 0 0000 Parameter Term DOFs negative ze positive defaults O 0000 O 6000 o 4000 o 0000 a A DOF zero is maximal Defuzzitication SOM MOM LOM for trapezoids Trapezoids are automatically transformed to singletons SOM MOM LOM 41 The output value is then computed in the same way as for singletons Recommendation e Use singletons to have easier insight to the output inference mechanism e No functionality is lost Publication LOGIX UMO04A EN P March 2007 42 FuzzyDesigner Component Library vankilana om Connections The input link of the output linguistic variable can be connected to a component providing the DOF value as a result of fuzzy inference that is the Rule Block component The output val
54. ab control page When you right click a tree view node a context sensitive menu appears When you right click the Projects node a context menu with the following commands appears see Projects Tree View Node Context Menu e New creates a new project see section Creating a Project e Open opens an existing project see section Opening an Existing Project e Close All closes all opened projects see section Closing a Project Projects Tree View Node Context Menu Mew Open Close All When you right click the opened project node a context menu with the following commands appears see Project Tree View Node Context Menu e Close closes the project see section Closing a Project e Save saves the project see section Saving a Project e Save As saves the currently active project to new file see section Saving a Project e Preview shows the project preview e Print prints the project see section Printing a Project Project Tree View Node Context Menu Close Save Save S Preview Print There are project component type nodes of the applied components under the node of each opened project When you right click the component type node a context menu with the following command appears see Component Type Tree View Node Context Menu Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com FuzzyDesigner Projects 69 e Delete All deletes all applied component
55. able Input Value group box Specify the range and input value of the Input Port selected from the Input Ports table e Minimum Display or set the lower limit of the selected Input Port e Value Display or set the value of the selected Input Port The track bar is an alternative to the entry field e Maximum Display or set the upper limit of the selected Input Port When you select the Input Port the range and current value are shown in the Input Value group box The Input Value group box is grayed out when you select a component from another table When you select an Input Port from the first table you can change its input value by clicking the slider in the Input Value group box or by entering a value in the appropriate entry field When you change the input value of the selected Input Port the fuzzy system project is statically evaluated that is the fuzzy system inference operation is computed When you disconnect the Input Linguistic Variable you can use the related Term Editor to set up its input value by the membership function editor slider You can use a variable as an additional input to the project Publication LOGIX UMO04A EN P March 2007 132 Fuzzy System Simulation 8 2011 www klinkmann com Close button Stops off line simulation and closes the Simulation Watch dialog Returns FuzzyDesigner to Design mode Simulation Watch Dialog a Simulation Watch ball Input Po
56. and control engineers to enhance the functionality of control and decision making systems in various branches of industry FuzzyDesigner includes a library of components you can use to design a fuzzy system that includes nonlinear input output mapping You can use a hierarchical structure to decompose a complex fuzzy system into smaller and simpler parts This reduces the internal complexity of a fuzzy model and results in fewer fuzzy rules and provides easier insight into the system operation Publication LOGIX UMO04A EN P March 2007 10 Get Started with FuzzyDesigner eanwildiniinanncom FuzzyDesigner is designed to work with Rockwell Automation s Logix5000 family of controllers A fuzzy system designed in FuzzyDesigner can be exported to an L5X Add On instruction AOD format You can then import the fuzzy AOI into any of your projects as needed Fuzzy AOIs can be used by any of the programming languages Function Block Diagram Ladder Logic or Structured Text With FuzzyDesigner you can also monitor and update the selected fuzzy AOI online directly in the running controller This is made available through the RSLinx OPC Server The Intended Use of FuzzyDesigner figure shows the underlying idea and intended use of the FuzzyDesigner software package used in designing Fuzzy Add On Instructions for Logix applications You can build smart components based on the expert knowledge encoded in fuzzy If Then rules You can use these compon
57. be IS Bn3 Wh where w 0 1 is rule weight of the k th consequent Schematically the rule can be rewritten as follows A13 A eee Apn gt B12 wy B77 w ony B 3 w This rule base format is very useful in manual design It can be represented in the form of a table where every column corresponds to one variable and rows of the table are filled with appropriate terms or optionally with their inversions applying the NOT operator FuzzyDesigner also supports the OR operator EEUU F X 1S A1 OR X IS Ayo OR JAND Xp IS A21 AND AND X IS A 1 THEN Y IS Byo You define the number of terms in the OR expression The NOT operator can be applied to the whole OR expression IF NOT IX IS A11 OR X IS A42 OR J AND X3 IS A gt AND AND X IS A 7 THEN Y IS B12 wy Yo IS B34 Wo ee The rule block component performs fuzzy logic inference based on fuzzy rules In a simplified way it computes degrees of fulfillment of consequent variables from degrees of fulfillment of premise variables by using fuzzy t norms and s norms t conorms FuzzyDesigner supports the following t norms fuzzy AND operators e Minimum Tyin x Vy min x y e Product Throg V X Y avoa kiinikinaniicoi FuzzyDesigner Component Library 49 FuzzyDesigner also supports this s norm fuzzy OR operator maximum Smax X V max x y The evaluation of the Rule Block is completed in three steps 1 DOFs of
58. between the Design mode and Monitoring mode see section FuzzyDesigner Control Basics e New Port New Input Port see section Input Port New Output Port see section Output Port e New Variable New Input Linguistic Variable see section Input Linguistic Variable New Output Linguistic Variable see section Output Linguistic Variable New Output Takagi Sugeno Variable see section Output Takagi Sugeno Variable New Intermediate Linguistic Variable see section Intermediate Linguistic Variable e New Rule Block see section Rule Block e New PID Controller see section PID Controller View FuzzyDesigner Main Menu View Structure Tool Bar Hide All Buttons w Status Bar Show All Bukbons w ree view w Create Mew Project w Open Project w Save Active Project w Undo w Redo w Refresh Active Project Go bo Monitoring Mode vw Preview w Print Hide Tree view 4 Mew Input Port Mew Input Linguistic Variable Mew Output Port Mew Output Linguistic Variable Mew Output Takagi Sugeno Variable Mew Intermediate Linguistic Variable Mew Rule Block Mew PID Controller 4 4 4 amp 4 amp amp amp amp amp 4 4 Help Publication LOGIX UMO04A EN P March 2007 woda hiinkinanicoi FuzzyDesigner Graphical User Interface 63 e Tool Bar Hide All Buttons hides all tool bar buttons of the application Show All Buttons
59. ch 2007 38 _ FuzzyDesigner Component Library MEA a Centroid Average CA generally An output value computed by this method is equal to the weighted average of the positions of the centroids of the output membership functions A weighted by their actual activation levels The output value is computed as follows where A c G e A cj is the maximum of the degrees of fulfillment over all y P cc the rules with the consequent A gt A c Cjis a position of the centroid of the membership function A which is calculated in advance e Mis a number of fuzzy sets A This method is used for applications when output is to be a continuous function of inputs for example a control system Maximum Center Average MCA generally This method is similar to the Centroid Average method except that c the center of maxima of B is calculated in advance This method is also continuous and allows the output value to reach the limits of the range Defuzzification CA MCA for singletons Term Editor ball angle O x YANGA eaa angle TE aia Sea negative DOF negative zero DOF zero positive DOF positive Output Ee DOF negatie DOF er DOF postive a 700 0 350 0 000 0 700 Output Default Value if all term DOFs 0 Output w alue 0 2667 B Term DOFs 0 0000 O 6000 O4000 0 0000 ee DOF positive Publication LOGIX UMO04A EN P March 2007 avew iinkinantico FuzzyDesi
60. control variable SO for angle actuator output add on instruction D pos_error REAL Float ball position error SP actual_position input velocity REAL Float velocity of the ball input bs h ct rt oar FURY ax je HH Ball_controller Ball Direct Nonlinear Fuzzy PD Controller of Ball amp Beam System F pos_error pos_error 00e velocity velocity 00e angle 00e e The order of parameters correspond to the port order of the fuzzy project angle you can optionally set it by the Tools gt Set Port Order menu item in FuzzyDesigner End i Ball A o x Direct MonlinefF ee PD Controller of Balg Bem System 7 a pos_ error O ate ha LL Non angle 0 f J n velocity velocity Monitoring and Updating a This section provides instructions on how to Project Online e establish online communication with a Logix controller to monitor and tune modify parameters the fuzzy AOI e what to do if online communication is not working correctly The communication with the Logix controller is via a RSLinx Classic OPC Server Before establishing online communication certain conditions must be met e The fuzzy AOI is generated by FuzzyDesigner e The fuzzy AOI is imported to an RSLogix 5000 project e An instance of the fuzzy AOI is created in the RSLogix 5000 project e The RSLogix 5000 project is downloaded to a Logix controller Publication LOGIX UMO04A E
61. ct Default Topic List Data Source Data Collection Advanced Communication Egg ashController M Autobrowse 2 Workstation EUCZPRGIEKOLINS1 een Linx Gateways Ethernet l AB_DF1 1 DFI Gl A OL 1756 163 LOGIxSS63 EggwashController H E Backplane 1756 4774 AB VBP 1 1789 417 4 Virtual Chassis E 00 Workstation EUICZPRGIKOLINS1 I fi 01 1789 L60 A Softlogix5860 Controller 1789 L 0 4 R13 31 Nem Clone Delete Apply Help The topic has to correspond to the correct controller path If not set the correct controller path in the communication path browser 4 Click Done Publication LOGIX UMO04A EN P March 2007 www klinkmann com Modifying Fuzzy System Parameters Online RSLogix 5000 Add On Instruction 145 5 To create a new OPC topic click New and enter a Topic name Set a correct path to the controller in the browser 6 Click Apply and then click Done When you modify fuzzy system parameters online parameters are overwritten directly in the selected controller that is in the selected fuzzy system Add On instruction tag The following conditions must be met e The number and the type of components have to be the same in both projects e The number of membership functions of every corresponding input linguistic variable output linguistic variable and intermediate variable has to be the same their names types and parameter values can vary e The number of consequent
62. e FuzzyDesigner is a project This chapter describes the concept of a project Topic Page Working with Projects 67 Designing a Fuzzy System 72 Fuzzy System Components 75 Term Editor 102 Term Properties Dialog 105 Rule Editor 108 Port Order Editor 113 Watch 113 History Graph 117 2D Graph 122 3D Graph 125 A project is a set of data containing information related to a particular fuzzy system Designing a fuzzy system involves several steps definition of components and links between components design of membership functions and creating fuzzy rules Most of the commands used for basic operations with projects can be found in the Project menu This menu contains the following commands e New creates a new project e Open opens an existing project e Close closes the currently active project e Close All closes all opened projects e Save saves the currently active project e Save As saves the currently active project to new file e Project Information shows the information about the currently active project e Preview shows the preview of the currently active project window Publication LOGIX UMO04A EN P March 2007 68 FuzzyDesigner Projects sioneidinkinanncom e Print prints the project window of the currently active project e Recent Projects shows the four most recent projects e Exit closes FuzzyDesigner All opened projects and their components are seen in the Tree View t
63. e a marked logical link from the appropriate list box of the selected tab dialog with applied logical links Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com FuzzyDesigner Projects 97 e Description tab dialog Specify the description of a new or existing RB Rule Block Properties Dialog Description Tab Dialog Rule Block bal Rules _ _ O x General Links Descnption Block Description rules Rule Editor Cancel Block Description Enter the description of the RB e Rule Editor button Click this button to open the Rule Editor see section Rule Editor where you can change the block rule base for example add or delete rules e OK button Accept the entered properties for the project e Cancel button Click this button to close the RB properties dialog Any changes made are not applied You can also click Close at the top right corner of the dialog PID Controller Use the Edit New PID Controller main menu command or the New PID Controller tool bar button to add a new PID Controller PIDC to the active project Alternatively use the New PID Controller project window context menu command to add a controller The PID Controller properties dialog appears and a default name is assigned to the component Click the OK button to add the component to the appropriate project All PID Controller names must be unique in the same project Publication LOG
64. e method of activation when multiple rules share the same term of a linguistic variable in the rule conclusion application A synonym for user programs Centroid Average CA A continuous defuzzification technique client An application that uses the services of an object crisp set A special case of a fuzzy set in which the membership function only takes two values commonly defined as 0 and 1 defuzzification A conversion of a fuzzy set into a numerical value Degree Of Activation DOA The degree to which the condition part of a rule is satisfied and consequently the degree to which the terms in the rule conclusion will be activated Publication LOGIX UM004A EN P March 2007 8 2011 150 Glossary www klinkmann com Degree Of Fulfillment DOF A general evaluation of the degree of truth of fuzzy logic expressions DOF values range from 0 to 1 For example e Term DOF the DOF for the expression X is A is the degree of membership of x in A e Rule DOF the DOF of the condition part of a rule extensible markup language XML Class of data objects and definitions that define how programs should behave XML documents are made up of parsed data characters forming character data or markup Markup defines the document s storage layout and logical structure FuzzyDesigner uses XML to define the format in which FuzzyDesigner projects are kept fuzzification Determination of degrees of membership
65. e of Fuzzy Logic FuzzyDesigner enables you to enhance the functionality of existing or new control and decision making systems in various branches of industry The fuzzy system designed and generated by FuzzyDesigner can be used in control systems for example as a direct nonlinear fuzzy rule based controller PID feedback control system supervisor or a process model in a Model Predictive Control scheme Input and output filters are used for signal preprocessing such as filtering deriving trends and many other functions that might add dynamics to the static I O map generated from fuzzy rules Input filters can also be designed in FuzzyDesigner Output filtering is an option and contains for instance a discrete integrator fed by the output of the Fuzzy Add On Instruction wow inkan eon Get Started with FuzzyDesigner 17 Nonlinear Fuzzy Rule Based Supervisor of a PID Controller Plant States FUZZY feedforward SUPERVISOR CONTROLLER The great advantage of fuzzy supervision is that it can be applied to existing control and there is little danger of making errors in design Most frequently used is a supervised PID controller where PID gains feedforward action or setpoints are being modified dynamically by rules depending on the process status and external conditions defined through setpoints Smart Switching Between Conventional Controllers Takagi Sugeno Controller FUZZY SUPERVISOR Plant State Schedule weight
66. e project e Parameter b Select the value of the parameter b entering the P b SP P V term of the PIDC equation The value of the parameter ranges from 0 to 1 e Dead Band Click the check box to activate the dead band of the controller and specify the dead band radius and dead band type e Dead Band Radius If the Dead Band check box is checked enter the dead band radius of the controller This value must be greater then zero e Dead Band Type If the Dead Band check box is checked enter the dead band type of the controller e Output Limiting with AntiReset WindUp Click the check box to enable this function and specify the upper and lower limits e Minimum When the Output Limiting check box is checked enter the lower saturation limit of the controller output e Maximum When the Output Limiting check box is checked enter the upper saturation limit of the controller output IMPORTANT H minimum value must be always lower then the maximum value Publication LOGIX UM004A EN P March 2007 wow inkan eon FuzzyDesigner Projects 101 Gain Forgetting Factor group box Specify the gain forgetting factor of a new or existing PIDC e Value Enter the controller gain forgetting factor value of the controller The factor default value is set to 1 for a new PIDC e Gain Forgetting Factor track bar Specify the gain forgetting factor setting from slow tracking minimum to exact
67. e variable DOFs are supposed e Temperature DOFtemp small 0 4 DOFtempdUarge 0 8 e Pressure DOFpress negative 0 1 DOFpress zero 0 9 DOFpress positive 0 5 Step 1 DOFs of all rules are computed Rule 1 DOF pje min DOF rempl smalh DOF press negative min 0 4 0 1 0 1 Rule 2 DOF e2 min DOF templ arge DOF press negative min 0 8 0 1 0 1 Rule 3 DOF jyje3 Min DOF press pOsitive min 0 5 0 5 Step 2 DOFs of all consequent variables for every rule terms are computed Rule 1 DOF 94 medium DOF je Weight o 0 1 0 9 0 9 DOF positive DOF Weight 0 1 1 0 0 1 Rule 2 DOF small DOF e72 Weight o 0 1 0 8 0 8 Rule 3 DOF small DOF je Weighty 0 5 1 0 0 5 DOF pyr positive DOF 1 Weight 0 5 1 0 0 5 Step 3 DOFs of all consequent variables terms for overall fuzzy system are computed DOF oie Small max DOF pp small for all rules max 0 08 0 5 0 5 DOF medium max DOF p medium for all rules max 0 09 0 09 DOF pr large 0 DOF Zero 0 DOF pyrr positive max DOF positive for all rules max 0 1 0 5 0 5 These steps are schematically shown on the following figures Publication LOGIX UMO04A EN P March 2007 wow klinkmnannicor FuzzyDesigner Component Library 51 Set of Rules Example Evaluation Procedure 1 small negative gt medium 0 9 positive 1 0
68. embership Functions appropriate output variables process according to fuzzy sets in the just active project e Set Port Order see section Port Order Editor e Watch see section Watch e Simulation simulate your process by entering values for the variables defined on input ports see Fuzzy System Simulation e 2D Graph see section 2D Graph e 3D Graph see section 3D Graph e Add On Instruction create or monitor fuzzy Add On Instructions see RSLogix 5000 Add On Instruction Publication LOGIX UMO04A EN P March 2007 wwa kilnikikianiicoi FuzzyDesigner Graphical User Interface 65 Window FuzzyDesigner Main Menu Window Tile Cascade Arrange Icons Minimize ll Maximize All Restore All 1 ball e Tile tiles all open windows in the application workplace e Cascade cascades all open windows in the application workplace e Arrange Icons arranges icons in the application workplace e Minimize All minimizes all windows in the application workplace e Maximize All maximizes all windows in the application workplace e Restore All restores all windows to their normal size in the application workplace Help FuzzyDesigner Main Menu Help Contents ShiFE F 1 Index ShiFE F2 Search ShiFt F 3 Product Activation About e Contents shows the help contents e Index shows the help index e Search shows the
69. en you select a fsp file type a file in the XML format with all project information will open When the selected file has no information about the project graphical representation or about FuzzyDesigner GUI the project will be opened with a default graphical representation and FuzzyDesigner GUI information without the project description for example 4 Click Open When the project opens successfully the appropriate tree view node will be added to the Tree View tab control page and the project window will be opened in the application workplace with appropriate recently opened dialogs FuzzyDesigner is a Multi Document Interface MDD application so you can open more then one project at a time Only one of the currently opened projects can be active so all the tool bar button commands associated with a project are applied to this active project The FuzzyDesigner active window does not automatically relate to the active project Changing the Active Project You can switch between projects by clicking the appropriate project window or project tree view node When a new or existing project is opened it is automatically set as the active project The name of the active project is shown in the title of the FuzzyDesigner main window Project Information Use the Project Project Information main menu command to open the Project Information dialog see Project Information Dialog for an active project Properties group box
70. ents in the many applications listed above Intended Use of FuzzyDesigner FuzzyDesigner From Expert Knowledge If x1 is low and x2 is medium then Y1 is fast Fuzzy Add on Generates Instruction XML Im po rted into RSLogix 5000 Control Systems controller supervisor process model Process Diagnostics process state classification Decision Making Forecasting decision support prediction model system For Publication LOGIX UMO04A EN P March 2007 www klinkmanin cor Get Started with FuzzyDesigner 11 A Fuzzy Add On instruction does not typically compete against standard controls found in Proportional Integral Derivative Controllers PID Fuzzy logic is a complementary tool and fills functional gaps not addressed in standard controllers such as PIDs or Model Predictive Controllers A development cycle of fuzzy logic solutions for Logix applications consists of multiple steps 1 Design the fuzzy system in FuzzyDesigner 2 Generate the fuzzy Add On Instruction 3 Integrate import and instantiate the fuzzy AOI to your RSLogix 5000 project 4 Monitor and tune the fuzzy AOI running in Logix online by using FuzzyDesigner Using FuzzyDesigner with RSLogix 5000 Software a RStogix 3000 FuzzyDesigner a JE a UEN es e Lae En pa i i z T phn eeri tone 1 0 iii 1 A Add On
71. erm properties Add Term Add a term The Term Properties dialog see section Term Properties Dialog appears A complete set of term properties is accessible Delete Term Delete a variable term You can also use the Delete key Inverse Term Add an inverse term for the currently selected variable term Input linguistic Variable only Convert Terms To Trapezoids Transfer all applied terms of the selected variable to trapezoids except for Intermediate Linguistic Variable and Output Takagi Sugeno Variable Convert Terms To S Functions Transfer all applied terms of the selected variable to s functions Unput Linguistic Variable only Convert Terms To Singletons Transfer all applied terms of the selected variable to the singletons Output Linguistic Variable only Next Term Selects the term adjacent to the current term TIP You can select a previous or next term from the currently selected one with the keyboard cursor keys To select a term directly click the term name area in the membership function editor Publication LOGIX UMO04A EN P March 2007 wow klinkmnannicors FuzzyDesigner Projects 103 Term Properties Show the Term Properties dialog see section Term Properties Dialog of the currently selected term IMPORTANT To show the Term Properties dialog of an already selected term double click anywhere in the membership function editor To cha
72. ermediate Linguistic Variable component Parameters e Name of the component e List of links to premise variables e List of links to consequent variables e Type of t norm PID Controller The fuzzy system PID component enables you to design an intelligent supervision of a conventional PID controller The following symbols and terminology are used in description of the component e PV process variable e CV control variable e SP set point e E error E SP PV e P proportional gain e J integral gain e D derivative gain e Man manually set value of CV e Mode controller mode Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com FuzzyDesigner Component Library 53 The component can be used as a conventional PID controller with supervised parameters defined by component input links The component output link provides a crisp value representing the control variable The component functionality is defined through the equation format with the option of using either independent gains or dependent gains When the independent gains option is used the proportional integral and derivative gains affect only their specific proportional integral or derivative terms respectively When the dependent gains option is used the proportional gain is replaced with a controller gain which affects all three parts Both formats are shown in the following equations CV A SP t PV t f
73. esigner Component Library Input Linguistic Variable trapezoid Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com Parameters e Name of the component e Vector b bo b1 Oml coefficients of the filter transfer function numerator b d optional e vector a a d coefficients of the filter transfer function denominator a d optional The fuzzy system Input Linguistic Variable component stores membership functions fuzzy sets of terms and is used for fuzzification classification of the component input a crisp value The component output is a vector of degrees of fulfillment of all terms for the crisp input or degree of overlapping for the input fuzzy set An Input Linguistic Variable component consists of linguistic terms Each linguistic term is defined by a fuzzy set that is by the membership function and the name There are four supported membership functions e Trapezoidal membership function e S shaped membership function e Inverse trapezoidal membership function e Inverse S shaped membership function inverse trapezoid a d s shape inverse s shape b oc a d x universe Linguistic terms are defined on specified range Xmin Xmax of discourse The component crisp input is fuzzified The result of fuzzification of the crisp input value x is a degree of fulfillment DOF of the terms which is computed for each term given by the membership function A x as follows
74. functions of every corresponding output Takagi Sugeno variable has to be the same The number of parameters of these functions have to be the same as well their values can vary e The number and structure of rules of every corresponding rule block have to be the same Weights of rules can vary e The names of fuzzy system input and output ports have to be the same as in the fuzzy system which was exported to the L5X file If these conditions are not met the fuzzy system in the controller may not be updated online Then you would have to export the fuzzy system to the L5X file and import it again to the RSLogix 5000 project If the fuzzy system parameters are successfully updated the following message appears x _l The project parameters were updated successfully Publication LOGIX UMO04A EN P March 2007 146 RSLogix 5000 Add On Instruction vaikina cam Modifying the fuzzy system parameters modifies a specific instance of the fuzzy Add On Instruction not the definition of the fuzzy Add On Instruction If you want to change the definition you should create a new version of the fuzzy Add On Instruction and import it onto your RSLogix 5000 project Importing an Add On FuzzyDesigner enables you to import the fuzzy Add On Instruction XML file L5X back to FuzzyDesigner To import the fuzzy Add On Instruction to FuzzyDesigner instruction click the menu item Tools gt Add On Instruction gt Import The following wi
75. g factor Output Port The Output Port component stores an actual output value of the HFS provided by a preceding component The output is a crisp value lan Connections The input link of the output port can be connected to all components providing a crisp value such as e Input Port component e Output Takagi Sugeno Variable component e Output Linguistic Variable component e PID component Parameters e Name of the component Publication LOGIX UMO04A EN P March 2007 www klinkmann com Introduction Chapter 3 FuzzyDesigner Graphical User Interface The FuzzyDesigner graphical user interface GUD is illustrated by a simple academic Ball and Beam experiment provided as one of the sample projects when you install FuzzyDesigner Topic Page setting Options 5 FuzzyDesigner Control Basics D9 The objective is to stabilize the ball at the desired position on the beam The main screen of the FuzzyDesigner GUI with the opened project named ball this is our Ball and Beam example can be seen below Ball and Beam Project in FuzzyDesigner FuzzyDesigner 16 00 ball Project Edit View Tools Window Help De hie oe PS RAS BISBR SRB BERRY Tree View AC Projects S E ball Input Forts a C9 velocity CF pos_error ea Input Linguistic Variables ee pos_error ball SON Mant POO 1 Non velocity velocity For Help press F1 Status Design Mode a Setting Options The
76. ght corner of the dialog Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com FuzzyDesigner Projects 79 Input Linguistic Variable Use the Edit New Variable New Input Linguistic Variable main menu command or the New Input Linguistic Variable tool bar button to add a new Input Linguistic Variable ULV to the currently active project Use the New Input Linguistic Variable project window context menu command to add a variable First the Input Linguistic Variable properties dialog appears Click OK to add the component to the appropriate project Input Linguistic Variable names must be unique in the same project e General tab dialog Specify the main properties of a new or existing ILV Input Linguistic Variable Properties Dialog General Tab Dialog AE Input Linguistic Yariable ballpos_ error Oo a x General Unit Range Terms Description Variable General Variable Mame pos_emron Input Link pos error id Term Editor Cancel Variable Name Specify the variable name Input Link Select a feasible ILV input link from the drop down list The link can realize the connection between the ILV and an Input Port Output Linguistic Variable Output Takagi Sugeno Variable or a PID Controller When the dialog for a new unconnected ILV is opened then all feasible input links are displayed Publication LOGIX UMO04A EN P March 2007 80 FuzzyDesigner Projects
77. gner Component Library 39 Defuzzification CA and MCA for trapezoids Trapezoids are automatically transformed to singletons CA centroid method MCA mean of maxima to allow reaching limit values of the range MCA CA CA MCA The output value is then computed in the same way as for singletons Mean of Maxima MOM generally This method computes the mean value of the interval at which the output fuzzy set reached the largest membership degree It is defined as follows A c max 4 c x y mean g INN F J Smallest of Maxima SOM generally This method is similar to the previous one Instead of mean value the minimum value of the interval is chosen The defuzzified output is computed as follows y smallest le A C Max A c J AD eco Publication LOGIX UMO04A EN P March 2007 40 FuzzyDesigner Component Library Publication LOGIX UMO04A EN P March 2007 www klinkmann com Largest of Maxima LOM generally The only difference to the previous method is that the maximum value of the interval is chosen The defuzzified output is defined as follows J _ largest e 4 c max A c Mean of Maxima Smallest of Maxima and Largest of Maxima methods are not continuous and are mainly used in applications on decision making and classification when the task is to choose from several alternatives If no term is activated DOF 0 then the inference result is set to
78. heck box and select the D gain link e D Gain Value When the D Gain Link check box is not checked enter the D gain constant value e Bias Link Click the check box and select the bias link e Bias Value When the Bias Link check box is not checked enter the bias constant value e Manual Control Select the manual control link e Mode Switch Select the mode switch link e Options tab control page Specify all remaining options of a new or already existing PIDC PID Controller Properties Dialog Options Tab Dialog PID Controller project0 PIDControllero Equation Format Independent Gains Dernvatiwe Input Process Variable Sampling Feriod fon Parameter b 1 00 Sore 4 bos the paramere in the foaloeing equaron Po See Dead Band Dead Band Radius om Dead Band Type Zero Lrossing O Output Limiting with Anti eset WindUp Minimum f Masimum fi Gain Forgetting Factor Slow Tracking e Features group box Set up all features of a new or already existing PIDC Publication LOGIX UM004A EN P March 2007 100 FuzzyDesigner Projects sioneidinkinanncom e Equation Format Specify the equation format of the controller e Derivative Input Specify the derivative input of the controller e Sampling Period Specify the sampling period of the controller The period must be greater then zero and is the same for all PIDs applied in the sam
79. inguistic Variable e Rule Block Publication LOGIX UMO04A EN P March 2007 76 FuzzyDesigner Projects Publication LOGIX UMO04A EN P March 2007 www klinkmann com e PID Controller Components of the same type used more than once in the same project must have a unique name The Input Port and the Input Linguistic Variable can share the same name The name of the Input Linguistic Variable Output Linguistic Variable Output Takagi Sugeno Variable and PID Controller used in the same project cannot be shared by the other components The Output Port can share names with the Output Linguistic Variable Input Port Use the Edit New Port New Input Port main menu command or the New Input Port tool bar button to add a new Input Port CIP to the currently active project Use the New Input Port project window context menu command to add a port First the Input Port properties dialog appears A default name is assigned to a new component Click OK to apply the component to the appropriate project All names assigned to Input Ports belonging to a single project must be unique e General tab dialog Specify the main properties of a new or existing IP Input Port Properties Dialog General Tab Dialog Input Port ballpos_ error E 5 x General Description Port General Fort Hame pos _eror Use Filter Butterworth Lowpass Filter Get Transter Function f Filter with Specific Transfer Function Bu
80. ion LOGIX UMO04A EN P March 2007 120 FuzzyDesigner Projects sioneidinkinanncom e Zoom out Zooms out of the graph The minimum size depends on maximum on an absolute scale value of all displayed curves e Zoom To Fit Window Changes the graph size to best fit the window e Zoom To Fit Base Vertical Range Changes the graph size to best fit the base vertical range in the window e Properties Opens the dialog with graph properties see Watch History Graph Properties Dialog The second context menu appears you right click the legend window In this menu you can change the parameters of the curve The menu structure is shown in History Graph Legend Context Menu The menu items are explained below History Graph Legend Context Menu Color Curve Vertical Range e Color Change the color of the selected curve e Curve Vertical Range Opens the dialog for changing the vertical range of the selected curve see Curve Vertical Range Dialog This range is used to map curves if the graph is relative and is valid only for this graph Curve Vertical Range Dialog 4 Set Vertical Range E B x Variable Description Name angle Unit Range Minimum E Masimum f Default Cancel Publication LOGIX UMO04A EN P March 2007 wow klinkmnannicor FuzzyDesigner Projects 121 History Graph Control Tool Bar The tool bar see History Graph Tool Bar has these three catego
81. iption Variable Description angle Term Editor Cancel Publication LOGIX UMO04A EN P March 2007 88 FuzzyDesigner Projects sioneidinkinanncom Variable Description Enter the description of the OLV e Term Editor button Click this button to open the Term Editor see section Term Editor where the default variable terms can be changed for example count and names e OK button Accept the entered properties for the project e Cancel button Click this button to leave the OLV properties dialog Any changes made are not applied You can also click Close at the top right corner of the dialog Output Takagi Sugeno Variable Use the Edit New Variable New Output Takagi Sugeno Variable main menu command or the New Output Takagi Sugeno Variable tool bar button to add a new Output Takagi Sugeno Variable OTSV to the active project Alternatively use the New Output Takagi Sugeno Variable project window context menu command to add a variable The Output Takagi Sugeno Variable properties dialog appears and a default name is assigned to the component Click OK to add the component to the appropriate project All Output Takagi Sugeno Variable names must be unique in the same project Publication LOGIX UMO04A EN P March 2007 wow diniinan eer FuzzyDesigner Projects 89 e General tab dialog Specify the main properties of a new or existing OTSV Output Takagi Sugeno Variable Properties Dial
82. isplays the upper limit of the y axis range e Grid Density group box Specify the grid density of the 2D graph Density Displays the selected grid density of the graph Select the density by clicking a point on the slider e Create button Generate the 2D graph see 2D Graph e Cancel button Close the 2D Graph Properties dialog without creating a graph You can also click Close at the top right corner of the dialog 2D Graph A 2D Graph No 0 ball 1 08 0 6 b4 O2 OF 02 OF G G 1 pos error The 2D graph can be controlled from the context menu or tool bar buttons Publication LOGIX UMO04A EN P March 2007 124 FuzzyDesigner Projects sioneidinkinanncom 2D Graph Control Context Menu The context menu appears when you right click the 2D graph and have the structure displayed in the 2D Graph Context Menu The meaning of all menu items is explained as follows 2D Graph Context Menu e Thin Thick Color Default Setting oom Ir Bounds Visible oom Out Bounds Color oom To Fit Window pee We ee arid Visible Graph Line d Grid Color Graph Ari b Piin hes y Values Visible Path b Values Color Default Setting Graph Properties 4 Active Point Visible Active Point Color w Path Visible Path Color Clear Path Default Setting e Zoom In Zoom in on the graph e Zoom Out Zoom out of the graph e Zoom To Fit Window
83. istic Variable OLV to the appropriate fuzzy system project Alternatively use the New Output Linguistic Variable project window context menu command to add a variable The Output Linguistic Variable properties dialog appears and a default name is assigned to the component Click the OK button to add the component to the appropriate project All Output Linguistic Variable names must be unique in the same project e General tab dialog Specify the main properties of a new or existing OLV Output Linguistic Variable Properties Dialog General Tab Dialog Output Linguistic ariable balh angle O x General Unit Range Terms Description Variable General Variable Hame Jangle Fuzzy Inference Algorithm MAM DAMI MINIMUM Detuzzification Algorithm CA Compute Output Fuzzy Set Term Editor Cancel Variable Name Specify the variable name Fuzzy Inference Algorithm Specify the fuzzy inference algorithm to be applied Defuzzification Algorithm Select the Defuzzification algorithm Compute Output Fuzzy Set Click this check box to generate a fuzzy set as the component output Publication LOGIX UMO04A EN P March 2007 wow inian eoi FuzzyDesigner Projects 85 e Unit tab dialog Specify the unit of a new or existing OLV Output Linguistic Variable Properties Dialog Unit Tab Dialog 4 Gutput Linguistic ariable ball angle
84. lolapsed y1 felapsed EXAMPLE The monitored fuzzy system is the Ball The data saved to the file are ball position error pos_error velocity of the ball velocity and the beam angle angle Every output variable has two columns the first one stores values computed by FuzzyDesigner the second one stores values loaded from the RSLogix 5000 project Data was saved at 12 37 27 4 time elapsed from the monitoring start was 37 seconds The defined history depth was 100 samples The both Sample Time CheckBox and Zero Based Sample Time in milliseconds CheckBox were checked Data was saved in the following form ms Controller Le T a TER es ee ee ee History Graph The History Graph is a Watch Tool component that displays measured or simulated data The creating and setting of initial parameters are explained in the section Watch The created graph can be either absolute or relative You can select the graph type by clicking the first button on the history graph tool bar and you can see this from a status bar at the top right corner of the graph window Publication LOGIX UMO04A EN P March 2007 118 FuzzyDesigner Projects Publication LOGIX UMO04A EN P March 2007 www klinkmann com If the graph is absolute then the displayed scaling of vertical axis is valid for all curves otherwise the scaling is valid only for the labelled base curve Its parameters are written in the separate window located in the left bottom co
85. lt OUTPUT LINGUISTIC VARIABLE NAME angle gt lt OUTPUT LINGUISTIC VARIABLE gt lt FUZZY SYSTEM gt Publication LOGIX UMO04A EN P March 2007 148 XML Format of a Fuzzy Project Publication LOGIX UMO04A EN P March 2007 www klinkmann com lt GRAPHIC MODEL VERSION 1 0 gt lt GRAPHIC MODEL gt lt GUI PROPERTIES VERSION 2 01 gt lt GUI_PROPERTIES gt lt FUZZYDESIGNER_PROJECT gt The prolog contains the XML declaration the XML description of the project designed in the FuzzyDesigner the version of the Fuzzy Core implemented in the FuzzyDesigner and time and date when the fuzzy project was exported to the XML document The Document element FUZZYDESIGNER_PROJECT contains information concerning the fuzzy system components element FUZZY_SYSTEM graphical representation of the fuzzy system element GRAPHIC_MODEL and GUI properties element GUI_PROPERTIES A more detailed description of the structure of the fuzzy system XML representation can be found in the FuzzyDesigner XML Project File Description document www klinkmann com Introduction Chapter 8 Glossary In this section you can find brief explanations on terminology used in this document activation A process by which the degree of fulfillment of a rule condition acts on an output fuzzy set aggregation An operation which combines several fuzzy sets to produce a single fuzzy set In the context of fuzzy control it is th
86. mbership over the universe of discourse for a given fuzzy set Multiple Document Interface MDD A Windows API for creating multiple window applications allows FuzzyDesigner to show multiple fuzzy systems or projects at the same time OLE for Process Control OPC An industrial standard allowing vendor independent access to industrial communication networks defined on OLE mechanism Output Port OP Stores values of a fuzzy system s output variables Publication LOGIX UMO04A EN P March 2007 152 Glossary Output Takagi Sugeno Variable OTSV Stores parameters of functional terms and computes the output value from degrees of fulfillment of terms Proportional Integral Derivative Controller PIDC Allows intelligent supervision of a built in PID controller Rule Block RB Stores rules and computes degree of fulfillment of rule conditions Rule Weight weighting factor RW A value from the interval 0 1 that states the degree of importance credibility confidence of a linguistic rule server The object that provides services Smallest of Maximum SOM A discontinuous defuzzification technique s norm The class of mathematical operations realizing fuzzy union OR t norm The class of mathematical operations realizing fuzzy intersection Publication LOGIX UMO004A EN P March 2007 AND Rockwell_ Software_RSLogix 5000_Fuzzy_Designer_en_0811 pdf Helsinki St Petersburg Moscow tel 358 9 540 4940 tel 7 81
87. mum of the corresponding input degrees of fulfillment Connections The input logical link of the Intermediate Linguistic Variable can be connected to a component providing a DOF value as result of fuzzy inference in a rule block that is the Rule Block component The output logical link of the Intermediate Linguistic Variable can be connected to components expecting a DOF value as result of fuzzification or fuzzy inference in a rule block such as the Rule Block component Parameters e Name of the component e List of terms defined by their names The fuzzy system Rule Block component stores rules performs fuzzy logic inference based on fuzzy rules and computes degrees of fulfillment of linguistic terms for consequent variables output logical links from degrees of fulfillment of linguistic terms used in the rule for premise variables input logical links Publication LOGIX UMO04A EN P March 2007 48 FuzzyDesigner Component Library Publication LOGIX UMO04A EN P March 2007 www klinkmann com Supported Format of Rules Multiple notations are used in the explanation of the supported format of rules e X4 X X premise variables e Y Y Y consequent variables e A 7 Aiz terms defined for the premise variable X e Biz Biz terms defined for the consequent variable Y C EXAMPLE IF X IS A33 AND X3 IS A2 AND AND X IS A1 THEN Y IS By gt w Y gt IS B71 Wol se
88. n E t dt D OO Bia CV P b SP t PV t g E dr DO se Bias Where parameter b can be defined by the user and has to be in interval 0 1 The default value is 1 This parameter dampens the influence of the setpoint on the proportional action The component allows two formats of derivative term O G Derivative input to the controller can be either the process variable PV or the error E t Use of the process variable eliminates output spikes resulting from setpoint changes Use of the error allows fast responses to setpoint changes when the algorithm can tolerate overshoots The algorithm is implemented in the discrete form Numerical integration is implemented as follows 1 Edt Ierm IE T lterm where T is the loop update time Numerical derivation is implemented as follows The calculation of the derivative term is enhanced by using a smoothing first order low pass digital filter This filter eliminates large derivative term spikes caused by noise in the process variable Publication LOGIX UMO04A EN P March 2007 54 FuzzyDesigner Component Library Publication LOGIX UMO04A EN P March 2007 www klinkmann com O AO 1 a ra aAO _ P l 162241 where D Finally the control variable CV is computed in the following way CV P E Iterm D AO Bias CV P E Iterm D AO Bias in the case of dependent gains and independent gains respectively The controller can be used in
89. ndow appears Import Add On Instruction Look in FuzzyD esigner c Ez My Recent Documents Deskho p EggWashMachine Lox My Documents ad oe EUILZPARGRPY TELKI User r io REE oo oo ae aa Places File name Files of type FSLogix 5000 Add On Instructions L5e Cancel Z Select the Add On Instruction L5X file you want to import to FuzzyDesigner Publication LOGIX UMO04A EN P March 2007 www klinkmann com Prolog Document Element Chapter XML Format of a Fuzzy Project FuzzyDesigner enables you to export and import the fuzzy project in the XML format In this chapter the basic structure of the XML document corresponding to the fuzzy project designed in FuzzyDesigner is shown An XML document consists of two main parts the prolog and the document element the root element lt xml version 1 0 gt ON a EPE E E E FE a PE 24E PAE ae se ES Sn TENSES me A EMESIS nee RATS TT lt FuzzyDesigner Project Description in XML format gt lt Version 16 00 05 gt lt Time 07 13 2006 14 41 00 gt Ced ge ee ee he ee ne Be een eae em mentee ee re _5 lt FUZZYDESIGNER_PROJECT SOFTWARE_VERSION 16 00 05 PROJECT_VERSION 2 01 00 gt lt FUZZY SYSTEM NAME FuzzyProject VERSION 2 01 gt lt INPUT PORT NAME velocity gt lt INPUT LINGUISTIC VARIABLE NAME velocity PROCESS FUZZY INPUT L gt lt INPUT_ LINGUISTIC VARIABLE gt
90. nection Panel 5 To stop monitoring click the Stop Monitoring icon The Watch window shows input and output values of the selected fuzzy system read from the selected controller Monitored Value as well as values computed in FuzzyDesigner Project Value Watch EgqgWashController Oo x ol i Le AX b Input Ports Intermediate Component Output Ports egg_dirt 0 40000 Water valve 0 15000 conveyorsp 0 00000 OU O0000 conveyorsp 0 2000 conweyor_ep 0 00000 Water Yale 0 15000 0 15000 water fowr 0 7000 oll valve pos 0 02587 oll valve pos 0 02587 0 02587 oil flow rate 0 6800 Publication LOGIX UMO04A EN P March 2007 144 RSLogix 5000 Add On Instruction aaeildinkmann com Configuring RSLinx OPC Server The OPC topic represents a specific path to a processor If an RSLogix Topi c 5000 project is saved then a default RSLinx topic is created with a default name that corresponds to the Controller Name To make sure that the selected RSLinx topic corresponds to the correct controller or to reconfigure or create a new RSLinx topic do the following steps 1 Start RSLinx software by double clicking the RSLinx icon on the desktop or by clicking the RSLinx Communication Service on the system Service Status toolbar a 2 In RSLinx software click the OPC DDE menu and choose Topic Configuration 3 Select the topic in the Topic List DDE OPC Topic Configuration re 2 x Proje
91. ner version 16 00 and later supports Rockwell Automation s Logix5000 family of controllers The fuzzy system designed using FuzzyDesigner can be exported to an RSLogix 5000 Add On Instruction AOD XML import file You can then import the fuzzy system into any of your projects as needed Fuzzy AOI can be used by any of the programming languages Function Block Diagram Ladder Logic or Structured Text With FuzzyDesigner you can also monitor and update the selected fuzzy AOI online directly in the running controller This is made available through RSLinx OPC Server Features Description Export fuzzy AOI Utility for export of designed fuzzy system into L5X file On line parameter change Changing parameters of a fuzzy system downloaded to the controller dynamically is enabled Real time fuzzy system monitoring Exact copy of the fuzzy system running on the PLC allows FuzzyDesigner to monitor all internal variables on the computer when both copies are fed with the identical inputs Integrated Design Some of the FuzzyDesigner features summarized in the preceding Environment IDE screen tables are shown in this section captures Publication LOGIX UMO04A EN P March 2007 www kllnkmanin eom Get Started with FuzzyDesigner 23 FuzzyDesigner Environment in Brief J FuzzyDesigner 16 00 Egg ashController Project Edit View Tools Window Help QR4 BISGERBE SR EBRR Y Dae lx PSg Tree View Gy EggWashController aa il
92. nge the parameters of the selected term membership function only right click the term parameter and ina click and drag operation move It to the new position Repeat as needed This applies to the Input Linguistic Variable and Output Linguistic Variable only Shift Selected Term Left Shift the currently selected term to the left Shift Selected Term Right Shift the currently selected term to the right Term DOFs Table Auto Arrange Optimize the width of the Term DOFs table columns Zoom Out Restore the membership function editor window setting meat Zoom in or out in the membership function editor position the mouse cursor anywhere in the editor area Hold down the right mouse button and drag your mouse to the right zoom in or left zoom out Release the button when the scale is at the desired level When you want to zoom any part of membership function editor click the begin point by the right mouse button and move the mouse right the right mouse button is still pressed down to the final point Release the right mouse button to rescale the membership function editor When the editor is zoomed click the right mouse button and move the mouse left or right to slide the membership function editor zoom To zoom out the editor click the right mouse button drag the mouse left and release the mouse button Discussed mouse operation must be performed inside the editor axes area Hide
93. nguistic Variable component e Output Takagi Sugeno Variable component only crisp values are considered e PID component only crisp values are considered The output logical link of the output Takagi Sugeno variable can be connected to components expecting a DOF value as result of fuzzification or defuzzification such as the Rule Block component Parameters e Name of the component e Range of the input value of the component Vin Vmax e List of functional terms described by The name The type of the function The vector of the function parameters dp 44 Apl for the linear function ago for the constant function e Default value Intermediate Linguistic The fuzzy system Intermediate Linguistic Variable component is used Variable as a buffer allowing logical chaining of rule blocks The component consists of linguistic terms with symbolic meaning Each linguistic term is defined by its name Degrees of fulfillment of all terms are results of previous logic inference in preceding rule blocks connected to this component Publication LOGIX UMO04A EN P March 2007 Rule Block 8 2011 www klinkmann com FuzzyDesigner Component Library 47 If the component input link is connected to a single rule block the output degrees of fulfillment just copy inputs If the component input is connected to several rule blocks the output degrees of fulfillment of stored linguistic terms are computed as a maxi
94. nnected to all components providing crisp values such as e Input Port component e Output Takagi Sugeno Variable component e Output Linguistic Variable component e PID component theoretically The output link of the PID component can be connected to all components expecting a crisp value such as e Output Port component e Input Linguistic Variable component e Output Takagi Sugeno Variable component theoretically e PID component theoretically Publication LOGIX UMO04A EN P March 2007 56 FuzzyDesigner Component Library vankilana om Parameters e Name of the component e Controller Gain parameters P I D Bias specified by links to components providing crisp values or by user defined constant values e Setpoint value SP specified by links to components providing crisp values or by user defined constant values e Sampling period T e Output limiting of control variable YES or NO e Output limits CV n CVa Gf YES min max e Equation format dependent or independent gains e Derivative input format error or process variable e Manual control None or the link to component which defines the manual control variable e Mode 0 automatic or 1 manual e Parameter b dampens the influence of the setpoint on the proportional action e Dead band YES or NO e Dead Band Radius Cif YES e Dead Band Type Zero Crossing or No Zero Crossing Cif YES e Gain forgettin
95. nnections The input link of the input linguistic variable is connectable to any of these components providing a crisp value e input Port component e output Linguistic Variable component e output Takagi Sugeno Variable component e PID component The output logical link of the input linguistic variable is connectable to components expecting a DOF value as a result of fuzzification or defuzzification such as the Rule Block component Parameters e Name of the component e Range of the input value of the component x min X mad e List of terms described by Name Type of membership function Vector of membership function parameters la b c dl Output Linguistic Variable The fuzzy system Output Linguistic Variable component stores output linguistic terms and is used for defuzzification The component has a logical input link degrees of fulfillment of all linguistic terms of the respective linguistic variable The link can be multiple meaning that the component can be connected to several rule blocks The component has two output links value and logical links Depending on the selected inference algorithm and defuzzification method the component computes a crisp value y Such a result provides an output value link The output logical link enables the connection of the component directly to another rule block If the component input link is connected to a single rule block the output degrees of f
96. nstruction instance Tag to monitor the OPC address space under a specific Topic the first branch under the tree root RSLinx OPC Server and click Connect On line Connection Wizard EggWashController Oj x Select Add On Instruction Instance Tag E ASLins OPC Server EqawashContraller _ Topic E Offline E Online 7 conveyor speed 7 conveyor speed di egg_dirt z oil flow rate i ol valve _pos_dif i water flow rate z Waker valve pos dit i Ee ee A AOI Tag E Program M amProgram a teres ae The AOI instance Tag can be accessed through a Topic tied to corresponding Logix controller If an RSLogix 5000 project is saved then a default RSLinx Topic is created with a default name that corresponds to the Controller Name To make sure that the selected RSLinx Topic corresponds to the correct controller or to reconfigure or create a new RSLinx topic follow the steps described in the Appendix K Configuring RSLinx OPC Server Topic If the AOI instance Tag cannot be found then check if the RSLinx topic is correctly configured as explained above If the selected AOI instance tag does not correspond to the fuzzy project in FuzzyDesigner the following message appears Error x x The selected OPC Item does not correspond to a proper Fuzzy Add On Instruction Instance Tag Publication LOGIX UMO04A EN P March 2007 142 RSLogix 5000 Add On Instruction aaeildinkimann com If the selected
97. o Oo x General Unit Range Terms Description Vanable Unit Predefined Vanable of none User Defined Unit rad Term Editor Cancel Predefined Click this radio button to select the unit of the variable from the list of predefined units e Variable of Select one of the predefined quantities which has the same meaning as the OLV e In Select one of the predefined units as the requested unit of the OLV User Defined Click this radio button to enter a user defined variable unit e Unit Enter the unit name of the OLV up to 100 characters Publication LOGIX UMO04A EN P March 2007 86 FuzzyDesigner Projects sioneidinkimanncom e Range tab dialog Specify the range of the variable and its default value Output Linguistic Variable Properties Dialog Range Tab Dialog Output Linguistic ariable balh angle _ O x General Unit Range Terms Description Variable Range Minimum 0 7 Masini 07 Default Value fo M Rescale Membership Functions of the Applied Terms Term Editor Cancel Minimum Specify the lower limit of the variable Maximum Specify the upper limit of the variable Default Value Set up the default value of the variable The default value must be within the specified range of the variable Rescale Membership Functions of the Applied Terms Click this check b
98. og General Tab Dialog ee pes variable Ee E General Uri Range Description Yanable General Variable Mame Takagig ugenoY ariablel Applied Input Links Pin Link Available Input Links none Add Pin Remove Pin Connect Term Editor Cancel Variable Name Specify the variable name Applied Input Links All pins and applied input links of the OTSV are shown here The component pins can be connected to the Input Ports Output Linguistic Variables Rule Blocks and the PID Controllers Available Input Links Select an available feasible input link for the variable Add Pin button Click this button to add the pin to the Applied Input Links table related to the OTSV Remove Pin button Click this button to remove the pin selected in the Applied Input Links table Connect button Click this button to connect the link selected in the Available Input Links combo box with the pin selected in the Applied Input Links table Publication LOGIX UMO04A EN P March 2007 90 FuzzyDesigner Projects vaikiinkmann com e Unit tab dialog Specify the unit of a new or existing OTSV Output Takagi Sugeno Variable Properties Dialog Unit Tab Dialog Output Takagi Sugeno ariable ball fakagiSugeno ariable 2 Oo x General Unit Range Description Variable Unit Predefined Variable of none it none C User Defined
99. oint Set Point Value i iv P Gain Link Ke P Gain Value Cc W Gain Link Ki hi Gain Value fo D Gain Link none D Gain Value fo W Bias Link Bias Bias Value fo Manual Control none Mode Switch aa Publication LOGIX UMO04A EN P March 2007 28 Get Started with FuzzyDesigner anwidininanncom Notes Publication LOGIX UMO04A EN P March 2007 sea i Chapter 2 FuzzyDesigner Component Library Introduction The FuzzyDesigner Component Library offers eight components from which you can efficiently build distributed fuzzy systems Topic Page Component Interface 29 Library of Components 30 Supported Membership Functions 30 Input Port 32 Input Linguistic Variable 34 Output Linguistic Variable 36 Output Takagi Sugeno Variable 42 Intermediate Linguistic Variable 46 Rule Block 47 PID Controller 52 Output Port 56 Component Interface The connection between components is called a link Generally a Hierarchical Fuzzy System HFS computes with data in the form of a crisp real value and or a fuzzy set Not all components enable both types of data to be transferred over the link The data type on both ends of a link should match FuzzyDesigner uses icons to define a link type as follows FuzzyDesigner Icons Icon Description Crisp value input or output value link input crisp values and crisp values resulting from defuzzification are transferred over the link Crisp value input or ou
100. oject Information menu item Description of input and output parameters The description is taken from the Description tab on the Input or Output Port properties dialog Publication LOGIX UMO04A EN P March 2007 136 RSLogix 5000 Add On Instruction aaeildinkimann com Importing Add On Instructions Yov can use fuzzy Add On Instructions in an RSLogix 5000 project to RSLogix 5000 Projects 1 Open the RSLogix 5000 project to which you want to import a fuzzy system Add On Instruction 2 Right click the Add On Instructions folder in the Controller Organizer Controlle Son Controller ball Controller Tags EE Controller Fault Handler ef Power Up Handler A 8 Tasks Sees MainTask ae MainProgram Unscheduled Programs Phases Motion Groups Add On I Data Typ New Add On Instruction in I Ne ee Trends L Import Add On Instruction Ta Conti P a nae mE FA Cut Certs Copy ctrl E FA Faste Ctri 3 Choose Import Add On Instruction The following dialog appears Import 4dd On Instruction Look in E FuzzpDesigner d re E Egg ashMachine Lox N Fz a hy Recent Documents Desktop My Documents ha EUCZPRGRPY TELKE1 Uzer r Ki File name i Notes Files of tyne RSLoaix 5000 Import E sport Files L amp Cancel Places Help E 4 Select the Add On Instruction L5X file you want to import to the RSLogix 5000 project
101. only the parameters vary Most often these functions are linear combinations of antecedent variables In control engineering each rule usually represents local dynamics in different state space regions and the consequent is given in the form of a state space or an ARX model The overall model of the system is achieved by fuzzy blending of these linear models FuzzyDesigner supports Takagi Sugeno fuzzy systems with linear functions in the rule consequents written in the following form R IF x is A and andx is A THEN y a a x 4 X or R IF x is A and andx is 4 THEN y a The Takagi Sugeno fuzzy system with the constant value in the rule consequents can be also considered as a fuzzy system with singleton membership functions in the rule consequents If the Centroid Average or Maximum Centroid Average defuzzification and Fuzzy Arithmetic Inference method is chosen than the behavior of both fuzzy systems is the same The fuzzy system Output Takagi Sugeno Variable component stores parameters of reference linear or constant consequent functions The component has two input links a logical input link degrees of fulfillment of all reference functions that can be multiple meaning that the component can be connected to several rule blocks and a value input link connectable to components that produce crisp values which can be multiple too The number of links depends on the number of consequent variables
102. ox to rescale the membership functions of all applied terms of the OLV IMPORTANT The minimum value must be always lower then the maximum value When the dialog for a new OLV is open the default range IS preset to 1 for minimum and 1 for maximum The default value is set to the middle of this variable range Publication LOGIX UM004A EN P March 2007 wow klinkmmannicors FuzzyDesigner Projects 87 e Terms tab dialog When the OLV properties dialog is open for defining the variable you can specify the variable terms in the same way as was explained for the Input Linguistic Variable Output Linguistic Variable Properties Dialog Terms Tab Dialog Output Linguistic ariable balh angle e _ OI x General Unit Range Terme Description Variable Terms Count f Type SINGLETON Hames negative zerg positive Term Editor Cancel Count the number of terms Type the type of terms Names predefined names of terms IMPORTANT When the OLV properties dialog is open for an existing variable then the terms count the type and the names are visible but you cannot change them in the properties dialog e Description tab dialog Specify the description of a new or existing OLV Output Linguistic Variable Properties Dialog Description Tab Dialog Output Linguistic ariable balh angle E 2 o x General LU rut Range Teme Descr
103. raph is shown immediately after you press the History Graph button The ranges of the relevant component predefine the axis and grid of the history graph automatically but you can change the range later You can change the curve parameters zoom of the graph graph legend visibility and vertical axis information as well The history samples are not interpolated but joined by abscissa e Save History Click this button to save the history into the specified file When a dialog see Watch Save History Data Dialog appears select the component values to be saved Other than the selected component value the time axis can also be stored Two time formats are supported relative and absolute time Click Save As at the bottom of the dialog to confirm the choice and the standard Save As dialog appears Select the file where you want to save the history and click Save e Clear History Click this button to clear the memory e Help Click this button to show the information about the Watch dialog Publication LOGIX UMO04A EN P March 2007 wow iniiaiee FuzzyDesigner Projects 115 e Close button Click this button to close the Watch dialog Watch Dialog TA Watch ball e al El me xX Input Ports Intermediate Componens Output Ports Name Project Value Name Project Value Name Project Value pos error velocity 0 000 Watch Data History Dialog Watch Data History ball x Watch Data His
104. rature and Humidity get their meaning In our case the current setting of the Temperature is such that it is both 85 Medium and 40 High Humidity is both 80 High and 50 Very High The first rule is thus 80 true for the current inputs while the second rule is 40 true when using minimum for the and operation The first rule states that if 100 satisfied the compressor should run at Medium speed Currently the first rule is only 80 fulfilled so one method of how to consider that the rule is only 80 fulfilled is to truncate the Medium fuzzy set for the output at the level 0 8 A similar situation happens with the second rule where High compressor speed is only 40 fulfilled As both rules are used at the same time their conclusions must be combined to get a fuzzy value for the output which is compressor speed The partially fulfilled Medium and High fuzzy sets are unified and a single fuzzy value is assigned to Compressor Speed As conventional control systems cannot deal with fuzzy values the fuzzy instruction includes conversion from a fuzzy to a crisp value For this case the center of gravity for the green area is computed and used to represent the original fuzzy value To summarize the designer has to e define input and output variables e cover the interval of the respective variable by fuzzy sets that is membership functions e write if then rules using labels of the fuzzy sets defined previously Potential Us
105. rature and humidity that are more suitable for describing the problem than numerical crisp values Fuzzy rules used in this example might be as follows lf Then Temperature is high and humidity is Speed is medium high Temperature is medium and humidity Speed is high Is very high Consider these factors when developing fuzzy rules e How do I specify High and other fuzzy values in fuzzy rules e How do the rules process numerical inputs provided by tags associated with sensors e How do the rules derive outputs from inputs e If the output generated is vague fuzzy how do I get the numerical crisp value at the output when needed Publication LOGIX UMO04A EN P March 2007 14 Get Started with FuzzyDesigner Publication LOGIX UMO04A EN P March 2007 www klinkmann com Crisp and Fuzzy For temperature readings you can classify a reading into three sets Low Medium and High Each set contains values in a given interval and the intervals do not overlap This means that a single reading or value is uniquely classified into one set degree of membership level of classification Classification Result Low Medium High 1 00 Medium 0 0 High Crisp Value Degree of membership DOM is a value describing how well the particular value of the variable in this case temperature fits the meaning of the label of the set Medium If the DOM is 1 the current temperature is understood as 100 Medium
106. re are some FuzzyDesigner features that can be configured according to personal preferences and for particular projects You can access the customization options from the View and Options menu commands Publication LOGIX UMO04A EN P March 2007 58 _ FuzzyDesigner Graphical User Interface T EE con Tool Bar The Tool Bar see FuzzyDesigner Tool Bar submenu on the View menu enables access to the commands for customizing the FuzzyDesigner tool bar A tool bar button can be set as visible or invisible by clicking the appropriate Tool Bar submenu commands FuzzyDesigner Tool Bar Deel PS AS TARE DERDREA R Status Bar The Status Bar see FuzzyDesigner Status Bar menu command on the View menu enables you to set the status bar of FuzzyDesigner as visible or as invisible The status bar shows FuzzyDesigner modes see FuzzyDesigner Control Basics and error messages FuzzyDesigner Status Bar For Help press F1 Status Design Mode E Iree View The Tree View see FuzzyDesigner Tree View menu command on the View menu enables you to set the Tree View tab control page visible or invisible see Working with Projects You can also do this by clicking the tool bar button with the same name shown as a tool tip To resize the Tree View tab control page use your mouse and drag the splitter on the right side of the tab control Publication LOGIX UM004A EN P March 2007 www klinkmann com FuzzyDesigner Control Basics
107. ries of buttons e Changing the nature of graph range relative or absolute e Zooming in and out of a graph e Working with graph main window The buttons are explained below History Graph Tool Bar Bidirectionally Horizontally Vertically e Relative Absolute Graph Switches the graph from relative to absolute and vice versa e Zoom In Zooms in on the graph You can select one of three zoom methods Zoom Bidirectionally Zoom Horizontally and Zoom Vertically The zoom method selected also applies to a zoom operation from dragging the mouse see section History Graph Control Mouse Dragging e Zoom Out Zooms out the graph See Zoom In for more details e Zoom To Fit Window Changes the graph size to best fit the window e Zoom To Fit Base Vertical Range Changes the graph size to best fit the base vertical range in the window e Show Hide Legend Shows or hides the legend window e Show Hide Base Curve Parameters Shows or hides the window with base window parameters Publication LOGIX UMO04A EN P March 2007 122 FuzzyDesigner Projects 2D Graph Publication LOGIX UMO04A EN P March 2007 www klinkmann com History Graph Control Mouse Dragging You can change the size of the displayed graph by using your mouse You can choose to zoom bidirectionally horizontally or vertically 2D and 3D graphs are useful tools for off line validation of a fuzzy system Assume
108. rner of the graph main window In the relative mode all curves are scaled to share the vertical range with the base curve The vertical range of each curve can be changed in the legend context menu This range is valid only for the current graph It is not the same as the range defined in the Linguistic Variable Editor Only the default range is the same The main graph window consists of three parts see History Graph the Real graph window the Legend and the Base Variable Property e Real graph window This window is located on the right side of the graph main window and it cannot be hidden Measured historical samples are plotted here Adjacent data points are linked by a line no interpolation is applied The number of curves displayed is limited to 1000 The grid of both axes is set automatically Range of displayed data can be changed by zooming in and out The units of the horizontal axis are written below the line The units of vertical range are not displayed as the plotted variables might not share the same units Right click the legend to display the graph context menu see History Graph Context Menu You can change the graph size or view the graph property window Legend This is located on the left top corner of the main graph window You can hide it by using a graph toolbar button The legend lists the displayed curves Related text items and curves share the same color If the curves overlap the color of the first related i
109. rojects 105 Output Term Editor for The Output Linguistic Variable Term Editor ball angle E Ioj x SOM OMA Arian E R angle negative Zero postive 0 700 0 350 0 000 0 350 OF 00 Output Value 0 0000 Parameter Term DOFs negative Ze positive lt default 1 0000 0 0000 0 0000 E a When you click the Add Term tool bar button of a variable Term Editor dialog the Term Properties dialog see Input Term Properties Dialog or Output Term Properties Dialog appears A default name is assigned to a new term e General group box Specify the main properties of the term Term Name Enter the term name Term Type Select the term type except for the Intermediate Linguistic Variable Parameters group box Specify the term parameters except for the Intermediate Linguistic Variable For the Output Takagi Sugeno Variable parameters are separated by a space Enter the linear term parameters in the order dO a1 a where a40 is a absolute term and a1 corresponds to the signal connected to pind For more information see the Output Takagi Sugeno Variable component description Publication LOGIX UMO04A EN P March 2007 106 FuzzyDesigner Projects sioneidinkinanncom e OK button Accept the entered properties for the project e Cancel button Click this button to leave the Term Properties dialog Any changes made are not applied You can also click Close at the top
110. roperties dialog is open for an existing variable then the terms count and the term names are visible but you cannot change them in the properties dialog e Description tab dialog Specify the description of a new or existing IMLV Intermediate Linguistic Variable Properties Dialog Description Tab Dialog Intermediate Linguistic ariable ball Intermediate aric _ Oo x fetes tL General Terms Description Variable Description Term Editor Variable Description Enter the description of the IMLV e Term Editor button Click this button to open the Term Editor see section Term Editor where you can change the default terms of the variable count names e OK button Accept the entered properties for the project e Cancel button Click this button to close the IMLV properties dialog Any changes made are not applied You can also click Close at the top right corner of the dialog Rule Block Use the Edit New Rule Block main menu command or the New Rule Block tool bar button to add a new Rule Block RB to the active project Use the New Rule Block project window context menu command to add a block Publication LOGIX UMO04A EN P March 2007 wow klinkmnannicor FuzzyDesigner Projects 95 The Rule Block properties dialog appears and a default name is assigned to the component Click OK to add the component to the appropriate project Rule Block names must be unique in the
111. rts Intermediate Components Output Ports Name Project Value Name Project Yalue pos eror velocity 0 0000 Name Project Value angle U 0000 angle 0 0000 Input Value Ml ininnuree 1 Value lo 0000 Masmum Publication LOGIX UM004A EN P March 2007 www klinkmann com Introduction Chapter 6 RSLogix 5000 Add On Instruction You can use Add On Instructions AOIs to deploy your fuzzy logic algorithm created with FuzzyDesigner Topic Page Generating a Fuzzy Add On Instruction 134 Importing Add On Instructions to RSLogix 5000 Projects 136 Monitoring and Updating a Project Online 138 Configuring RSLinx OPC Server Topic 144 Modifying Fuzzy System Parameters Online 145 Importing an Add On Instruction to FuzzyDesigner 146 With FuzzyDesigner you can use a fuzzy system in RSLogix 5000 software and run it on Rockwell Automation controllers Logix5000 family FuzzyDesigner enables you to export the designed fuzzy system to an Add On Instruction XML import file L5X You can then import the fuzzy system as a fuzzy instruction into any RSLogix 5000 projects The fuzzy instruction can be used by any of the programming languages Function Block Diagram Ladder Logic or Structured Text Publication LOGIX UMO04A EN P March 2007 134 RSLogix 5000 Add On Instruction saaiildinkmann com Generating a Fuzzy Add On You can export a fuzzy system to an Add On Instruction XML import file L5X Inst
112. ruction E To generate the Add On Instruction AOD for the active fuzzy project click Tools gt Add On Instruction gt Instruction Generator The following dialog appears Add On Instruction Generator Egg ashMachine 2 _ O x Properties ADI Name EgawashM achine Major Revizion Minor Revizion j Vendor IF ockwell Automatior Revision Mote Create In C NRSLogix AUOU Projects FuzzyDesigne Browse Create Cancel The following conditions must be met to generate a fuzzy AOT e All blocks of the fuzzy system must be connected e The fuzzy system must include at least one input port and one output port e All rules must be complete there is no rule with empty conditions or consequents e Each rule block must include at least one rule e Each linguistic variable must include at least one term The default file name of the Add On Instruction is taken from the name of the fuzzy system in FuzzyDesigner but you can change the name The instruction name must follow to the Add On Instruction naming rules It must e Be no more than 40 characters long e Start with a letter or underscore _ All other characters can be letters numbers or underscores Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com RSLogix 5000 Add On Instruction 135 In RSLogix 5000 software the Add On Instruction names are stored in the same namespace as all other built in instructions and data
113. s 0 1 CONTROLLER 1 Setpoints fh A ia 2 x laa lim i CONTROLLER EE lt x 3 PLANT Process Variables Another popular control structure with fuzzy logic is smart switching between local controllers A local controller is an analytical controller designed to work around specific process operation conditions Once the conditions change the rule based supervisor decreases the influence of one controller and gives more weight to another controller that has been designed to work in the new conditions Publication LOGIX UMO04A EN P March 2007 18 Get Started with FuzzyDesigner oanwildinidnanncom Feedback Control System with Direct Fuzzy Controller Control system status Primary controls Variables CONTROLLER Input filter Output filter Process Variables A fuzzy controller with the above structure typically handles multiple inputs and generates multiple outputs This system is recommended for experienced designers since control variables are direct functions of rules The number of rules increases rapidly with the number of inputs and fuzzy terms for inputs The problem of dimensionality can however be reduced by hierarchical structuring of the rule base of the controller which is supported by FuzzyDesigner Specifications and Features FuzzyDesigner features and specifications are summarized in the following tables For details refer to the subsequent chapters Fu
114. s A value that you must enter exactly as shown A variable that you replace with your own text or value Example programming code shown in a monospace font so you can identify each character and space A keyboard key Publication LOGIX UMO04A EN P March 2007 8 2011 8 Preface www klinkmann com Notes Publication LOGIX UMO04A EN P March 2007 a sancom Chapter 1 Get Started with FuzzyDesigner Introduction Topic Page Understanding FuzzyDesigner 9 Fuzzy Logic and Fuzzy Control Essentials 12 Specifications and Features 18 Understanding FuzzyDesigner is a software package for designing a fuzzy system to FuzzyD esi gner be implemented as a Hierarchical Fuzzy System HFS Fuzzy systems can be used in the following applications e Industrial automation and control systems e Process diagnostics and intelligent monitoring systems e Artificial intelligence e Decision making and forecasting Hierarchical Fuzzy System Non NON ee AON eg a Non lan DON ce an System_Tank_Level Dag Pump_1 Water_Regquest Level_ Modified Pumping Effortt2 new Water_Request 7 OON Pump_Control12 Fe O CE FE n Pump_2 Predicted Demand Predicted Demand Past_Pumping 2 Past_Pumping12 PO fet CS ca POA O m c n me n Mon NCA c aa yeaa Energy_Cost Level_ Manipulation pave He me a an a Pump_3 Pumping Effort34 OY Pump_Control34 FE AE aan Pump_4 Past_Pumping34 Past_Pumpings4 FuzzyDesigner enables application
115. s with the appropriate type from the project Component Type Tree View Node Context Menu Delete All There are component nodes of the applied components under the appropriate component type node of each opened project When you right click the component node a context menu with the following commands appears see Component Tree View Node Context Menu e Delete deletes the selected component from the project e Properties shows the properties dialog of the selected component see section Fuzzy System Components Component Tree View Node Context Menu Delete Properties Creating a Project Create a new project by using the Project New main menu command or the New context menu command of the Projects tree view node When a new project is created the appropriate tree view node and project window will be added A new project is not automatically saved To save a recently created project use the Save As command see section Saving a Project in the Project menu Opening an Existing Project Follow these steps to open the existing project 1 Select the Project Open main menu command or the Open context menu command of the Projects tree view node 2 From the Open dialog choose the appropriate file type fsp and click or type the name of the project file Publication LOGIX UMO04A EN P March 2007 70 FuzzyDesigner Projects Publication LOGIX UMO04A EN P March 2007 www klinkmann com 3 Wh
116. same project e General tab dialog Specify the main properties of a new or existing RB Rule Block Properties Dialog General Tab Dialog Rule Block bal Rules General Links Description Block General Block Name T norm Type Rule Editor Cancel Block Name Enter the block name T norm Type Set up the t norm type for the block Publication LOGIX UMO04A EN P March 2007 96 FuzzyDesigner Projects aaiildinkmann com e Links tab dialog Set up the RB input and output logical links Rule Block Properties Dialog Links Tab Dialog F Rule Block ball Rules SS lolx General Links Description Block Links Applied Input Logical Links Applied Output Logical Links velocity New Logical Link pos efor Add Link Delete Link Rule Editor Cancel Applied Input Logical Links Tab Dialog The list box shows all applied input logical links of the block Applied Output Logical Links Tab Dialog The list box shows all applied output logical links of the block New Logical Link combo box Select one of the available input or output logical links that can be used for the block Add Link button Click the button to add a new logical link selected by the New Logical Link combo box to the appropriate list box of the selected tab dialog with applied logical links Delete Link button Click the button to delet
117. sauwiildinkimann com e Unit tab dialog Specify the unit of a new or existing ILV Input Linguistic Variable Properties Dialog Unit Tab Dialog Input Linguistic Yariable ball pos_error General Unit Range Terms Description Variable Unit Predefined Vanable of Length Ir m C User Defined U rit linone Term Editor Cancel Predefined Click this radio button to select the variable unit from the list of predefined units e Variable of Select one of the predefined quantities which has the same meaning as the ILV e In Select one of the predefined units as the requested unit of the ILV User Defined Click this radio button to select a user defined variable unit e Unit Specify a unit of the ILV up to 100 characters e Range tab dialog Enter the operating range of the variable Input Linguistic Variable Properties Dialog Range Tab Dialog Input Linguistic Yariable ball pos error 2 s iol x General Unit Range Terms Description Variable Range Minimum E Maimun i M Rescale Membership Functions of the Applied Terms Term Editor Cancel Minimum Specifies the lower limit of the variable Maximum Specifies the upper limit of the variable Publication LOGIX UMO04A EN P March 2007 wow klinkmnannicors FuzzyDesigner Projects 81 Rescale Membership Functions of the Applied
118. sely the process variable matches the setpoint without changing the value of CV There are two choices of dead band type zero crossing and no zero crossing dead band 8 2011 www klinkmann com FuzzyDesigner Component Library 55 Zero crossing dead band control stops changing control variable ACV 0 when the process variable crosses the setpoint The control variable is not changed as long as the process variable remains within the dead band interval Zero crossing dead band control can be written as follows Once PV reaches SP E 0 and as long as PVe DBI use ACV 0 consider 0 No zero crossing dead band control stops changing control variable immediately when approaching the setpoint and crossing the band limits The control variable is not changed as long as the process variable remains in the dead band interval No zero crossing dead band control can be written as follows IF PVe DB THEN ACV 0 consider E 0 When the gain parameters of the PID controller are supervised by fuzzy rules rapid changes of premise rules variables may cause rapid changes of controller gains To avoid these changes the gain forgetting factor g can be used The value of g has to be in the interval 0 001 1 Value g 1 corresponds to exact parameter tracking no forgetting factor is applied and g 0 001 corresponds to very slow parameter tracking The default value is 1 Connections The Input links of the PID component can be co
119. ser Interface Main Menu Publication LOGIX UMO04A EN P March 2007 www klinkmann com FuzzyDesigner has two main modes Design mode and Monitoring mode FuzzyDesigner defaults to the Design mode where you can design the project set or reset options and use all application tools without any restriction Use the Monitoring mode when you need to change the project parameters only and leave the project design unchanged Options and tools that can enable project design changes are restricted in the Monitoring mode The current mode is indicated in the application Status Bar The Monitoring mode can be switched on or off directly from the Edit Go to Design mode Go to Monitoring mode main menu item or the tool bar with the same name in the tool tip The Monitoring mode is automatically enabled when the following dialoges are opened e Watch e Simulation Watch e Project Installation Wizard e Online Connection Wizard Panel e Monitoring Stand Alone Component After you close all of the open dialoges FuzzyDesigner switches to the Design mode You can minimize maximize or restore all windows opened in the FuzzyDesigner work area from the menu commands in the Window menu This section lists all menu items in the main menu bar of the FuzzyDesigner main window FuzzyDesigner Main Menu Project Edit wiew Tools Window Help www klinkmann com FuzzyDesigner Graphical User Interface 61 Project FuzzyDesigner
120. tem from the legend is applied Use a drag and drop operation to reorder the list of items The first item in the list is a base curve taken as a reference if the graph is relative in scale Right click the legend and the legend context menu appears see History Graph Legend Context Menu From this menu you can change the color or the vertical range tick marks related to a curve 8 2011 www klinkmann com FuzzyDesigner Projects 119 e Base Variable Property This is located on the left bottom corner of the main graph window You can hide it by using a graph toolbar button The window displays the base curve information the curve on top of the legend Unit name color and range are displayed History Graph fe History Graph ball Wi e S8A 8 8 w NOS err r angle variable angle Base Variable Absolute Graph Time 3 History Graph Control Context Menu Two context menus are available for history graph control The first menu shown below appears when you right click the graph The menu items are explained below History Graph Context Menu foom In E Bidirectionally Zoom Out d Horizontally zoom To Fit window foom To Fit Base Vertical Range Vertically Properties Bidirectionally Horizontally Vertically e Zoom In Zooms in on the graph The maximum size depends on the base vertical range and maximum on an absolute scale real displayed value Publicat
121. ter 5 Introduction 0 0 00000 a Chapter 6 IOMOCUCHOM pesa epa 6 hee ate ye Sues Be hed wee eh hd Generating a Fuzzy Add On Instruction Add On Instruction Parameters 0008 Importing Add On Instructions to RSLogix 5000 Projects Monitoring and Updating a Project Online Configuring RSLinx OPC Server Topic Modifying Fuzzy System Parameters Online Importing an Add On Instruction to FuzzyDesigner Chapter 7 POO sees eee bode ae Ye Eee ES Oe ee ee ee ee Document Element 0 a a cee eens Chapter 8 Introduction 0 00000 es www klinkmann com www klinkmann com About This Publication Who Should Use This Publication Conventions Preface Use this manual to understand how to best use the features in RSLogix 5000 software version 16 FuzzyDesigner This manual describes the necessary tasks to e build fuzzy systems as block diagrams from components of the FuzzyDesigner Component Library and use FuzzyDesigner functions to complete the project e use execute and monitor the designed fuzzy system on Rockwell Automation Logix5000 controllers e understand the fuzzy project and how you can export it to the XML format This manual is for application and control engineers to enhance functionality of control and decision making systems Text that is Bold Italic Courier Enclosed in brackets Identifie
122. ters the graphical and the internal parameters All the parameters are accessible from the block context menu see Blocks Context Menu This menu appears if you right click a block or a group of selected blocks You can change block appearance graphical parameters foreground color the color of the border background color font and line width www klinkmann com Fuzzy System Components FuzzyDesigner Projects 75 You can change the internal parameters that define the block function by clicking the Block Properties menu item or by double clicking the object You can also modify internal parameters by using the Tree View Blocks Context Menu Delete Foreground Color Background Color Font Line Width Default Setting i Thin Thick Block Properties Working with Text Use text to make comments in a fuzzy model To insert text select the Text Properties command in the Design Window context menu If you enter only spaces or don t enter any text the text object is not created Text objects are not selectable from the tree view To edit a text object double click the object or right click the text object and select the Text Properties command Texts Context Menu Delete Color Font Test Properties There are eight fuzzy system component types e Input Port e Input Linguistic Variable e Output Port e Output Linguistic Variable e Output Takagi Sugeno Variable e Intermediate L
123. tive 0 0000 0 1250 0 2500 0 3750 0 5000 H 3 fw 0 1871 g low high Input Value 0 11739 Parameter 4 0 3261 low low positive Term DOFs Ea low acceptable high extra_high 5 m 0 6739 acceptable zero 0 3261 0 6739 0 0000 0 0000 4 gt E For Help press F1 Status Design Mode Publication LOGIX UM004A EN P March 2007 24 Get Started with FuzzyDesigner www klinkmann com Project Tree view Tree View Se ml Projects E E Eqg ashController l a Input Ports en C4 egg_dirt ae C9 conveyor speed vee Ca water flow rate C9 oll flow rate ea Input Linguistic Wariat ef egg_dirt i conveyor speed ie water flow rate if oil flowrate Output Ports ED conveyorspeed sE water valve_pos ae D oilvalve_pos_dit E E Output Linguistic Vari eof water valve_pos ef conveyor speed eee oll valve_pos_dit S Rule Blacks fg RuleBlock1 FuzzyDesigner Environment Component examples Input Linguistics Variable 0417 Input Port elas conveyor speed conveyor_speed 0 426 A Publication LOGIX UMO04A EN P March 2007 Output Liguistics conveyor speed cit Dan fe s water valve _pos cif rE conveyor _ 3 0032 El water valve 0 010 wow klinkmnannicorn Get Started with FuzzyDesigner 25 FuzzyDesigner Membership Functions H Term Editor FuzzyProject Temperature E o x NANAS 8 e ew Temperature ver
124. to use the Main Menu or the Tool Bar Publication LOGIX UMO04A EN P March 2007 74 FuzzyDesigner Projects Publication LOGIX UMO04A EN P March 2007 www klinkmann com Selecting a Block Click an existing block to select it Other previously selected objects are automatically unselected To select multiple graphical objects draw a bounding box around them with your mouse or hold the CTRL key while clicking them To select all graphical objects blocks and texts use the Select All item from Fuzzy System Project Window Context Menu see section Fuzzy System Project Window You can also select a block through the Tree View Removing a Block A block can be removed from the graphical model using the keyboard or the Tree View Select a block and press the Delete key Blocks are removed from the project completely Moving a Block Graphical objects can be moved by using a drag and drop operation with your mouse or by using the CTRL and directional arrow keys on a selected object The final position of the object must lie inside the working area If one or more dragged objects are moved out of the permitted area then the objects are moved by default to the nearest permitted position Resizing a Block Select an object and move your mouse over to any corner of object Resize the object by clicking and dragging on a corner You can resize only one block at a time Block Properties A block has two sets of parame
125. tory Limit data samples to last OU Watch History Data Dialog S Watch History Data ball OCx Check History Data for Watch Dialog os Data Intermediate Data Output Data velocity Publication LOGIX UM004A EN P March 2007 116 FuzzyDesigner Projects saaiildinkmann com Watch History Graph Properties Dialog Watch History Graph Properties ball o 3 x Available Watch History Data Input History Data Intermediate History Data Output History Data od Show Cancel Watch Save History Data Dialog i Watch Save History Data ball x Available Watch History Data Input History Data Intermediate History D ata Output History Data o T d Additional Watch Data IW ero Based Sample Time W Sample Time ron Monitored data is saved in the ASCII format The structure of the file is shown in the following table where e Ny is the history depth e T is the sampling period folapsed S the relative time elapsing from the monitoring start factual is the absolute computer time Publication LOGIX UMO04A EN P March 2007 wow klinkmnannicors FuzzyDesigner Projects 117 Sample Time Time Output Name 1 vie ms lelapsed Ny 1 T actual Ny 1 T U lolapsed Ny 1 73 uz elapsed Ny 1 73 E y l telapsed Ny 1 T5 lelapsed I actual I na l Ud lelapsed l ae he y 1ltelapsed l telapsed Uy lolapsed up
126. tput value link crisp values are transferred over the link DOF value input or output logical link degrees of fulfillment of fuzzy terms of a fuzzy variable are transferred over the link to a rule block DOF value input or output logical link degrees of fulfillment of fuzzy terms resulting from rule block evaluation are transferred over the link to a fuzzy variable ABET Publication LOGIX UM004A EN P March 2007 30 FuzzyDesigner Component Library Library of Components E Supported Membership Functions Publication LOGIX UMO04A EN P March 2007 www klinkmann com The FuzzyDesigner Component Library offers the following components from which you can assemble fuzzy systems ranging from single input single output systems to multiple input multiple output systems with complex hierarchical structure of rules FuzzyDesinger Component Library Icons Icon Name Description H Input Port Preprocesses and stores values of a fuzzy system s input variables Output Port Stores values of a fuzzy system s output a m variables Input Linguistic Stores linguistic terms and is used for Variable classification of the actual component Input represented by a crisp value into the fuzzy sets defined for the respective linguistic terms In fuzzy control the process where the input is converted from a crisp value is commonly called fuzzification Rule Block Stores rules and computes degree of fulfillment of r
127. tterworth Lowpass Filter Filter Order Cutoff Frequency Filter with Speciic Transter Function Numerator Coefficients i Denominator Coefficients fo Reset Filter State Cancel www klinkmann com FuzzyDesigner Projects 77 Port General group box Enter general parameters of a new or existing IP Port Name Specify the IP name This name will be used as the input parameter name when the fuzzy algorithm is compiled to an Add On Instruction Use Filter The IP input value can be filtered by a user defined filter Click the Use Filter check box to set up the Input Port filter Butterworth Lowpass Filter Click this radio button to set the Butterworth Lowpass Filter parameters Filter with Specific Transfer Function Click this radio button to set the user defined filter parameters Get Transfer Function When you want to read the transfer function of the specified Butterworth Lowpass Filter ts numerator and denominator click this button The required numerator and denominator will be shown in the appropriate text boxes of the Filter with Specific Transfer Function group box Butterworth Lowpass Filter group box Specify parameters of a Butterworth Lowpass filter Filter Order If you selected the Butterworth Lowpass Filter radio button enter the required filter order here Cutoff Frequency If you selected the Butterworth Lowpass Filter radio button enter the
128. two different modes Manual mode and Automatic mode default mode During Manual mode the parameter Mode which is defined by the input link has to be set to 1 In this mode the controller calculates the user defined control variable which is connected to the input link Man During this mode the controller calculates the internal state of the integrator from the user defined control variable to achieve a bumpless transfer when the operator changes the control mode from manual to automatic In the Automatic mode when the parameter Mode is set to 0 the controller provides the computed value of the control variable If input links Man and Mode are not fed the default automatic mode is applied The component also provides the user additional features output limiting with anti reset windup dead band control and gain forgetting factor Output limiting allows applies limits CV_min and CV_max to the control variable If the output limiting is enabled and the computed control variable exceeds the limits the CV is saturated When the value of the computed control variable reaches or exceeds limits the integration is paused until the value of the computed control variable comes back into the range The adjustable dead band DB lets the user select an error interval DBI SP DB SP DB around the setpoint where the controller output does not change as long as the error remains within this range This dead band lets you control how clo
129. types system or user defined Therefore the Add On Instruction name must be unique Otherwise an error message will appear during the import of the fuzzy Add On Instruction When exporting the fuzzy Add On Instruction you can select the folder to which you want to save the L5X file You can also enter these optional parameters e Add On Instruction Major Revision e Minor Revision e Vendor e Revision Note Add On Instruction Parameters The Add On Instruction parameters correspond to the fuzzy system input and output ports The ordering of parameters follows the order of the ports Port order is determined by default by the order in which the ports are added to the project The port order can also be set from the Tools gt Set Port Order menu item The following parameter attributes are set to default values e Name The name of parameters are the same as the names of fuzzy system input and output ports e Usage The usage is set to Input for input ports or Output for output ports DataType The DataType ofallinput and output parameters is set to REAL Required The attribute Required is set to true that is the attribute requires an argument e Visible The attribute Visible is set to true that is the attribute is visible on the RLL and FBD Instruction face when the AOI is invoked Description The description is taken from the Description field on the Project Information dialog Project gt Pr
130. ue link of the output linguistic variable can be connected to components expecting a crisp value such as e Output Port component e Input Linguistic Variable component e Output Takagi Sugeno Variable component only crisp values are considered e PID component only crisp values are considered The output logical link of the output linguistic variable can be connected to components expecting the DOF value as a result of fuzzification or defuzzification that is the Rule Block component Parameters e Name of the component e Range of the output value of the component Vins Ymax e List of terms described by The name The type of membership function The vector of membership function parameters la b c d for the trapezoidal membership function c for the singleton membership function e Type of fuzzy inference e Type of defuzzification method e Default output value Output Takagi Sugeno Variable The classical model by Takagi Sugeno offers a fuzzy rule based smooth switching between analytical functions The consequent is a crisp function of the antecedent variables rather than a fuzzy proposition A general form of a Takagi Sugeno model is Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com FuzzyDesigner Component Library 43 The consequent functions f are typically chosen as instances of a suitable parameterized function whose structure remains equal in all the rules and
131. ule conditions Intermediate Bridges logical chaining of rule blocks Linguistic Variable Output Linguistic Stores linguistic terms and computes the output Variable value from degrees of fulfillment of stored terms defuzzification It implements the process of activation of output linguistic terms defined as fuzzy sets Output Stores parameters of functional terms and Takagi Sugeno computes the output value from degrees of Variable fulfillment of terms PID Controller Allows intelligent supervision of a built in PID controller Library blocks let you work with fuzzy sets as defined by membership functions Let x be the linguistic variable and A x be the degree of membership of x to the fuzzy set A defined by the sketched membership function FuzzyDesigner works with the following types of membership functions www klinkmann com FuzzyDesigner Component Library 31 Trapezoidal Membership Function with Parameters vertices a b c d 0 if x lt a x a b a if xe l a b A x 1 if xelb c x d c d if xe c d 0 if x gt d If a b then Ala 1 If c d then A c 1 Trapezoidal membership functions can be used in input and output linguistic variable components S shape Membership Function cubic spline with Parameters a b c d 0 if x lt a 2 2 3b a aye G 2 A 1 if xel b c 2 2 _3e d dao x d 5 if xe c d 0 if x gt d If a b then Ala 1 If c d then A c 1 S
132. ulfillment are the same as the input degrees of fulfillment If the component is connected to several rule blocks the output degrees of fulfillment of linguistic terms are computed as a maximum of the corresponding input degrees of fulfillment Publication LOGIX UMO04A EN P March 2007 8 2011 www klinkmann com FuzzyDesigner Component Library 37 The Output Linguistic Variable component stores linguistic terms Each linguistic term is defined by its fuzzy set that is the membership function and the name The following membership functions are supported e Trapezoidal membership function e Singleton membership function Linguistic terms are defined on the specified range lyin Ymax universe of discourse Defuzzification Defuzzification converts fuzzy sets to a crisp value taking into account their degrees of fulfillment FuzzyDesigner supports the following defuzzification methods Centroid Average Maximum Center Average Mean of Maximum Smallest of Maximum and Largest of Maximum y output variable Y universe of output variable defined by an interval AQ y crisp output value after defuzzification WWV A membership function the output term that is its fuzzy set A fuzzy set which is being defuzzified obtained as a union of all clipped output membership functions A y membership degree of variable y in fuzzy set A Publication LOGIX UM004A EN P Mar
133. y low medium high very_high TX NY XX _ ua 100 0 Parameter Value Input Value Term DOFs very_low low medium high very_high 0 5063 0 5196 0 0000 0 0000 Degree of Fulfillment window FuzzyDesigner Rule Base Rule Editor J Rule Editor Egqg ashController RuleBlock1 loj x i ar m 0 0000 high low Ze positive m 0 0000 high high high negative ZEO f mo 0187 E low high 1 negative 1 pz 0 3261 J law low positive 1 1 i cal 0 6739 E acceptable Zero 1 Zero 1 Publication LOGIX UMO04A EN P March 2007 26 Get Started with FuzzyDesigner inwildinianannccom FuzzyDesigner Rule Interfacing J Rule Block bal Rules x General Links Description Links Applied Input Logical Links Applied Output Logical Links velocity Hew Logical Link pos_erron Add Link Delete Link Rule Editor FuzzyDesigner Defuzzification Methods angle angle negative zero positive 40 negative zero positive bey NZ YSN A A 1 000 y y 0 000 0 500 1 000 y y y 0 500 0 000 0 500 1 000 MCA CA SOM MOM LOM Publication LOGIX UM004A EN P March 2007 wow klinkmnannicor Get Started with FuzzyDesigner 27 FuzzyDesigner PID Controller PID Controller WetBlend PID SS lolx General Options Description PID Controller Hame PID Input Links Process Y arable Temperature M Set Point Link Setp
134. zzy System Components Components are graphical objects blocks you work with to design a fuzzy system Component Membership AND Aggregation Inference Detuzzification functions Activation Type method if applicable Input Port Input Linguistic Trapezoidal Variable S shape and their inverses Rule Block Min product Max t norms Output Trapezoidal Max s norm Mamdani Fuzzy CA MCA Linguistic singleton Arithmetic MOM SOM LOM Variable Publication LOGIX UMO04A EN P March 2007 wow klinkmnannicor Get Started with FuzzyDesigner 19 Component Membership AND Aggregation Inference Detuzzification functions Activation Type method if applicable Intermediate Max s norm Linguistic Variable Output T S Max s norm Variable PID Controller Fuzzy System Analysis Tools Tool Description 2D 3D mesh plots Visualization of input output static mappings generated by the fuzzy system or Its specified subsystem Interactive plot control Color grid texture zoom and viewpoint management Tracing fuzzy system evaluation Marks output on the mesh when Input Is being changed FuzzyDesigner Mesh Plot A 3D Graph No 0 FuzzyProject E O x CPA AN RAA oo Velocity i e i Agt Publication LOGIX UM004A EN P March 2007 20 Get Started with FuzzyDesigner inwildinianannccom FuzzyDesigner Mesh Plot with Simulated Path 3D Graph No 0 FuzzyProject pi ATR a re ie
135. zzyDesigner Projects Publication LOGIX UMO04A EN P March 2007 www klinkmann com e Close button Click this button to close the Rule Editor dialog Generate Rules Dialog of Rule Editor Z Generate Rules ball rulebase Options Select the Premise Variables r M velocity Masmum Number of Aules E Mumber of Rules J T M Remove Just Applied Rules za Target Position Dialog of Rule Editor Choose Target Position Position Mo 3 Rules As Text Dialog of Rule Editor Rules as Text ball rulebase a o x Applied Rules 1 IF _ pos_eror 5 zero AND velocity 1S zero THEN angle IS zeralfi 2 IF pos _eror 5 negative THEN angle IS negativel 3 IF velocity 5 negatiwe THEN angle IS negatiwelli 4 IF pos_eror S positive THEN angle 15 posttiyel 1 5 IF velocity 15 positive THEN angle IS positivel wow klinkmnannicor FuzzyDesigner Projects 113 Port Order Editor Use the Tools Set Port Order main menu command to open the Port Order Editor dialog see Port Order Editor Dialog where you can change the order of the Input and Output Ports of the active project Port Order Editor Dialog 2 Port Order Editor ball Input Ports Output Ports Indes Name Index Name pos error 2 velocity ro e Input Ports table Select one of the applied Input Ports be e Output Ports table Select one of the applied Output Ports e

LOGIX-UM004A-EN-P RSLogix 5000 Fuzzy Designer User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents