ENPORT Model Builder: An Improved Tool for Multiport Modeling of
Contents
1. sion DSC VoL58 665 672 3 Breunese A P J J L Top J F Broenink and J M Ak kermans 1998 Libraries of Reusable Models Theory and Application Simulation 71 7 pp 7 22 4 MathWorks 1999 MATLAB 5 3 Simulink 3 0 Natick Massachusetts 5 Otter M and F Cellier 1996 Software for Modeling and Simulating Control Systems The Control Hand book CRC Press Boca Raton FL pp 415 428 6 Broenink J F 1999 20 SIM Software for Hierarchical Bond Graph Block Diagram Models Simulation Practice and Theory No 7 pp 481 492 7 Controllab Products 1999 20 Sim Reference Manual Version 3 0 Enschede Netherlands 8 Dynasim AB 1999 Dymola Dynamic Modeling Labo ratory Online Available http www dynasim com November 16 1999 9 Lorenz F 1997 Modeling System 1 User Manual Lorenz Simulation Liege Belgium 10 Integrated Systems Inc 1994 SystemBuild User s Guide Version 4 0 Santa Clara CA 11 Hales M K 1999 Computer Aided Engineering Tools for Structured Modeling of Mechatronic Systems Ph D Dis sertation Michigan State University 12 Hales M K and R Rosenberg 2000 Structured Model ing of Mechatronic Components Using Multiport Tem plates to appear Proceedings ASME IMECE 2000 Automated Modeling 13 Hales M K and R Rosenberg 2000 A Plan For Con trolled Access to Multiport Model Properties Proceed ings First IFAC Conference on Mechatronic Sy2. a Multiport Modeling envi ronment for mechatronic systems The environment is called ENPORT Model Builder or MB The modeling envi ronment is an attempt to allow for models to be con structed in ways that are congruent with the ways in which engineers like to work Additionally several novel features are supported including Multiport Templates filter based library organization and browsing tools and controlled access to Multiport Model properties The result of this work is a functional environment that enhances the model building effort System equations are formatted for solution in the MATLAB environment An earlier version of MB was successfully used in an under graduate mechatronics modeling course for mechanical engineering students at Michigan State University Future development of the MB environment includes plans to make the software more robust provide additional Graphical User Interface features enhance the Template capabilities and handle a larger class of modeling prob lems A concurrent effort is being made to complete a complementary environment for mmerically solving the system equations and displaying the solutions REFERENCES 1 Top J Akkermans H 1993 Layered Modeling of Physical Systems DCS Vol 47 Automated Model ing for Design pp 95 107 2 Rosenberg R Hales M and Minor M 1996 Engi neering Icons for Multidisciplinary Systems Pro ceedings ASME Dynamic Systems and Control Divi
3. dialog used for specifying the mem bers of the Owner Group 1 and Group 2 Groups This dialog can only be accessed if the Current User is a member of the Owner Group The Group for which the Users are to be specified is selected from the list on the top of the dialog Figure 7 indicates that the list of Users for Group 1 is currently being specified The list of Users of the currently selected group appears in the bottom half of the display Figure 7 also indicates that three Users currently belong to Group 1 Eda Group Lists Skaken Figure 7 Interface for Editing Group Members During the model building process an Owner User can set the access values for any access controllable object Figure 8 shows the dialog for defining access settings for a Comp o nent s Equations On the left hand side of the dialog the Group for which Access Values are to be set is selected On the right hand side of the dialog the access values for the currently se lected Group are specified The current options shown in Figure 8 indicate that any user belonging to Group 1 can only view the Equations associated with the currently selected Core Component Similar dialogs are available for specifying access settings for a Subsystem Component and a Model File Core Equation Acopss Settings Figure 8 Restricted Access to Equations CONCLUSIONS AND FUTURE WORK In this paper we have presented the results of the de sign and implementation of
4. general proper ties of the environment are discussed Section 3 presents the design and implementation of User Defined Model Types The functionality of a library of model types is discussed in Section 4 The concepts of controlled access and their implementation are given in Section 5 Conclusions and statements of future work are given in Section 6 GENERAL FEATURES In order to make the MB environment easy to use and ac cessible to a general user the design is based on a common Windows architecture The environment supports stan dard features such as an event driven Graphical User s Interface top level drop down menus toolbars and scrol lable resizable windows Whenever possible elements of the design such as the menu structure were chosen to mirror typical windows applications In addition to supporting common features of Win dows based applications MB also supports common mu l tiport modeling tools The fundamental model building block is a General Multiport as discussed by Rosenberg et al 2 Some characteristics of a General Multiport are that it can represent a physical system subsystem proc ess or effect it interacts with its environment through various types of interface structures called Ports it sup ports hierarchical model structuring by defining Subsystem Components that can contain other Components and it has display properties such as an icon that can be used in a graphical modeling environment The
5. Published in Proc Intl Conf On Bond Graph Modeling 2001 SCS v 33 n 1 152 157 ENPORT Model Builder An Improved Tool for Multiport Modeling of Mechatronic Systems Michael K Hales michael hales delphiauto com Delphi Automotive Systems 3900 Holland Road Saginaw MI 48601 Keywords Model Tools Model Types Software Tools Abstract Mathematical modeling based on multiport concepts is a powerful approach for simulation analysis and design of mechatronic systems Multiport concepts can be more fully exploited when they are embedded in a software envi ronment An effective modeling environment should not only support core ideas of multiport modeling but should also be easy to use flexible in design customizable and provide practical support for component libraries The environment should be implemented so that the engineer does modeling of mechatronic systems in a manner con gruent with his or her thinking In this paper we present the results of an effort to de sign and implement a multiport modeling environment for mechatronic systems many of the features described above have been realized Additionally several novel fea tures have been incorporated into the environment n cluding support of customizable user defined model types methods for organizing and browsing the contents of a model type library and tools for controlling access of classes of users to model information The result is a fully functional modeling en
6. cludes a pump and load The power of these modeling concepts is enhanced when embedded in software There are various classes of modeling software widely available and used today The MATLAB Simulink environment 4 is a typical example of a simulation tool that is widely used in industry today However while pro viding may useful features this tool has limited support for the modeling concepts referred to above Otter and Cellier 5 dis cuss several limitation of block diagram type of environments in general for physical systems modeling including the quirement of representing equations in a fixed causal form and the lack of close correlation between a physical system and its model Many other software tools have been developed that are more appropriate for modeling of mechatronic systems by di rectly supporting modeling paradigms that include combina tions of power and signal connections A representative list includes 20 Sim 6 7 Dymola 8 and MS1 9 These software packages provide many excellent tools In this paper we pres ent the results of research design and implementation efforts to produce a modeling environment that supports mechatronic design The modeling environment is called ENPORT Model Builder or MB Several novel and unique features have been realized including Templates for User Defined Model Types library organization and browsing tools and controlled access to selected model properties In Section 2 the
7. earching through a list of User Defined Model Types These meth ods are based on sorting by keywords by constraints and by generation history Sorting by Keywords When a user creates a new model type using a Mul tiport Template a set of Keywords may be associated with the Template Under standard conditions the list of Tem plates available to the user includes all Templates that have been previously defined However this list can be reduced by specifying a set of Keyword Filters The list of available keyword filters is automatically generated based on the keywords that have been defined in all available Templates After a user selects a subset of the available keyword filters and requests an update of the current list only those Templates that have been defined with match ing keywords are displayed Figure 4 shows a MB list of com ponents after applying the Signal keyword filter OMPOMANtS Keyeord Filters Power Conaenang Actetion la Source LJ iz Transducer I eyword Fitters J Figure 4 Sorting by Keywords There are several benefits to filtering by keyword First a set of Templates can be classified in multiple ways not just placed in a single group Second it is possible to find a Tem plate in a Library in multiple ways Third keyword associations are made according to the desires of an individual further per sonalizing the modeling environment Sorting by Constraints One
8. ected to signals and the output variable is the integral of the input variable When a new model is being built and an integrator is needed the user simply selects the integra tor block from the list of available components PDMTs are limited because the number of available types is fixed when the software development is completed UDMTs provide greater user flexibility The modeling environment can be specialized by defining a set of UDMTs that are the most useful to a par ticular individual The concept of UDMT is not new and has been imple mented in other software Examples are MATLAB s S functions Matrixx s User blocks 10 and 20 Sim s model types In the MB environment UDMTs are implemented and supported using Multiport Templates A Multiport Template hereafter referred to as a Template is an organization of three sets of data 1 General Multiport Definition All information needed to support a definition of any instance of a General Multiport 2 Constraints A set of Constraints specify the manner in which an Component instance may be modified and 3 Defaults Default values define an initial configuration of a Component instance A simple example will help to illustrate the use of the above information Consider a Bond Graph 1 Port C Com ponent defined as a UDMT in the MB environment illus trated in Figure 2 To support the commonly expected be havior of this component the following information is needed in
9. le be cause the classification structure is not unique For exa m ple consider the task of initializing an environment with a set of electrical components that only have Power Ports One way to go about this task is to first define a Parent Template that adds the single constraint that all the power ports must be electrical The next step might be to derive a new Template from this Parent Template and add a single Constraint specifying that there are no Signal Ports An equally valid option for obtaining the same results would be first to define a Template with no Signal Ports and then to use this Template to derive one that adds the Con straint that all Ports are electrical Both methods result in the nearly identical Templates only the generation history is different CONTROLLED ACCESS TO MODEL ATTRIBUTES A developing trend in industry is to outsource modeling efforts and to increase the sharing of models between custom ers and suppliers One of the concerns with this trend is the protection of proprietary information A simple solution com monly used is to prevent access to the entire contents of a model thus making the model a black box In this case use of such a model is greatly restricted Such an approach while protecting the investment of the model provider can be frus trating to engineers who use the model Models may be diffi cult to use correctly due to a lack of understanding of model details Additionally if changes t
10. limitation to using keywords as a searching device is that there is no automatic control over which keywords get associated with a Template This situation has the potential to be misleading and frustrating with searches leading to inap propriate Templates To overcome these difficulties the idea of using keywords a filter was extended to using the Con straints that are associated with the Template For Constraints that are sufficiently general it is be possible to search a library for Templates that have that Constraint For example a library could be searched for Templates that have the Constraint that No Power Ports are Allowed or for Templates with an Only Algebraic Equations constraint This scheme has many of the same benefits discussed with the keyword method An additional advantage to this idea is that the Constraints associated with a Template directly influ ence its functionality This result means that it isn t possible to have a search that produces a poorly matched Template That is the Templates found as a result of a Constraint search are guaranteed to exhibit the behavior specified by the Con straints Another advantage is that classifying models based on what Constraints are applied to them is a natural way to organize one s thinking about a set of models One weakness of this scheme is that the list of Constraints on which to search is limited to the types of Constraints known to the system Also the Const
11. o the model become neces sary due to design changes the model provider must be em ployed to implement the change The concerns described above were addressed in the de sign of the MB environment by supporting a concept termed Controlled Access Instead of restricting access to an entire model file a more sophisticated approach was used to control access to individual aspects of a model Instead of simply locking up an entire model only critical parts are protected More flexibility is provided and greater utility from a model is possible A detailed description of the controlled access fea tures of MB can be found in Hales 11 and Hales and Rosen berg 13 A condensed description follows The current MB implementation provides for controlling access to three modeling structures called Access Controllable Objects These objects and the types of access values they may assume are shown in Table 1 A None ac cess value for a particular component s Equations means that the component s equations cannot be viewed or modified a Read value means that equations can be viewed but not modified a Read Modify value means that Equations have no access restrictions A Subsystem with a None acess value means that the contents of the subsystem can not be view or modified A Open value means that the contents of a subsystem can be examined but not modified a Open Modify value means that there are no access re
12. raints are strictly limited to func tionality abstract classifications that are possible using Key words are not possible For these reasons it is useful to sup port searching both by keywords and by Constraints Sorting by Generation History A third classification scheme takes advantage of the Parent Child relationship between two Templates see Figure 3 This relationship sets up a natural ordering of Templates using a tree structure instead of a flat list An example of this sorting method in the MB environment is shown in Figure 5 0 Power E Capacitance c E Common Ettcrt E Cannon Flow EY Gyrete Ol heris R Resistance Disgiey B Generction History Figure 5 Sortin g by Generation History One benefit of this type of organization is the guidance given when searching for a component with a particular functionality To understand this benefit note that a de rived Template is a specialization of a parent Template Sorting by Generation History takes advantage of this fact If a candidate model located in a library is too general then a child Template can be sought that specializes the behavior in an appropriate way Conversely if a model found in the library is too restrictive the parent model can be considered Another benefit to this classification arises from the fact that the classification structure can be patterned after the thinking of the person building it This is possib
13. rent from a list of existing Templates The General Multiport Template is always available as a parent Tem plate 2 Specify Template General Properties General Template properties included defining whether the component sup ports model hierarchy or contains equations directly how the icon representing component instances is displayed a default label and keywords associated with the Template 3 Specify Power Port Properties Constraints on the num ber direction power type and causal properties of Power Ports are specified along with default values for each category 4 Specify Signal Port Properties Constraints on the number and direction of Signal Ports are specified Default values for these categories are specified 5 Specify Template Parameter Properties A set of default parameters can be associated with a Template Constraints can be given to limit the range of parameter values and to indicate whether or not the parameter can be deleted from a Component instance 6 Specify Template Equation Properties A set of de fault equations can be associated with a Template Constraints can be given to limit equation editing to the right hand side limit an equation to be linear al gebraic and or time invariant and limit the list of vari ables that may be used in an equation LIBRARY TOOLS The UDMT tools supported by MB help to create a flexible and customizable modeling environment By allow ing users to create and
14. rt with an existing Template A new template inherits constraints from the parent template The user defines addi tional constraints to define a more specialized Template This method produces a natural ordering among a set of Templates as shown in Figure 3 a The possible configurations possible from a given Template are always a strict subset of the possi ble configurations of its parent Template as shown in Figure 3 b Add Constraints EJ a Derivation Tree General Multiport Template General Multiport Template Template 1 ey b Possible Instance Configurations Figure 3 Derivation of Templates Consider again the Template for the Bond Graph 1 Port C Component This Template could be used to define a more restrictive Component a Bond Graph Linear 1 Port C Compo nent For the most flexible definition of a new Template a Gen eral Multiport Template is always available The General Mul tiport Template is a special Template that has no constraints or defaults defined For a more comprehensive treatment of the concept of Multiport Templates see Hales 11 and Hales and Rosenberg 12 In the MB environment a Template Wizard supports the Template definition process This tool simplifies Template crea tion by taking a user through a series of dialogs A summary of the Template definition process is given below 1 Select Parent Template This step allows the user to select a Pa
15. save sets of their own personal ized model types MB can be tailored to meet individual styles and preferences However this additional power creates a need for dealing with a large number of User Defined Model Types When building a new model find ing a useful model type for a current modeling need by searching through a large flat list with only the compo nent name and icon as a reference can be burdensome A typical method for handling this issue is to classify and group similar models in a library Often this ap proach amounts to grouping a set of models in one loca tion While useful to some degree using this organization scheme has several disadvantages First generally there are only three pieces of information that can be used to help locate a useful model the library name the model name and the model icon Second although a large list of models is divided up into smaller groups the contents of a single library can still become large Third there is only one way to locate a model by finding it in its library Mul tiple location paths are not possible Fourth the library organization reflects the thinking of the person who built the library which may not be the same as that of users of the library Many of these issues of library structure and organiza tion and use were addressed by Breunese et al 3 To assist the modeler in dealing with this issue MB provides three additional methods for organizing and s
16. ssignment of causality is made and checked the system equations are generated Sym bolic equation manipulation is applied where needed The current implementation of the MB environment formats system equations for solution in MATLAB A companion application for solving the MB system equations is cur rently under development i otage Se 1 GyY 4 1 Speed R Fricton heria Figure 1 The MB Environment USER DEFINED MODEL TYPES A powerful concept for storing engineering knowledge and reducing the model building effort involves reusing past mo d eling efforts in current modeling tasks One implementation of this idea is the use of User Defined Model Types UDMTs A UDMT is a generalized model definition that can be specialized for a particular purpose The type definition allows future modelers to take advantage of previous modeling efforts by automatically providing certain model details and reducing the required input to define a usable model A good model type provides sufficient details to capture the attributes of a class of related models but is flexible enough so that multiple realiza tions that are substantially different from each other are possi ble An analogous concept is a Pre Defined Model Type PDMT supported by most modeling environments available today A block diagram Integral block is common example of a PDMT The model type definition specifies that the model can only be conn
17. standard bond graph elements form a subset of the modeling components that can be represented by a General Multiport block dia gram components form another useful subset Fig ure 1 shows a representation of the MB environment and demonstrates several of the features discussed above In this figure a set of pre defined General Multiport mo deling Components appears on the left hand side of the screen Models are built by selecting Components from this list and placing them in a model window The figure also shows the top level representation of a model in the top right window labeled Modell This window contains a model of a common closed loop feedback system The model is hierarchical meaning that more detailed modeling information is found by looking inside the contents of a Subsystem Component As an example the contents of the Subsystem Component labeled Plant are shown in the window labeled Model1 Plant on the lower right hand side of the screen At some point in the model building process specific mathematical relationships and parameter values need to be defined MB provides an equation editor for specifying constitutive equations and parameter values The editor accepts equations in acausal form providing for flexible component definitions MB defines a set of common pre defined functions such as sin integrate etc and pro vides syntax checking After a closed system model is defined and an automatic a
18. stems Darmstatdt Germany September 2000 699 704 14 Stallings W 1998 Operating Systems Internals and Design Principles Prentice Hall Upper Saddle River NJ
19. stric tions A Model File with a None access value means that a model cannot be opened in the MB environment a Load value means that the Model File can be opened but not modi fied a Load Save value means that there are not access re strictions Table 1 Access Controllable Objects and Values Object Access Values Equations None Read Read Modify Subsystem Contents None Open Open Modify Model File None Load Load Save To implement these ideas the concepts of Groups and Users patterned after the UNIX operating system 14 was introduced At all times during a model building session a variable defining the Current User exists The Current User is a system parameter used to identify the person currently working in the modeling environment The value of the Current User Variable can be changed at any time which is equivalent to logging off and log ging on Next as shown in Figure 6 each Model File contains a set of Groups Four groups appear in the current design Owner Group 1 Group 2 and World In principle the number of groups is variable but is considered fixed in the current design of MB Associated with the first three groups is a unique list of Users i e each User is a member of exactly one group Model File Figure 6 Organization of Groups and Users Using this classification scheme the Current User is always identified as belonging to exactly one Group If the Curren
20. t User is among the Users on the Owner list then the Current User is considered an Owner If the Current User is among the Users on the Group 1 list then the Cur rent User is consider a Group User If the Current User is not contained on any list then the Current User is consid ered a World User The functional purpose of the data structure above is to classify the Current User as a member of one of the Groups When a new Model File is created the Current User is automatically added to the list of Users of the Owner Group The lists associated with the other Groups are initially empty When an existing Model File is opened then the Current User is identified as an Owner Group 1 Group 2 or World User depending on the user informa tion stored with the model Each access controllable object maintains an inde pendent set of access values for the Group 1 Group 2 and World Groups For example a Component can specify that Group 1 Users can read and modify its Equations that Group 2 Users can only read its equations and that Word Us ers can neither read nor modify its equations A User classified as an Owner always has complete access to all data in a model In addition an Owner can add and delete Users from any of the Groups The sole restriction is that the Current User cannot be removed from the Owner List If this option were available then it would be possible for a Model to be permanently locked up Figure 7 shows the MB
21. the Template 1 General _Multiport The General Multiport Definition contains power ports port variables causality prefer ence mathematical equations the equate integration multiply and divide operations and an icon definition 2 Constraints When an instance is created template constraints specify that a user cannot add or delete ports the state equation cannot be modified and the constitutive equation must specify a relationship be tween the port variable e and the energy variable q Constraints are not enumerated at this level of inter face The manner in which component constraints are accessed is discussed below 3 Defaults The default equation is a linear constitutive equation General Parametere Variati Emustons Defniiert ef qt iI Ql intaga o Figure 2 A Bond Graph 1 Port C Component The Template data defined above allow for different variations of the C Component to be created while pre serving the fundamental properties of its definition For example the constraints ensure that the constitutive equa tion cannot be modified to represent a dissipation effects that are associated with an R Component The use of constraints is a unique extension of the power of UDMT in the nultiport modeling methodology The Template data structure can be further exploited when deriving and organizing a set of new Templates In the MB environment when creating a new Template the user must always sta
22. vironment called ENPORT Model Builder An overview of the design and implementation of this software is presented INTRODUCTION Analytical tools are becoming increasingly important in industry Mathematically based tools are key to provid ing significant enhancements to product designs with decreased development lead time Mathematical modeling of mechatronic systems is at the forefront of this trend Companies are looking for methods that make investiga tion of the dynamic behavior of complex systems with components from multiple power domains as simple as possible Multiport modeling is an excellent method for modeling the dynamic response of complex mechatronic systems One benefit of multiport modeling is the ability to develop a model at distinct abstraction levels This concept was discussed by Top and Akkermans 1 further developed by Rosenberg et al 2 and applied to the structure of a Ronald C Rosenberg rosenberg egr msu edu Michigan State University 2555 Engineering Building East Lansing MI 48824 model library by Breunese et al 3 These concepts guide the model building process in a natural manner by letting an engi neer focus on relevant information at the appropriate time For example when modeling a hydraulic line the precise mathe matical relationship that defines the pressure drop across a bend in the line need not be considered when making deci sions about how to organize structure of a system that in
Download Pdf Manuals
Related Search