MICHBACK User`s Manual - Michigan State University
Contents
1. MICHBACK User s Manual Version 1 0 for DOS A d AVEN Load Deflection Ronald S Harichandran Cynthia M Ramon Gilbert Y Baladi MICHBACK USER S MANUAL Version 1 0 for DOS by Ronald S Harichandran Associate Professor Cynthia M Ramon Graduate Student and Gilbert Y Baladi Professor Department of Civil and Environmental Engineering Michigan State University East Lansing MI 48824 1226 March 2000 1995 Michigan State University Board of Trustees Table of Contents 1 Introduction oiecs ses vasescdscciaetesesspeseeacdevesevsdessneseabnshevsescd ones stock poseusstonoisesysdeossacegabienes 1 2 Summary of Modeling and Backcalculation scccccssssssscccsssssssssecsssesees 2 2 1 Modeling of the Pavement and Forward Calculations ceeeeeeesseceseeeeeeeeeenaeens 2 2 2 Backcalculation and Convergence Control eeccesscecesececeseeeceeceeceeececseceecseeeeeeaeees 2 3 System Requirements e ssseseecssssoccecessssccecessoosecessssoocecssssosocesssssccecessososcessssossse 3 4 Configuring the Computer essssssecoecssssscceccssoosececsssoocccssssoooccessssocecessossssesssesssee 3 41 JTyst llatign Procedure s 423 coccnestat snaeasensmecunas a a TE E aaseaeaaaes 3 42 The conto sys Pile ganestnnd meiir ene A ee ete See led re ee lah 3 4 3 Required Amount of Free MeMOry eecceesecesceceseeeceeeeeceeeeeceeceecseeeeceeeecsteee2. then this statement will need to be removed and the computer re booted by simultaneously press ing the CTRL ALT and DEL keys before running MICHBACK If available use of an ansi sys compatible device driver that can be unloaded from memory on demand is convenient since it elim inates the need to re boot the computer 4 3 Required Amount of Free Memory The MICHBACK program requires about 550 KB of free memory to run If the output file is to be viewed or sent to the printer from within MICHBACK then about 585 KB of free memory is required before running MICHBACK DOS and memory resident programs reduce the amount of free memory for use by other programs The amount of free memory available can be checked by using the DOS command MEM or CHKDSK If there is insufficient free memory then memory resident programs will need to be removed before running MICHBACK If there is insufficient memory to load the program a message similar to that shown below will be displayed Program too big to fit in memory 4 4 Printing Graphics Graphic screens produced by MICHBACK can be dumped onto an attached printer if the DOS command GRAPHICS COM is issued after the computer is turned on and before MICHBACK is used It may be convenient to include the command in the AUTOEXEC BAT file so that it is issued every time the computer is turned on To download graphics that are on the screen to the printer simply press the SHIFT and PRSCR keys simulta
3. 1 2 Do you want to change default weights y n n SENSOR LOCATION Default Layout 1 Sensor No 1 2 3 4 5 6 7 Sensor Location inches 0 8 0 12 0 18 0 24 0 36 0 60 0 Default Layout 2 Sensor No 1 2 3 4 5 6 7 Sensor Location inches 0 12 0 24 0 36 0 48 0 60 0 72 0 nter 1 or 2 to accept one of the default layouts nter 3 for any other configuration w m Figure 10 Form for sensor locations and weights User Defined Layout 3 Sensor No 0 I 2 3 4 5 6 Sensor Location inches 0 8 0 12 0 18 0 24 0 36 0 60 0 a Sensor layout 3 and default weights User Defined Weights Sensor No 0 I 2 3 4 5 6 Weight 1 0 1 0 1 0 1 0 1 0 1 0 1 0 b Sensor layout 1 or 2 and user defined weights User Defined Layout 3 and Weights Sensor No 0 1 2 3 4 5 6 Sensor Location inches 0 8 0 12 0 18 0 24 0 36 0 60 0 Weight 1 0 1 0 1 0 1 0 1 0 1 0 1 0 c Sensor layout 3 and user defined weights Figure 11 Possible extensions to Sensor Location Weight form 12 STATION IDENTIFICATION Station 5 Starting Mile Post 2 0 Ending Mile Post 8 0 Number of Data Sets Max 4 4 Figure 12 Form for station identification and number of data sets 5 5 6 Returning to the main menu Selecting Return to Main Menu option 0 from the Data File Menu returns you to the main menu Throughout MICHBACK the 0 option returns you to the parent menu 5 6 Entering Deflection Data using the Keyboard
4. Deflection data may be entered manually by selecting Enter Deflection Data using Keyboard option 4 from the Main Menu The form shown in Fig 12 is then displayed Entries for the station number starting milepost and ending milepost are optional The number of deflection basins that are to be entered manually is specified in the last field Up to four deflection basins can be entered at a time When the F1 key is pressed a form similar to that shown in Fig 13 is displayed The number of columns will equal the number of data sets specified on the previous form The sensor readings for each deflection basin are entered in the columns Deflections may be positive or negative the oretically negative deflection can arise for thin pavements The average FWD drop load and ECTION DATA mils Sensor No Data Set 1 Data Set 2 Data Set 3 Data Set 4 24 37000 22 27000 15 04000 19 63000 20 87000 18 79000 11 63000 17 03000 18 50000 16 48000 9 43000 15 42000 15 58000 13 49000 6 67000 13 33000 13 10000 11 04000 4 53000 11 59000 9 50000 7 42000 1 71000 8 93000 5 53000 3 50000 33000 5 66000 Load lbs 9000 9000 9000 9000 AC Temperature F 79 77 80 65 Figure 13 Form for specifying deflection data 13 pavement temperature for each test are also entered on this form The pavement temperature must be entered if temperature correction was requested otherwise it is optional Upon exiting from this form the manually ent
5. 5 5 4 Specifying layer names thicknesses and Poisson s ratios Selecting Specify Layer Thicknesses and Poisson s Ratios option 4 from the Data File Menu displays the form shown in Fig 9 for entering the layer names thicknesses and Poisson s ratios A thickness cannot be input for the last layer even if it overlies a stiff layer The depth to COMPOSITE PAVEMENT Which layer is PCC slab 2 Subbase Type 1 GRANULAR 1 STABILIZED 2 Figure 8 Extension to form for pavement type 2 10 ER THICKNESSES AND POISSON S RATIOS Layer Thickness Poisson s No Name inches Ratio 1 AC 13 5 35 Base 11 0 40 3 Roadbed 45 Allow program to improve layer thickness y n n Depth to stiff layer can be specified on Stiff Layer Properties screen SUGGESTED VALUES FOR POISSON S RATIO AC 0 35 Concrete 0 20 Granular 0 40 Cohesive 0 45 Figure 9 Form for layer names thicknesses and Poisson s ratios the stiff layer is specified in a different form All thicknesses must be greater than zero and Pois son s ratios must be less than 0 5 MICHBACK allows one layer thickness to be backcalculated along with the layer moduli Typically this will be the roadbed soil or AC thickness which have the most impact on the back calculation If the roadbed soil thickness i e the stiff layer depth is fixed and another layer thick ness is to be improved type y for the last
6. Graphically on Screen Statistical Information Return to Main Menu Selection _ Figure 24 View results menu COMn where n is the parallel or serial port number to which the printer is connected The output file can always be printed from outside MICHBACK using the PRINT command Option 4 Used to view results graphically on the screen when backcalculation has been performed at two or more test locations You may view overlaid profiles similar to deflec tion profiles which show the percent variation of each layer modulus from the correspond ing modulus backcalculated using the representative deflection basin or a single profile which shows the variation of any layer modulus along the analysis segment Typical exam ples are shown in Fig 26 a and b On color systems each curve in the overlaid profiles is shown in a different color along with a legend e Option 5 Used to plot cumulative distribution functions CDF s of the AC modulus and the peak deflection Examples of these are shown in Fig 27 a and b The CDF s are SUMMARY OF BACKCALCULATED MODULI psi Layer Average Standard Maximum Minimum Layer Name Modulus Deviation Modulus Modulus 1 AC 567508 187015 918713 251865 2 Base 47163 3332 9 139658 6927 3 Roadbed 18259 4719 24498 8784 Backcalculated at individual test locations Lto Ten Average Std Dev Maximum Minimum Stiff Layer Depth inches 603 L98 720 154 Press an
7. a test location that best represents the collec tion of deflection basins at all test locations in a pavement segment A repaired pavement in which a previously constructed concrete slab has been rubblized prior to pouring the AC layer Weights assigned to each sensor measurement during backcalculation Weights smaller than 1 0 may be used to indicate that a particular reading is perceived to be inaccurate Weights assigned using the sen sor layout form are used at all test locations One or more sensors at arbitrary test locations may be assigned a zero weight by performing deletions Initial estimates of layer moduli that are used to begin the backcalcu lation process Good seed moduli usually reduce the computation time For well behaved deflection basins the backcalculated layer properties are not sensitive to the seed moduli However for some de flection basins usually involving stiff composite pavements the back calculated layer properties may depend on the seed moduli A treated subbase A real stiff layer usually consists of bedrock However some analysts prefer to use an artificial stiff layer to approximately account for the stress stiffening nature of the roadbed soil The stiff layer depth is measured from the pavement surface i e top of the AC layer A location at which an FWD test is performed 28
8. effort of many individuals Conceptual de velopment and overall guidance for the development of MICHBACK Version 1 0 was provided by Ronald S Harichandran Tariq Mahmood and Gilbert Y Baladi Most of the coding was performed by graduate students at Michigan State University Cynthia M Ramon coded the user interface pre processing of deflection data and post processing of results Weijun Wang coded most of the analysis routines An initial formulation of a gradient based approach for backcalculating layer moduli was provided to the research team by Robert A Raab from the Strategic Highway Research Program Lynne Irwin from Cornell University improved the numerical quadrature in our version of the CHEVRON program Baoyan Wu assisted in porting the program to UNIX systems MICHBACK Version 1 0 for DOS was developed with funding from the MDOT and the University of Michigan Transportation Research Institute as part of the project entitled Reduction of Rutting under Heavy Vehicle Loads with Gilbert Y Baladi as Principal Investigator and Ronald S Harichandran as Co Principal Investigator Members of the Technical Advisory Com mittee for this project especially Dave Smiley Jack DeFoe Kurt Bancroft Jerry Sweeney and Ish varlal Patel provided helpful suggestions regarding the user interface Appendixes A Temperature Correction When requested correction of the backcalculated AC modulus to the standard temperature of 77 F is performed usi
9. fewer sensors are used then one of two default sensor configurations may be selected later g 10 Temperature Corrections Required y n If the backcalculated asphalt concrete AC modulus is to be corrected to the standard temperature of 77 F then type y otherwise type n If temperature corrections are requested then the AC modulus corrected according to a commonly used version of the Asphalt Institute equation see Appendix A is reported in the output file Please be warned that deflection data in general does not support the As phalt Institute temperature correction equation a warning screen advising the user of this is displayed when creating a new data file 11 Change Convergence Control Parameters y n If the default parameters that control the modified Newton algorithms and convergence are to be changed then type y The de 66 29 fault is n If y is typed for item 11 then when the Initial Data form is completed the Convergence Control form shown in Fig 6 will be displayed with the default values shown in bold typeface The convergence control parameters may be changed as outlined in Section 2 2 The last line of the Convergence Control form relates to the generation of seed values for the moduli and stiff layer CONVERGENCE CONTROL 1 Calculate Gradient Matrix After 3 Iteration s 2 Maximum Number of Iterations 10 Min 10 3 Modulus Tolerance 1
10. question on the form The form will then be extended requesting you to specify the layer whose thickness is to be improved and the thickness specified for this layer is used as an initial seed value 5 5 5 Specifying sensor locations and weights Selecting Specify Sensor Locations option 5 from the Data File Menu displays a form sim ilar to that shown in Fig 10 if seven or fewer sensors are used To select one of the default sensor layouts type 1 or 2 for item 1 To select a custom layout type 3 By default MICHBACK weights all sensors equally in the backcalculation process If you suspect that one or more sensors are erroneous and wish to de emphasize their readings at all loca tions in the backcalculation then type y for item 2 Depending on the values typed for items 1 and 2 of the form an extended version of the form may be displayed when the F1 key is pressed The possible extensions to the form are shown in Fig 11 The extension in Fig 11 a is displayed if only the sensor layout is to be changed the extension in Fig 11 b is displayed if only the weights assigned to the sensors are to be changed and the extension in Fig 11 c is displayed if both the sensor layout and the weights are to be changed If the specified number of sensors is greater than seven then a form similar to that shown in Fig 11 c replaces the one shown in Fig 10 11 SENSOR LOCATION WEIGHT DATA 1 Select sensor layout 1 2 or 3
11. then the seed moduli specified on the form are used for the first backcalculation and the backcalculated moduli are used as seed values for the subsequent backcalculation MICHBACK uses internal regression equations to obtain an initial estimate of the depth to a stiff layer The estimated value is displayed with a prompt as shown in Fig 21 You may type y or n at the prompt to include or exclude a stiff layer in the analysis If the backcalculation is being performed for a deflection basin generated by an elastic layer program without using a stiff layer or if the estimated stiff layer depth is very large say more than 400 in then you may wish to ex clude the stiff layer If a stiff layer is to be included in the analysis then a form similar to that shown in Fig 22 is displayed You may change the modulus Poisson s ratio and estimated depth of the stiff layer and indicate whether MICHBACK should optimize the estimated stiff layer depth The stiff layer depth cannot be optimized if the thickness of a layer other than the roadbed soil is to be improved since MICHBACK currently allows the optimization of only one layer thickness In many cases allow ing the program to vary the stiff layer depth gives better results However for deep stiff layers the backcalculated moduli may not be too sensitive to the stiff layer depth MICHBACK does not al low the stiff layer depth to exceed 720 inches since larger depths have a negligible effect on t
12. value for is 1 0 Often only for theoretical deflection basins generated by an elastic layer program can the iteration be carried on until the RMS error in the deflections is extremely small For deflection basins measured in the field it will usu ally not be possible to obtain an arbitrarily close match between the computed and mea sured deflections If a larger RMS error is acceptable then can be increased 2 The percentage error in each successive moduli and perhaps a layer thickness estimate is smaller than a specified tolerance This criterion may be expressed as Bit 1 BL x 100 lt e k 1 2 n EX Gi fit 1 _ Fi E x 100 lt e where i is the estimated modulus of layer k at iteration i and 7 is a layer thickness at iteration 7 The default value for is 1 0 This criterion indicates that the iterative scheme has stabilized and that the unknown variables are not changing significantly from one iteration to another Increasing the value of is not recommended For most problems the default of 10 for the maximum number of iterations is sufficient However if convergence does not occur after 10 iterations then you can increase the maximum number of iterations and try again 3 System Requirements The MICHBACK program was originally written for personal computers running the DOS operating system Soon it will also be available for Sun and Hewlett Packard workstations running the UNIX ope
13. 00 4 RMS Tolerance Desired 1 00 5 Suppress Internal Generation of Seed Values y n n Figure 6 Form for changing convergence control parameters PAVEMENT TYPE Flexible L Composite 2 AC on Rubblized Concrete 3 Pavement Type to be Analyzed 1 Figure 7 Form for pavement type depth to begin the iteration procedure and is used only as a safety mechanism Usually the seed values for the moduli are generated internally using built in regression equations but may be changed before backcalculation begins While every effort has been taken to ensure that the regres sion equations do not cause a run time failure for unusual situations a run time failure causing the program to crash may occur If this should happen then the internal generation of seed values may oo 99 be turned off by typing y for item 5 5 5 3 Specifying pavement type Selecting Specify Pavement Type option 3 in the Data File Menu displays the form shown in Fig 7 The type of pavement for which backcalculation is being attempted is specified by typing 1 2 or 3 respectively If 2 is typed to specify a composite pavement then when the F1 key is pressed the form is extended by appending the lines shown in Fig 8 Data may now be entered only on the extended portion of the form You must identify which layer consists of the PCC slab by typing 2 or 3 and whether the subbase is granular or stabilized by typing 1 or 2
14. The data file must have deflections for at least two locations for the profile to be displayed If the file has deflections only at one location then the deflection basin is displayed rather than the deflection profiles Occasionally a deflection data file may have embedded comments duplicate drops or other anomalies that may cause problems when the data is read If this should happen an error message is displayed at the bottom of the screen and a message such as the following is displayed below it Delete line containing FORCE DIFFERENCE DETECTED MEASUREMENTS MAY BE IN The deflection data file must then be edited to remove the line containing the string displayed with in quotes so that it can be read by MICHBACK Note however that such errors may indicate a 1ECTION DATA FILE Created by KUAB old Software Created by SFWD new Software User Created ASCII file See User s Manual for format Return to Main Menu Selection _ Figure 14 Deflection data file type menu 14 Route MSU 29F Deflection mils 18 16 14 12 18 8 6 4 2 6 4 18 26 36 46 56 66 Test Location Number 1 Zoom In Out 2 Identify Analysis Locations 3 Delete Readings 4 Backcalculate 6 Return to Main Menu Selection _ Figure 15 Typical plot of deflection data potentially serious problem at the related test location and you should carefully consider whether to use the data at this loc
15. ation for backcalculation Each curve in the profile plot see Fig 15 shows the deflections measured by a sensor along the length of pavement that was tested and is plotted in a different color on color monitors In order to have a meaningful comparison between deflections at different test locations the deflections at each test location are linearly scaled to correspond to the target load This scaling is done only for the profile plot Backcalculations at different test locations are actually performed using the aver age drop load at each location The X axis indicates the test location number which is sequentially assigned to the data in the file The test location in feet or meters from a reference point as en tered by the FWD operator during testing is displayed above the X axis at major tick marks In Fig 15 the test location does not increase monotonically but is somewhat erratic This figure is selected to emphasize that the test location entered by the FWD operator may not be meaningful It is recommended that the FWD operator be instructed to enter the test location distance from a fixed reference point and that the tests be performed in sequence beginning from the closest to the farthest location This will facilitate the physical interpretation of the data presented in the plot In essence the deflection profiles indicate the variation in the measured deflections along the pavement section The plot is useful for identifying weak l
16. cludes stiff layer Maximum number of layers is 4 if stiff layer is to be incorporated in analysis Figure 5 Form for initial data 2 Filename to Output Results to The output from the analysis will be directed to this file The output is in standard ASCII form and may be viewed or edited using any text editor Output needed for graphical plots are stored in a file having a plt file name extension and the same base name i e the name without the extension as the output file For example if the output file is named sample out the plot file will be named sample plt Route District Control Section Descriptive alphanumeric data Lane Job No Descriptive alphanumeric data Description 1 Descriptive alphanumeric data Description 2 Descriptive alphanumeric data pe SY at A Number of Layers The number of layers in the pavement section A maximum of five layers are permitted If a stiff layer such as bedrock is to be included in the analysis then the number of layers is limited to four and does not include the stiff layer Load Plate Radius Radius of the FWD loading plate Number of Sensors The number of sensors to be used in the backcalculation Some FWD devices have a series of sensors behind the loading plate and another sensor in front or to the side of the loading plate Only the sensor at the middle of the plate and the remaining sensors behind the plate should be included in the analysis If seven or
17. converted to the chosen system When this option is chosen the following prompt appears below the menu Select System of Units as Press E for English Units or S_ for SI Units Press E to select English or British units or S to select the international system of units For new data files the default is English units Avoid changing back and forth between English and SI units since this can cause some changes in the data due to rounding errors 5 5 2 Specifying initial data and convergence control Selecting Initial Data option 2 in the Data File Menu displays the form shown in Fig 5 typical data is also shown in bold typeface The data that should be entered into the fields are described below 1 Filename to Save Data to All data that is entered in this and other forms is stored in this file The data may be recovered at a later time and modified if necessary INITIAL DATA 1 Filename to Save Data to sample dat 2 Filename to Output Results to sample out 3 Route MSU 29F District 12 Control Section 3 4 Lane 2 Job No 6 5 Description 1 Test pavement section designated MSU 29F 6 Description 2 Pavement has moderate alligator cracking 7 Number of Layers 3 Max 5 8 Load Plate Radius 5 91 inches 9 Number of Sensors 7 Max 10 10 Temperature Correction Required y n n 11 Change Convergence Control Parameters y n n Number of layers includes roadbed and ex
18. d data and manipulate the data prior to backcalculation MICHBACK also computes useful statistics of the backcalculated results Alternatively you may analyze individual deflection basins by manually entering the measured deflections The following files are included in the release of MICHBACK Version 1 0 readme An ASCII file containing installation instructions michback exe The executable backcalculation program sample dat Sample input data file kuab fwd Sample KUAB FWD deflection data file created by the KUAB software The sample dat file should be used with this deflection file ae gic re 5 sample out Sample output file generated by using sample dat and sample fwd and backcalculating at test locations through 18 The stiff layer depth was optimized at each test location 6 sample plt Sample plot file corresponding to sample dat and sample out 7 ascii fwd The ASCII template for translating deflections measured by devices other than the KUAB FWD The supplied files can be used to generate screen images of most of the figures in this User s Man ual without having to perform an analysis Once these files have been copied into your system the program is started by typing michback Four basic procedures are required when using the program 1 Input of the initial data 2 Input of the deflection data 3 Backcalculation and 4 Viewing the results Each of these procedur
19. e sensitivity of each calculated surface de flection to variations in the pavement layer moduli and or a thickness A file containing deflections measured by a KUAB FWD Two soft ware products are available for producing deflection files using the KUAB FWD the KUAB software and the SFWD software MICHBACK supports both software products Acronym of a Swedish FWD manufacturing company Michigan Department of Transportation Plots of backcalculated layer moduli or percent variation of moduli from those backcalculated at the representative basin along a pave ment segment The acceptable percentage error between successive estimates of each layer modulus see Section 2 2 If a layer thickness is also being op 27 Newton method pavement segment PCC slab RMS tolerance representative basin rubblized pavement sensor weights seed moduli stabilized subbase stiff layer test location timized then the tolerance used for the layer thickness is the same as that used for the moduli A numerical method of solving non linear equations using a gradient matrix If the gradient matrix is not updated after every iteration then the method is referred to as the modified Newton method A length of pavement selected for backcalculation Portland cement concrete slab The root mean square percentage error between the measured and cal culated surface deflections that is acceptable see Section 2 2 The deflection basin at
20. electing Zoom In Out option 1 allows you to obtain a clearer view You are then prompted for a beginning and ending test location number for zooming in or out You may enter appropriate numbers for zooming or enter R at either prompt to re display the entire deflection profiles When viewing a deflection profile that is zoomed such that successive test locations are dis played at each major tick mark along the X axis each sensor reading is shown with a circle symbol on the profiles The circle symbol is not shown for deleted readings see Section 5 7 3 below 5 7 2 Identifying test locations for analysis The Identify Analysis Locations option option 2 is used to begin specifying test locations for backcalculation Typically backcalculations are performed over a given pavement segment You will be prompted for the beginning and ending location numbers for selecting the pavement segment Three options are then presented for selecting deflection basins used for backcalculation 1 Representative Basin This option selects one deflection basin from among those mea sured at the test locations along the segment as the most representative one Backcalcu lation is performed using the deflections measured at this single location The basin is selected using the criterion outlined in Appendix B 2 Individual Locations This option performs backcalculation at each of the test locations in the selected segment If the segment has several tes
21. ements into an ASCII file having the format described in this section MICHBACK can then read this ASCII file and plot deflection profiles enable segments to be se lected for backcalculation and allow data manipulations as described in Sections 5 7 1 5 7 2 and 5 7 3 The sample ASCII deflection data file shown in Fig 19 is supplied with MICHBACK and can be used as a template for preparing a data file The file has two parts the first part contains comments the second part contains the required data and the two parts are delimited by a separator line consisting of at least 20 signs The required data must be formatted as follows 1 An E or s must be typed as the first non blank character of the first line following the sep arator line to identify English or SI units Any comments entered after the first character are ignored The data in the file may be in either system of units and if this is different from the system of units being used in MICHBACK then the data is converted appropriately Def lection Basin at Location 8 Def lection Basin at Location 8 Def lection mils Def lection mils 28 38 4i 28 38 Radial Distance inches Radial Distance inches Press any key to continue Press any key to continue a Before Deletion a After Deletion Figure 18 Deflection basins for example before and after deletion at test location 8 18 61 MICHBACK V1 0 ASCII DEFLECTION DATA List of Abb
22. eneeeeteeeeeeeenees 9 Formi f r Pavement ty p sissn icenian g acct urate ea ea a iana 10 Extension to form for pavement type 2 sssseseseeseesersseseresressesererressrseresressessresressesee 10 Form for layer names thicknesses and Poisson s ratios sssseeeeesereererserersereseee 11 Form for sensor locations and weights eeeseseesssseessesesesresseseresressessresresseseresressesee 12 Possible extensions to Sensor Location Weight form 0 0 0 0 cee eeeeeseeeeneeeeeeeneeeees 12 Form for station identification and number of data Sets eee eeeeeeeeesseeesseeeeeeees 13 Form for specifying deflection data ccccdeates conieczeieal ceieticdenedweancslacaiiin 13 Deflection data file type mentsen a a eee 14 Typical plot of deflection data sseeeeeeeeeeeeseseseeeesseesseseresresseserssresstssresressessresressesee 15 Typical listing of drops for which the load differs by more than 10 trom the target lodd assin n a a E het A etd geet ad 16 Deflection profiles before and after deletion for example cee eeeeseesteeeeeeeees 17 Deflection basins for example before and after deletion at test location 8 18 Sample ASCII deflection data file with instructions sesessseesssseesseesseessersseeesseee 19 Form for seed moduli and moduli ranges eseeeseeeeeseseseeeesseseesressessrerrersessresressesee 20 Prompt to request whether stiff layer should be incorporated in the analysis 21 Form fo
23. entries in each field and save data If some of the data is invalid prompts will be issued for corrections Esc Discards any changes made on the current screen and return to previous screen 5 2 Cursor Movement and Editing Keys The data entry forms have several fields into which data is typed The field in which the cur sor resides is highlighted on IBM compatible systems The functions of the cursor movement and editing keys within a form are described in Table 1 5 3 Title Screen When MICHBACK is loaded the title screen shown in Fig 1 is displayed Pressing the F1 key displays the credits screen shown in Fig 2 while pressing any other key displays the main menu 5 4 Main Menu The main menu is shown in Fig 3 Any one of the eight options shown on the menu may be selected by typing a number from 0 to 7 These options are described below e Option 1 Used to input data relating to a new backcalculation problem e Option 2 Used to change data for the problem currently being worked e Option 3 Used to read the data from a previously defined problem and modify it if neces sary The name of the file in which the previous data was saved will be requested e Option 4 Used to enter deflection data manually using the keyboard e Option 5 Used to read a KUAB FWD file or an appropriately formatted ASCII file con taining measured deflections from a series of tests Option 6 Prompts for run time data and begins the backcalculatio
24. er the deletion are shown in Fig 18 In MICHBACK deflection basins are drawn using cubic splines Note that the circle symbol is not shown for the deleted reading For this example the deflection basin after the deletion at test location 8 seems more reasonable than that before the deletion 17 5 7 4 Manually editing deflection basins After selecting test locations for analysis you may manually view edit the sensor readings by returning to the Main Menu and selecting Enter Deflection Data using Keyboard option 4 If the representative basin or average basin was selected for backcalculation then only that single ba sin may be viewed edited If individual locations were selected for backcalculation then the first four basins may be viewed edited Basins that are viewed edited in this manner are saved in the data file and may be recovered at a later time If you wish to perform backcalculation using the edited basins then you must select Perform Backcalculation option 6 from the Main Menu 5 7 5 Backcalculation After identifying the analysis locations the Backcalculate option option 4 is used to initiate the backcalculation This option is similar to the Perform Backcalculation option option 6 in the Main Menu Please see Section 5 8 for details on this option 5 7 6 Format of user created ASCII deflection data file If you use an FWD device that is different from the KUAB FWD then you will need to trans fer your deflection measur
25. ered deflection basins may be plotted one after another or overlaid on the same plot The visual display is useful for finding gross errors in the in put data and for examining the general shape of the basins Deflection basins entered manually are saved in the data file and may be recovered at a later time 5 7 Reading Deflection Data from a File Deflection data for up to 400 tests may be read from a file generated by the KUAB FWD de vice operated by MDOT or from a specially formatted ASCII file by selecting Read Deflection Data from a File option 5 from the Main Menu The deflection data may be viewed graphically and manipulated prior to backcalculation MICHBACK computes various useful statistics of the backcalculated results When this option is selected the menu shown in Fig 14 is displayed Options 1 is used to read a file created by the KUAB software option 2 is used to read a file created by the newer SFWD software or option 3 is used to read a file having the format specified in Section 5 7 6 For all three options you will be prompted for the name of the deflection data file Enter the name of the deflection data file or press RETURN to return to the Main Menu For a valid deflection data file the data is read unit conversions are performed if the data in the file is not in the same system of units as that being used in MICHBACK and the deflection profiles are displayed graphically in a plot similar to that shown in Fig 15
26. es is described in this manual Most of MICHBACK is written in FORTRAN 77 Graphics and screen manipulations are performed using the FORTRAN callable GRAFMATIC graphics library marketed by Microcom patibles Inc 301 Prelude Drive Silver Spring MD 20901 2 Summary of Modeling and Backcalculation This section gives a summary of the modeling and backcalculation so that you are aware of the capabilities and limitations of the MICHBACK program Further details about the modeling and backcalculation and various sensitivity studies are given in the publications by Harichandran et al 1993 and 1994 and Mahmood 1993 2 1 Modeling of the Pavement and Forward Calculations Each layer in a pavement cross section is assumed to extend infinitely in the horizontal direc tions and the last layer which may possibly be a stiff layer such as bedrock is assumed to be infinitely deep All the pavement layers are assumed to be fully bonded so that no slip occurs due to applied load The load applied by the FWD or other device is assumed to be a uniform pressure over a circular area Due to the assumptions used the problem is reduced to an axisymmetric one The pavement layers are assumed to be composed of linear elastic material An extended precision CHEVRON program is used for the forward calculations 2 2 Backcalculation and Convergence Control MICHBACK uses a modified Newton method for the backcalculation Several parameters control the backca
27. esaes 4 AA Printige Graphies s e occ Sexi a Rove a A E mecca EEN 4 5 Use MICHBACK cccsccccctecscscececocsesszesccnteces einai a staats sadete ossut adod sa eee 4 DL VES NAMES So aana E E e T A canes ai laa Ns 4 5 2 Cursor Movement and Editing Keys ceecccssscecsenceceeececesececesccecssececssceeeseeeeesaeers 5 5 3 Title Screen issiniteh ieee esi aces ela eld eels 5 5S4 Man Men n a a a e E E E A E 5 5 5 Data File Menus and Associated Forms ccccsccccesscecescecesececssccecssccecsccecsseeeenscees 7 5 5 1 Specifying system of units ssssssessseeeseeeesseessessessseeesseeesseesseesseesseeesseee 8 5 5 2 Specifying initial data and convergence control ssseesseessessseesseeeesee 8 5 5 3 Specifying pavement type ssssssssssesssesessseessressrtsseesseresseeesseesseesseessee 10 5 5 4 Specifying layer names thicknesses and Poisson s ratiOS eeeeee 10 5 5 5 Specifying sensor locations and weights sssessesssesssesessseessresssesseessee 11 5 5 6 Returning to the main Menu ssssssesseiesrseeesresseserestesseseresresseseresressesees 13 5 6 Entering Deflection Data using the Keyboard essseeseseseeeeesersreseresersrrseresressrseresee 13 5 7 Reading Deflection Data from a File cs cccas sp scandgsscagetsagasdeaaiqeieaieeceantaneteiclenstssaises 14 STI Zooming es eee eni erene n EE TE REEE E 16 5 7 2 Identifying test locations for analysis ssessssesseessesseeesssesss
28. essresseessee 16 5 1 3 Deleting readings eane a E R AE E NSE 17 5 7 4 Manually editing deflection basins 5J sesiedapqeseth aliedaeareatiddeaeades 18 5 1 Backealc lati on i etann erens e p E a Eni ESEE 18 5 7 6 Format of user created ASCII deflection data file eeeeeeeeeeeeeteees 18 5 8 Performing Backealeulaton c s ccchiassdigh eiaiaeadvs cared inde eed aoe 20 DOPE WAC WOR SUM IVS EI sc gies cca ps a Seas essa ee Suan roew O NE 22 Problem Reporting sssssicsissssssissssssississssssssssosesssssssissiseisesossssssosissit sesiis soseen 25 Credits and Acknowledgments seesssooossesssoocecessssocecessosooceesssooecessssoooeessssssseeesso 25 eN J D IE LD lt PT E E 25 EAKA A TE S A OE E 26 eE Ta E A E E E E ETT 27 iii Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 Figure 18 Figure 19 Figure 20 Figure 21 Figure 22 Figure 23 Figure 24 Figure 25 Figure 26 Figure 27 Table 1 List of Figures IEC SCTE CMs ce ctcstes scaler s E R actors arcudites stoaden SES 6 Credits sereen oiri etet eet pulse E A E lene et A E E 6 M in ment moa n er are e a NS tl tae has eaan a aaa e tat 7 Wer data file menire isep reen a E E as E T Ea des SA ETEA N 7 Formi GORMAN a a deseen an E Gs haat T A AO cametenn sd 8 Form for changing convergence control parameters 00 0 0 csceeeeeeeesree
29. he backcalculated layer moduli STIFF LAYER PROPERTIES 1 Modulus 2000000 0 psi 2 Poisson s Ratio 45 3 Estimated Depth to Stiff Layer From Top Surface 600 0 inches 4 Allow Program to Vary Above Depth y n y Figure 22 Form for changing stiff layer properties 21 RMS Estimated Moduli psi Stf Layer Calculation Error Base Roadbed Depth in Time sec 500000 432148 795023 454054 526295 528461 525222 519612 531109 0 L 2 3 4 5 6 7 8 m HNR G O Hj Backcalculating at station 1 out of 18 stations Test Location 1 Total Calculation Time at End of Previous Iteration 0 min 34 sec Figure 23 Run time summary screen During the backcalculation process run time summary information similar to that shown in Fig 23 is displayed on the screen Estimates of the layer properties during each iteration and the calculation time for each iteration are shown in tabular form The test location at which the back calculation is being performed and the total calculation time at the current test location are dis played below the table The table is cleared and rewritten when backcalculation is begun at the next test location The calculation times shown in Fig 23 are for a 90 MHz Pentium PC 5 9 View Results Menu The View Results menu shown in Fig 24 is displayed when option 7 is selected from the Main Menu The file with the same base name as the ou
30. ient memory Unloading any TSR s i e terminate and stay resident programs before running MICHBACK may alleviate the problem Percent Locations with AC Modulus Less Than or Equal to M Percent Locations with Peak Deflection Less Than or Equal to D CDF Percent CDF Percent 108 108 a 1 4 4 1 Pi 1 1 a 1 L 1 f 1 268 388 408 588 688 768 888 988 1600 5 5 6 8 6 5 7 0 7 5 8 8 8 5 9 8 9 5 18 8 AC Modulus M ksi Def lection D mils Press any key to continue Press any key to continue a CDF of backcalculated AC modulus b CDF of peak deflection Figure 27 Cumulative distribution functions for AC modulus and peak deflection 24 Problem Reporting Although MICHBACK has been tested quite extensively it is possible that errors causing the program to terminate abnormally may still be encountered if a haphazard sequence of options is used To report a problem note down the number and message displayed when the program termi nates abnormally and send it along with a diskette containing the input data file and deflection file to Dr Ronald S Harichandran or Dr Gilbert Y Baladi Department of Civil amp Environmental Engineering Michigan State University East Lansing MI 48824 1226 Alternatively send the input data file deflection file and a description of the error by electronic mail to harichan cee msu edu or baladi cee msu edu Credits and Acknowledgments The MICHBACK program reflects the combined
31. lculation procedure and the convergence checks Default values for these param eters are automatically supplied in MICHBACK but you can change these if desired For unusual or difficult problems control parameters that are different from the default values may yield better results In the modified Newton method a gradient matrix is used to determine how to change the current estimates of the layer moduli and perhaps a layer thickness Calculation of this gradient matrix is time consuming since it requires several calls to the CHEVRON program Therefore rather than calculating the gradient matrix after every iteration by default it is usually updated only after three iterations Although this procedure slows down convergence overall efficiency is at tained If the program appears to be converging to an incorrect solution for a particular problem then calculating the gradient matrix after every iteration may yield better results Two criteria are used to check for convergence Convergence is assumed to have occurred if one of the following conditions are met 1 The root mean square RMS percentage error between the measured and calculated sur face deflections is smaller than a specified tolerance This criterion is expressed as JE where w is the deflection measured by sensor j and w is the deflection calculated at the location of the jth sensor using the current estimates of the moduli and perhaps a layer thickness The default
32. moduli and moduli ranges Default values are generated by MICHBACK but you may change these The default val ues are generated using the representative basin the first basin in the identified segment or the av erage basin depending on whether the backcalculation is being performed using the representative basin at each individual location or using the average basin respectively The seed moduli pro vide an initial starting point for the backcalculation and the moduli ranges constrain the backcal culated results For well behaved deflection basins the backcalculated results are not sensitive to the seed moduli and very wide moduli ranges may also be specified However the convergence SEED MODULI psi Layer Layer Seed Minimum Maximum No Name Modulus Modulus Modulus 1 AC 500000 0 50000 0 3000000 0 2 Base 10000 0 300 0 500000 0 3 Roadbed 7500 0 3000 0 100000 0 Figure 20 Form for seed moduli and moduli ranges 20 STIFF LAYER DEPTH 1 Program estimates the stiff layer depth to be 192 1 inches 2 Incorporate stiff layer y n Figure 21 Prompt to request whether stiff layer should be incorporated in the analysis will occur sooner if good seed moduli are provided For ill defined deflection basins e g from badly damaged composite pavements the final backcalculated results may be dependent on the seed moduli If backcalculation is being performed at several individual locations
33. n process Copyright C Fl to list credits Backcalulation of Flexible and Composite Pavement Layer Properties Using a Modified Newton Method Developed for Michigan Department of Transportation Dept of Civil amp Environmental Engineering Michigan State University East Lansing MI 48824 1226 For further information call Version 1 0 MICHBACK by 517 355 5107 1995 MSU Board of Trustees Press any key to start Figure 1 Title screen Contributions Sponsers Michigan Department of University of Michigan Project Directors MICHBACK Version 1 0 June 1995 Conceptual Development and Overall Guidance Ronald S Harichandran Tariq Mahmood and Gilbert Y Baladi Cynthia M Ramon and Weijun Wang Programming A Robert Raab Collaboration Lynne Irwin Corrected CHEVRON Program Transportation Transportation Research Institute Gilbert Y Baladi Prinicipal Investigator Ronald S Harichandran Co Principal Investigator Press any key to continue Figure 2 Credits screen MAIN MENU Create a New Data File Change Current Data File odify an Existing Data File Enter Deflection Data using Keyboard Read Deflection Data from a File Perform Backcalculation View Results Exit OANDOABWNE Selection _ Figure 3 Main menu e Option 7 Used to view the backcalculated results e Option 0 Exits from the MICHBACK program and ret
34. neously 5 Using MICHBACK MICHBACK is designed to be user friendly Menus are used to perform the required steps in the backcalculation and data entry forms facilitate data input In addition extensive checking of input data is performed and appropriate error messages are displayed upon completion of each form 5 1 File Names The names of files in which the data and results are saved may include an absolute or relative path name if necessary e g a sample dat to save the file sample dat on the diskette in drive A job1l sample dat to save the file in subdirectory job1 job1 sample dat to save the file in the subdirectory job1 relative to the default directory etc If no path is specified the file will be saved in the default directory in which MICHBACK was started TABLE 1 KEYPAD FUNCTIONS WITHIN DATA ENTRY FORMS KEY Function Return Enter Move cursor to next field Tab Move cursor to next field on the right Shift Tab Move cursor to previous field on the left Home Move cursor to first field in the form End Move cursor to last field in the form T or PgUp Move cursor to field above current one J or PgDn Move cursor to field below current one Backspace Delete character before cursor Del Delete character at cursor Ins Insert space at cursor gt Move cursor one space to the right Move cursor one space to the left Fl Check validity of
35. ng the following commonly used version of the Asphalt Institute equation logE logE 1 47362x10 12 12 where Eo corrected AC modulus in psi E backcalculated uncorrected AC modulus in psi t test temperature in degrees Fahrenheit and to reference temperature 77 F 25 The user is cautioned that deflection data in general does not support the Asphalt Institute temperature correction equation A proper temperature correction should be based on the temper ature gradient and the average temperature in the asphalt concrete layer B Representative Deflection Basin MICHBACK allows the backcalculation to be performed at a basin that is most representa tive among a series of basins measured along a pavement segment The percent variations in the backcalculated moduli plotted in the graphical display are also calculated relative to the moduli backcalculated using the representative basin The representative basin is selected based on the deviation from the average of three mea sures a the peak deflection under the load b the deflections measured by all sensors and c the area under the deflection basin Alexander et al 1989 A measure of the deviation at any test location is IsM iSM 2_ G w wY aA a HE L J 4 J ISM FN W A where ISM impact stiffness modulus applied load deflection under the load w deflection at sensor j m no of sensors A area under the deflection basin and
36. ocations erroneous measurements and for selecting test locations or segments for analysis The deflections stored in the file must all be measured for the same target load and the num ber of drops at each test location must be identical If the FWD deflection data file violates these requirements then an error message is displayed Error messages are also issued if there are unex pected entries in the data file in which case you may need to edit the file and delete inappropriate 15 Test location s at which loads are outside 10 of target load of 9000 lbs are listed below Location Location Drop Difference from Number Distance Number Target Load 3 20 2 36 6 4 40 1 STel 4 40 2 36a 4 40 3 36 5 Finished reading data file Press any key to continue Figure 16 Typical listing of drops for which the load differs by more than 10 from the target load lines before trying to read the file again Additional tasks are performed using the menu displayed below the plot Occasionally an FWD file may contain drop loads that differ significantly from the target load and in this case drops for which the load differs by more than 10 from the target load are listed on the screen when the file is read Fig 16 shows a typical listing 5 7 1 Zooming If deflections are measured at a very large number of test locations then the initial deflection profiles shown after the file is read may be highly compressed and difficult to interpret S
37. r changing stiff layer properties s ssssssesssesesseessseessresseesseeesseressresseesseessee 21 RUG summary s ri sssricsretiietn eiit ii eiis 22 View results Menus 24 gentniien tiated dati bai aa EE aE ESAE ESS iA aiis 22 Summary results from backcalculation seeeeseseseeseeseseresressessresrersersresrreseesreseesee 23 Typical profiles of the backcalculated moduli sseeseeeseeeeeseeesesressessresressessrsrresseese 24 Cumulative distribution functions for AC modulus and peak deflection 24 List of Tables Keypad functions within data entry forms 0 eee eee eeeeceseeeseeeeseecaeceaeesseeesaeeenaeens 5 iv 1 Introduction MICHBACK is a menu driven program which uses a modified Newton method for the back calculation of flexible and composite pavement layer moduli thicknesses and stiff layer depth from measured surface deflections The backcalculation procedure is an iterative one requiring several forward calculations MICHBACK Version 1 0 has the capability of automatically processing falling weight de flectometer FWD measurement files created by the KUAB FWD device used by the Michigan Department of Transportation MDOT For other measuring devices the data files should be translated into an ASCII file having the format specified in Section 5 7 6 before processing The advantage of using data files containing several sets of measurements is that MICHBACK allows you to visually inspect the measure
38. rating system For DOS systems the following hardware and software are required e MS DOS version 3 0 or higher e 640 KB of random access memory RAM e A hard disk e A color graphics adapter CGA EGA or VGA and compatible monitor Although not strictly required for the use of the program the following hardware is strongly recommended e A 386 or later PC e A math co processor 8087 80287 or 80387 Running time will be greatly increased if a math co processor is not installed e A printer for obtaining hard copies of plots and output 4 Configuring the Computer 4 1 Installation Procedure The MICHBACK program is initially supplied on one high density diskette To install the program on a hard disk first make a subdirectory to hold the program e g MD MBACK and copy all the files in the supplied diskette to the hard disk 4 2 The config sys File In the root directory there is a file named config sys which configures the PC system and loads any requested device drivers when the computer is turned on The following statement will need to be added to the config sys file if it does not already exist FILES 20 The MICHPAVE program uses a Fortran callable graphics package called GRAFMATIC Unfortunately this package is not compatible with the ansi sys device driver used by some other programs for screen manipulations Thus if the config sys file has the statement DEVICE ANSI SYS
39. reviations The data before the separator line below is optional The separator SL Station location distance from a fixed reference point line consisting of 80 signs must appear in each file followed by TEMP Pavement temperature the required data formatted according to the instructions given below LOAD Actual load applied during FWD drop Dn Deflection at sensor n n 1 2 10 Instructions Se ee ae Optional Information 1 Make a backup copy of this file for a permanent record 2 2 2 2 2 eanan h ieee ee Project Number 2 Arrange your deflection data in the column format shown at the bottom Test Location of this file The data following the separator line are as follows Additional Comments a An E or S to identify English or SI units for all data in this file must be typed in the first column of the first line If the Replace the example data below with your data system of units specified in this file is different from that specified in MICHBACK then the deflection data in this file will be converted to correspond to the units being used in MICHBACK E Type E for English or S for SI units b The number of drops per test must be typed at the beginning of the 3 lt Number of FWD drops per test second line 9000 lt Target load c The target load must be typed at the beginning of the third line All deflections are scaled linearly to correspond to the target SL TEMP LOAD D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 load prior
40. s deletions FWD gradient matrix KUAB FWD file KUAB MDOT modulus profiles modulus tolerance Asphalt concrete A file containing only characters in the American Standard Code for Information Interchange An ASCII file can contain only standard keyboard characters and a few non printing characters carriage re turn new line etc The process of estimating layer properties from measured surface de flections Cumulative distribution function The CDF F x is equal to the pro portion of all observations for which X lt x Thus if M denotes the AC modulus then F y 100 000 psi is the proportion of all test locations in the analyzed segment for which M lt 100 000 psi An elastic layer program developed by the Chevron Oil Company The numerical quadrature used in the original CEHVRON program is not accurate enough especially for stiff pavements and an extended precision version is used in MICHBACK A pavement containing a concrete slab as the first uppermost or sec ond layer The deflected shape of the surface measured by a FWD Plots of sensor deflections along a length of pavement Measurements at particular test locations and particular sensor loca tions that are deleted by the user Deleted measurements are assigned a zero weight during backcalculation Test locations with deletions are not considered when determining the representative basin Falling weight deflectometer A matrix that indicates th
41. t locations then the time required for the backcalculations may be considerable A sub menu is presented for selecting how to view the deflection basins You may omit the viewing of basins view only the first basin or view the basin at each test location in the selected segment 16 Route EXAMPLE Route EXAMPLE Deflection mils Deflection mils 8 8 SEN We ODN SrEN Wem DN 46 86 26 68 46 88 28 68 6 1 fi L L L fi 1 ri L ri 6 8 18 12 14 16 18 6 8 18 12 14 16 18 Test Location Number Test Location Number 1 Zoom In Out 2 Identify Analysis Locations 3 Delete Readings 1 Zoom In Out 2 Identify Analysis Locations 3 Delete Readings 4 Backcalculate Return to Main Menu 4 Backcalculate Return to Main Menu Selection _ Selection _ a Before Deletion b After Deletion Figure 17 Deflection profiles before and after deletion for example 3 Average Basin This option computes an average deflection basin by averaging the deflec tions measured by each sensor over all test locations in the selected pavement segment A single backcalculation is then performed using the average basin 5 7 3 Deleting readings Selecting Delete Readings option 3 allows you to delete the readings of one or more sensors at one or more test locations This option may be useful if one or two sensors seem to give erratic measurements at some locations and you suspect that they are in error Yo
42. the over bar denotes the average of the corresponding quantity over all test locations in the se lected pavement segment All deflection basins are linearly scaled to correspond to the same target load when calculating A The scaled deflection basin having the smallest value of A is selected as the representative one References Alexander D R Kohn S D and Grogan W P 1989 Nondestructive resting techniques and evaluation procedures for airfield pavements In Nondestructive Testing of Pavements and Backcalculation of Moduli Bush A J and Baladi G Y eds American Society for Test ing and Materials Philadelphia 502 524 Harichandran R S Mahmood T Raab A and Baladi G Y 1993 A modified Newton al gorithm for backcalculation of pavement layer properties Transportation Research Record 1384 15 22 Harichandran R S Mahmood T Raab A and Baladi G Y 1994 Backcalculation of pave ment layer moduli thicknesses and bedrock depth using a modified Newton method In Nondestructive Testing of Pavements and Backcalculation of Moduli volume 2 American Society for Testing and Materials Philadelphia PA Mahmood T 1993 Backcalculation of pavement layer properties from deflection data Ph D Thesis Michigan State University East Lansing 26 Glossary AC ASCII file backcalculation CDF CHEVRON program composite pavement deflection basin deflection profile
43. tion but the number of drops must be the same for all test locations and all the drops must be for the same target load i e the drop height must not be changed 3 The target load must be typed at the beginning of the third line Any comments entered after the load are ignored Each measured deflection is linearly scaled to correspond to the target load 4 The fourth fifth and sixth lines are ignored and may contain optional headers identifying the columns of data 5 The deflection data is entered in columns separated by spaces and or tabs Up to 10 sensors may be used and in addition to the sensor readings the station location pavement temper ature and actual drop load must be entered Deflections for each test are grouped together The station location and pavement temperature must be entered on the first line of the group along with the deflections measured for the first drop Deflections for subsequent drops must be entered on consecutive lines Deflections for different test locations must be sepa rated by one blank line 5 8 Performing Backcalculation Backcalculation is initiated by selecting Perform Backcalculation option 6 from the Main Menu or by selecting the Backcalculate option in the deflection profile display when a deflection data file is used Several forms are presented so that you may enter additional data required for the backcalculation A form similar to that shown in Fig 20 is displayed for specifying seed
44. to backcalculation d The fourth fifth and sixth lines are ignored 0 532 9663 19 37 10 56 10 31 6 93 4 96 2 87 1 63 e The deflection data begins at the eight line and n 3 columns are 9655 18 95 13 78 10 25 6 89 4 98 2 88 1 62 required where n number of sensors 9648 18 80 13 75 10 21 6 91 4 94 2 87 1 62 f Columns may be separated by spaces or tabs g The number of drops at each station must be the same 100 52 5 9769 10 22 7 84 6 25 4 76 3 78 2 52 1 48 h Deflections for each test are grouped together The station 9760 9 98 7 69 6 12 4 73 3 79 2 52 1 47 location and pavement temperature must be entered on the first line of 9736 9 88 7 65 6 09 4 69 3 75 2 51 1 50 the group along with the deflections measured for the first drop Deflections for subsequent drops must be entered on consecutive 200 56 2 9672 10 40 7 87 6 36 4 81 3 84 2 58 1 49 lines Deflections for different test locations must be separated 9671 10 17 7 74 6 29 4 78 3 83 2 54 1 50 by a blank line 9648 10 06 7 68 6 26 4 74 3 83 2 53 1 50 3 Cross section data number of sensors and other pertinent information must be input directly in MICHBACK Units English SI Deflections mils microns Temperature Fahrenheit Celsius Load pounds kN Figure 19 Sample ASCII deflection data file with instructions 2 The number of FWD drops per test must be typed at the beginning of the second line Any comments entered after the load are ignored Up to 5 FWD drops may be recorded at each test loca
45. tput file but with the extension plt must exist in order to view plots This plot file is created during backcalculation The options on the View Results menu are explained below Option 1 Used to display the summary results on the screen Typical results are shown in Fig 25 The average standard deviation maximum and minimum moduli for each layer based on the analysis of all identified test locations are displayed If the stiff layer depth or a layer thickness is also optimized then statistics for this are also displayed Option 2 Used to type the output file on the screen The file is displayed using the DOS command MORE The display pauses after each page and you may display the next page by pressing any key You can terminate the display without paging through to the end of the file by pressing CTRL BREAK at any time The output file can always be viewed from outside MICHBACK using any text editor or pager Option 3 Used to send the output file to the printer when it exists The output file is spooled to the printer through the DOS command PRINT If the printer port was not set since the computer was booted up you will be prompted for the port You may respond by pressing RETURN if the printer is connected to LPT1 also called PRN or typing LPTn or 22 VI EW R ESULTS Plot 1 Type 2 Type 3 Send 4 View 5 0 Summary Results on Screen Output File on Screen Output File to Printer Results
46. u may then wish to delete suspicious measurements prior to backcalculation in order to obtain a more uniform deflection pro file During backcalculation MICHBACK assigns a zero weight to deleted measurements Note that the original deflection data file is not changed in any way Test locations with one or more deletions are not considered when determining the representative basin for a pavement segment Deleted sensors are not considered when the average basin is computed for a segment Up to 50 deletions may be made You are prompted for a test location number and sensor number for each deletion You may type a at the prompt for the sensor number to delete all sen sors at the specified location Note that sensors are numbered starting with 0 To quit without sav oe 99 ing the deletions in memory type go at either prompt To keep the deletions and replot the 66 599 modified profiles type x at either prompt Fig 17 a shows an example deflection profile in which the deflections measured by the sec ond sensor at test location 8 appears to be high and that measured by the 5th sensor at test location 15 appears to be low Prior to performing backcalculation one may wish to explore the effect of deleting the readings of the second sensor at test location 8 and the fifth sensor at test location 15 The deflection profiles after the two deletions are shown in Fig 17 b The deflection basins at test location 8 before and aft
47. urns to DOS 5 5 Data File Menus and Associated Forms Data File Modify Current Data File and Modify Existing Data File menus are displayed when selecting options 1 2 or 3 respectively from the main menu All three menus are identical in structure and the first is shown in Fig 4 The only difference between these menus is that in those used for modifying data files existing data is modified instead of specifying new data It is recommended that the options in the data file menu be followed in sequence DATA FILE zj ES ENU 1 Specify System of Units Current System is English 2 Initial Data Input Output Filename Tolerances etc 3 Specify Pavement Type 4 Specify Material Properties 5 Specify Sensor Location 0 Return to Main Menu Selection _ Figure 4 New data file menu When modifying an existing file it is mandatory to first use option 2 and specify new names for the required file names or to indicate that the input and output files used earlier should be over written The other options may be performed in any sequence 5 5 1 Specifying system of units Selecting Specify System of Units option 1 in the Data File Menu allows you to change the system of units in which the data is entered displayed and stored The system of units that are cur rently active are shown within parentheses in item 1 of the Data File Menu Whenever the system of units is changed all data is automatically
48. y key to continue Figure 25 Summary results from backcalculation 23 Percent Variation From Modulus at Representative Location 15 Backcalculated Results for Layer 1 Percent Variation of odul i Modulus ksi 158 18088 988 168 p 888 58 F 708 688 7 saa 50 408 388 100 1 1 L 1 L f 200 1 1 1 f f 1 1 1 6 2 4 6 8 16 12 14 16 18 6 2 4 6 8 16 12 14 16 18 Test Location Test Location Number AC S Roadbed iaa Base 1 Zoom In Out Exit 1 Zoom In Out Exit Selection _ Selection _ a Percent variation of all layer moduli b Variation of AC moduli Figure 26 Typical profiles of the backcalculated moduli computed only at the points indicated in the plot and these are connected by straight lines CDF s can only be plotted if backcalculations are performed at 10 or more locations For any modulus value M along the X axis the Y ordinate of the CDF of the AC modulus shows the percentage of basins having backcalculated AC moduli smaller than or equal to M The CDF of the peak deflections is interpreted similarly The peak deflections are the values read from the deflection data file and are input quantities rather than output quanti ties The CDF plots indicate the proportion of test locations at which the pavement is weak and can help in deciding whether to overlay the analyzed pavement segment If options 2 or 3 do not function then the most likely cause is the lack of suffic
Download Pdf Manuals
Related Search