Enventive Monte Carlo Analysis Tutorial Version 4.1.x

Transcription

1 Enventive Monte Carlo Analysis Tutorial Version 4.1.x Copyright 2017 Enventive, Inc. All rights reserved.

2 Table of Contents Welcome to the Monte Carlo Analysis Tutorial 1 Getting started 1 Complete the assembly 2 Set up and run a Monte Carlo simulation when the wheel is on the line 4 Run a Monte Carlo simulation when the wheel is close to the arc 7

3 Welcome to the Monte Carlo Analysis Tutorial Welcome to the Monte Carlo Analysis Tutorial In this tutorial, you will learn how to set up and run a Monte Carlo analysis. Monte Carlo uses statistical random sampling to calculate values for an analyzed dimension, including mean, sigma, accuracy, largest/smallest value, and percent in tolerance. Enventive s standard tolerance analysis uses a Root Sum of Squares (RSS) calculation, which is appropriate for most needs, but Monte Carlo analysis may be preferable in cases such as: Nonlinear sensitivity values Changing contact locations (different contributor list) Non-Gaussian contributors Transition fit for pin-in-hole assembly For more information about use cases for Monte Carlo, see Enventive Online Help. To learn about Enventive's Monte Carlo analysis, we'll be studying the angle distribution of a moving bracket part. We will be working with an existing assembly that has been constructed with three model instances: a moving bracket, a fixed bracket, and a wheel. Getting started Before beginning the tutorial, download the following files from the Enventive website by going to (you must first log in as an Enventive customer to access the files) or from the Enventive installation directory under..\doc\monte_ Carlo. Move_bracket Fixed_bracket Wheel Assy_MC_2 Monte Carlo - 1

4 Enventive Monte Carlo Analysis Tutorial Version 4.1 Complete the assembly First, we ll complete the assembly by defining the orientation of the "Move_bracket" model instance. 1. Open Assy_MC_2. 2. Zoom into the slot in the Move_bracket instance. 3. Create a construction circle inside the slot and change its color to yellow. 4. Constrain the circle tangent to the sides of the slot hole. Monte Carlo - 2

5 Complete the assembly 5. Place a Floating Pin in Hole (FPIH) constraint between the yellow construction circle and the blue circle on the plate. 6. Select the FPIH constraint and note that the Automatic tolerance associated with the FPIH constraint is currently 0.028, which is not correct. The FPIH constraint automatically calculates its tolerance value based on the tolerances applied to the two circles. If a diameter dimension is not present and/or the appropriate tolerance value is not assigned, then the tolerance value for the FPIH constraint cannot be computed correctly. In this case, the blue circle does have a diameter dimension and tolerance assigned to it ( /-0.028) but the yellow construction circle does not. 7. Apply a diameter dimension to the yellow circle. (Use the right-click selection menu to select the correct circle, which is at the assembly level. The other (blue) circle is under the Fixed_bracket instance.) The diameter should be To determine the tolerance on the width of the slot, generate a tolerance analysis report on the derived diameter dimension. You will see that the nominal slot width has been assigned a tolerance of / Assign this same tolerance value to the derived diameter dimension. 9. Select the FPIH constraint and verify that the tolerance value is now Monte Carlo - 3

6 Enventive Monte Carlo Analysis Tutorial Version 4.1 Set up and run a Monte Carlo simulation when the wheel is on the line In this section, we ll set up and run a Monte Carlo simulation when the wheel is not at the intersection of the line and the radius of the polycurve. 1. Zoom All to see the entire assembly. Change the 29.0 CTRL dimension to Run tolerance analysis on the derived ANGLE dimension. The tolerance analysis report shows that there are eight contributors, the sigma value is , and 95.47% of the parts are within tolerance specifications. 3. Select ANGLE, and from the Analysis menu, choose Setup Monte Carlo Simulation. The Modules area opens the Monte Carlo Simulation tool, which automatically identifies the contributors for ANGLE. Note: The Monte Carlo simulation shows 12 contributors rather than the 8 reported in standard tolerance analysis, because Monte Carlo analysis also includes the four diameter dimensions associated with the two Pin-In-Hole contributors. Monte Carlo - 4

7 Set up and run a Monte Carlo simulation when the wheel is on the line 4. The Stop After settings in the Run area of the Monte Carlo simulation lets you specify the accuracy or number of trials (iterations) that must be reached to complete the simulation. Select the # Trials option, and keep the default setting of 20,000 trials. With this setting, the simulation will stop after 20,000 iterations. Note: The Accuracy value specifies the desired range for the 95% confidence interval based on a Student s t-test, where the units for accuracy are the same as the units of the analyzed parameter. For example, if we specify an Accuracy value of for the Stop After setting, the mean value will be when the simulation is complete. We can be 95% confident that the actual mean value is within the range of and Run the simulation by pressing the Play button. The sketch will be updated every 100 trials by default, as specified in Update Display Every: field. 6. When the simulation is finished, compare the results with those of the standard Enventive RSS tolerance analysis you did in step 2. (You can view the tolerance analysis report by clicking the Excel button in the Modules area, and switch back to the Monte Carlo Simulation by clicking the Monte Carlo button.) RSS Results Monte Carlo Results (Typical Values) Mean Sigma % in Tolerance 95.47% 95.1% % Although slightly different, both the Monte Carlo and RSS methods give a mean value around 0.000, a sigma value around 0.250, and a % in Tolerance value close to 95.4%. The two methods are roughly equivalent in this case. Note: Your values will differ from the one shown in our illustration, since Monte Carlo is based on random trials, and will never repeat exactly the same results. Monte Carlo - 5

8 Enventive Monte Carlo Analysis Tutorial Version 4.1 Monte Carlo - 6

9 Run a Monte Carlo simulation when the wheel is close to the arc Run a Monte Carlo simulation when the wheel is close to the arc In this section, we ll rerun the Monte Carlo simulation after setting the CTRL dimension to 32, where the wheel is close to the arc. 1. Change the CTRL dimension to 32.0 with a tolerance of ± Perturb the CTRL dimension from 32.0 to 28.0 (if needed, adjust the step size to 1 or 2), and observe that the wheel transitions from the flat section of the fixed bracket onto the arc of the polycurve. This means that the configuration is subject to changing contact locations, and consequently, changing contributors. However, Enventive's RSS analysis is based only on the configuration at the nominal value. It does not take into consideration the dimensions and constraints associated with other geometric objects that could be involved if contact conditions change. Those extra contributors will affect the results and must be added in a Monte Carlo analysis. 3. Change the CTRL dimension value from 32 to 28.0, and update the tolerance analysis report for the ANGLE dimension by right-clicking in any cell in the Excel spreadsheet and selecting Update Current Analysis from the Enventive menu. Monte Carlo - 7

10 Enventive Monte Carlo Analysis Tutorial Version 4.1 Changing the contact status has resulted in 14 contributors instead of only 8, as were reported in the RSS tolerance analysis report, when CTRL was set to (The additional 7 contributors in red are "construction" contributors, which are orientation constraints that do not have tolerances and do not affect the results. See Enventive Online Help for more information.) Note that the Sensitivity value for CTRL is In the tolerance analysis report, change the nominal value for CTRL to Note that the Sensitivity value has changed to 0.00 for CTRL and five other contributors Dim15, Dim3, Dim17, Dim12, and Dim16. These are the six new contributors that become relevant when the contact condition changes as the wheel moves onto the arc. The sensitivity value for many of the other contributors also changed. Monte Carlo - 8

11 Run a Monte Carlo simulation when the wheel is close to the arc Note that the mean value is 0.0, the sigma value is , and the % in Tolerance value is %. 5. With the CTRL value still set to 32.0, select ANGLE, and setup the Monte Carlo Simulations. Note that twelve contributors are automatically added, but the contributor list does not include CTRL, Dim15, Dim3, Dim17, Dim12, and Dim Change the nominal value for CTRL to 28.0 and click Auto Add Contributors from the Setup menu. We now have 18 contributors, including CTRL, Dim15, Dim3, Dim17, Dim12, and Dim Change the nominal value of CTRL back to Note that the contributor list still indicates that the Monte Carlo simulation will use 28.0 as the mean value for CTRL. 8. Click on the Refresh Monte Carlo - 9

12 Enventive Monte Carlo Analysis Tutorial Version 4.1 button to update the value for CTRL in the Contributor list. 9. Run the Monte Carlo simulation. Note that the results appear to be slightly non-gaussian, as the tail on the right (positive) side of the distribution is larger, due to the transitional contact condition. 10. Compare the results from the Monte Carlo simulation with those of the RSS study: Monte Carlo - 10

13 Run a Monte Carlo simulation when the wheel is close to the arc RSS Results Monte Carlo Results (Typical Values) Mean Sigma % in Tolerance % 94.0% % 11. Select the CTRL dimension, and then in the Tolerance area of the Properties Explorer, set the MC Distribution to User Specified and select Uniform as the distribution type. 12. Click on the Refresh button to update the distribution type for CTRL in the Contributor list. 13. Click on the Clear simulation results button to reset the Monte Carlo simulation. This will clear the results of the previous simulation without deleting the contributor list. 14. Run a new simulation. Monte Carlo - 11

14 Enventive Monte Carlo Analysis Tutorial Version 4.1 Note that the results are now clearly non-gaussian; the tail on the right (positive) side of the distribution is even more pronounced. 15. Compare the results between the RSS tolerance analysis report, the Monte Carlo simulation that used a Gaussian distribution for the CTRL dimension, and the simulation that used a Uniform distribution for the CTRL dimension: RSS Results Monte Carlo Results - Gaussian (Typical Values) Monte Carlo Results - Uniform (Typical Values) Mean Sigma % in Tolerance % 94.0% % 83.0% % In this case, both of the Monte Carlo simulations have produced results that are significantly different from the RSS tolerance analysis report. The RSS method relies on a few key assumptions: Each contributor s sensitivity value is computed for the Nominal configuration, and that sensitivity value is assumed to be constant throughout the tolerance Monte Carlo - 12

15 Run a Monte Carlo simulation when the wheel is close to the arc range. The list of contributors is also identified based on the Nominal configuration, and no additional contributors are involved in the stack-up (points of contact don t transition onto different features). All of the dimensional contributors have Gaussian distributions. The above assumptions do not apply to this tolerance study, and therefore the RSS method will not give accurate results, as evidenced by the following: The transitional contact location introduces additional contributors. As the CTRL dimension changes from 32 to 28, the roller circle transitions from the bottom line onto the arc, and six new contributors appear in the RSS tolerance report. The transitional contact location also results in changes in the sensitivity values for most of the contributors. By comparing the two Monte Carlo simulations, we can observe a significant difference between the two sets of results, depending on whether the CTRL dimension has a Uniform distribution or a Gaussian distribution. For more information about RSS and sensitivity calculations, see Enventive Online Help. Congratulations! You've completed the Monte Carlo Simulation Tutorial. Monte Carlo - 13