One of the weak areas in designing parts is deciding tolerances of various parts. Most engineers are familiar with Arithmetic Tolerance stack up analysis wherein they check impact of maximum and minimum values of various tolerances on assembly of parts. However, this can often result in high manufacturing cost. Thus, it may be more appropriate to analyse tolerances using statistical tolerance stack up approach. This can be performed using Monte Carlo Simulation. In one of the previous videos, I had shown how to predict reliability using Monte Carlo simulation. In this video, I will explain how to perform statistical tolerance stack up analysis using Monte Carlo Simulation. I have used Simular software to demonstrate this with a practical example. Simular is a free software (emailware) which can be downloaded from https://www.simularsoft.com.ar/. However, one can use other software such as Crystal Ball, @Risk etc. for such analysis. Statistical tolerance stack up is usually an integral part of Design for Six Sigma (DFSS) projects. I hope viewers will find this video useful. Feedback is welcome!
Telkom University says
What are the key methodologies employed in statistical tolerancing using Monte Carlo simulation?
Hemant Urdhwareshe says
The main concept used in statistical tolerancing is property of additivity of variances. We cannot add standard deviations but squares of standard deviations (i.e.variances) can be added. To perform Monte Carlo simulation, one needs to know the distribution of the variables, and a mathematical model of their relationship. In statistical tolerancing, the variables are the dimensions, and the model is based on how these are assembled. In some cases, the variables can be properties: for example, in electronics, these variables can be capacitance, resistance, inductance etc. In Monte Carlo simulation, we generate large number of simulated data of the variables and use the model to predict the resulting distribution. Hope this helps to some extent. You may like to watch my video on use of Monte Carlo in design of a spring.
Here is the link: https://youtu.be/8-taU-UKnLk?si=gomsO5pEjYEHAnMF