In English

Toolbox for Statistical Analysis of Load and Strength in Vehicle Engineering

Efrem Efremov ; Benjamin Grozdanic
Göteborg : Chalmers tekniska högskola, 2018. Master's thesis - Department of Mechanics and Maritime Sciences; 2018:20, 2018.
[Examensarbete på avancerad nivå]

In the fatigue reliability assessment of the automotive industry there is a need for capturing and accounting for the variation in customer load measurements. This is what the variation mode and e ect analysis(VMEA) aims to do, but there is no easy and systematical way of using it and there is little information on the methods robustness in the face of random variations. This thesis is implementing the VMEA method, that has been modi ed for the vehicle industry, by using the programming language Python, thus creating a VMEA toolbox and then testing it on customer load and component strength data, provided by Volvo Group Trucks Technology. To test the robustness of the VMEA method a parameter study and a sensitivity study are performed. In the parameter study the four di erent input parameters, which are prone to change between cases, are tested. The parameter study is performed in order to provide a framework for how each input parameter a ects the end results for the VMEA method. The aim of the sensitivity study is to test the robustness with regard to fewer customer measurements. The rst part of the sensitivity study consists of strategically removing customer load data in order to determine both the amount of data and which data is needed to receive reasonable results. In other words, to determine how the variation of magnitudes, of the customer loads, a ects the VMEA results. The second part of the sensitivity study consisted of randomly removing di erent amount of customer load data multiple times. From the results of this thesis it is observed that VMEA is reliable and a robust method for doing fatigue reliability assessments in the automotive industry, on a component level. When it comes to the four input parameters, it can be concluded that the VMEA method is robust when using reasonable estimations of input parameters. For the sensitivity study, VMEA shows robustness when missing customer data and only at a few points have a striking deviation, but even then provides conservative results. Overall, the conclusion is that the variation has the largest impact on the safety factors and thus it is of great importance to capture it as good as possible in the full scale test. Finally, it is safe to say that VMEA is a good way of doing fatigue reliability assessments as it is taking into account all statistical uncertainties and variations when doing so. Additionally it provides the option of adding uncertainties for which there exists no rigorous statistical data ones and providing established safety factors.



Publikationen registrerades 2018-07-05.

CPL ID: 255553

Detta är en tjänst från Chalmers bibliotek