In English

Using Ergonomic Criteria to Adaptively Define Test Manikins for Assembly Operations

Peter Mårdberg
Göteborg : Chalmers tekniska högskola, 2014. 51 s.
[Examensarbete på avancerad nivå]

Digital Human Modeling software is an important tool in virtual manufacturing that allows simulation of manual assembly work. Thus, it is possible to evaluate the ergonomics of the workers in an early stage of the development, long before any physical product has been built. Furthermore, it has also been shown that the ergonomics is related to the production quality, and is thereby an important factor to consider in order maintaining a sustainable high quality production. In order to proper evaluate an assembly station, a set of manikins that are capture all the relevant ergonomic aspects has to be used. However, how to select these test manikins is a non-trivial task. This, since it is a tedious and time-consuming process to identifying which anthropometric variables that need to be considered in the assembly simulation. When several anthropometric dimensions are considered, the designer must either use a set of predefined test manikins, or manually determine which anthropometric variables that should be used when the test manikins are generated. However, when a more complex assembly task is evaluated, how can the designer ensure that the test manikins accommodate the desired population? Or does there exist manikins that suffer from bad ergonomics even though all the tested manikins turned out well? In this thesis, we propose a new algorithm for automatically building a set of test manikins called Adaptive Ergonomic Search (AES). Different manikins perform the assembly operation and the ergonomics is evaluated. The anthropometric variables that affect the ergonomics are identified and used to iteratively build up the set of test manikins. The algorithm considers the whole assembly operation and test manikins are generated from the entire set of anthropometric data. The algorithm has been compared with a percentile and to boundary methods on assembly cases from the automotive industry. The results shows that the AES algorithm generates test manikins that have worse ergonomics or may not be able to perform the assembly operation, than those manikins generated by the compared methods.

Nyckelord: Digital Human Modeling, Sampling Algorithm, Response Surface Method



Publikationen registrerades 2014-02-26. Den ändrades senast 2014-02-26

CPL ID: 194181

Detta är en tjänst från Chalmers bibliotek