Investigation of Diagnostics for Buses

Sammanfattning: As vehicle functions are getting more complex, the difficulty of diagnosing work is increasing. Currently, at Volvo Bus Corporation, there are two main methods to do fault tracing on the body, but there is no guided diagnostics, which means if something wrong happens, we can get a warning but we have no idea how to locate the error component efficiently, the diagnosis work depends largely on the expert’s experience. In this thesis, the front door system of a Volvo bus is chosen to be a target system to try the model-based diagnosis method. The system is modeled with a model-based reasoning software called RODON. With the help of the tool, the model can be created, modified and simulated. Based on the model, the corresponding database can be generated automatically after some constraints are set, then we can generate the decision tree based on the database, that is what we expect from the model. And with the help of the decision tree, the mechanics can be guided to the right direction to find out the erroneous component precisely and efficiently. The results of guided diagnostics were good and all the faults injected on the bus are detected and the guided diagnostics method was proved more efficient than other existing methods. Based on the result presented in the thesis, we can draw a conclusion that the guided diagnostics can be integrated to the already existing method as a complement.