Automotive Fault Diagnostic Research And Design

The car has become an indispensable means of transport in our society now. Due to the reasons such as wear and aging and so on, automotive fault often happen, brought huge threat to safety of life and property of people. In the 1940s～1950s, automotive fault diagnosis had began abroad. There have been more and more auto in our country recently and auto fault diagnosis is developing quickly.Firstly, this paper introduces the contents of auto fault diagnosis. Then we study the auto mechanical fault diagnosis from two aspects that is fault characteristics and the theoretical of auto fault diagnosis, and in the while we analyze the reason of mechanical fault and diagnostic methods detailedly. Auto fault diagnosis use sensor technology for obtaining automotive real-time data, use time or frequency domain numerical analysis method to extract fault features, use artificial intelligence (such as expert systems, neural networks, etc.) for fault diagnosis, improving auto fault diagnosis greatly.Among neural network, BP neural network is used for automotive fault diagnosis. In practice, there have been a logical relationship between the fault phenomena and fault, fault and fault characteristics, and BP neural network can remember these logical relationships by learning sample and can be good for automotive fault diagnosis.Finally the design about hardware and software of this automotive fault diagnostic instrument had been completed, and realized part diagnostic function.