Multi sensor data fusion technology has great potential to improve system performance, and its application is more and more widely. With the deepening of research, many researchers are actively applying it to other fields in recent years, such as automatic target recognition. For fault diagnosis, the application of sensor also shows its importance and superiority.
Usually, a fault source of a system may have multiple fault representations. Therefore, it is not necessary to obtain all the characteristics of the system fault in fault diagnosis, and it is also impossible. Because it requires a large amount of detection, it will increase the cost of detection. For accurate and reliable fault diagnosis, sufficient fault characterization is still needed, at least covering all fault sources. Generally speaking, the more the amount of detection, the more reliable the diagnosis. If the number of detection is limited, it needs a very detailed knowledge of system model, fault source and fault representation.
Therefore, fault diagnosis is actually a process of finding the fault source of the system according to some fault representations obtained by the detection quantity and the mapping relationship between the fault source and the fault representation of the system. In order to make full use of the information provided by the detection quantity, a variety of diagnostic methods can be used for each detection quantity when possible. This process is called local diagnosis. It is called local fusion to synthesize the results of various diagnostic methods. By further synthesizing the local diagnosis results, the overall result of system fault diagnosis is obtained, which is called global diagnosis
Integration.
