The process of finding out whether the system and equipment have faults by various inspection and testing means is fault detection; The process of further determining the approximate location of fault is fault location. Fault detection and fault location belong to the category of network survivability. When maintenance is needed, the process of locating the fault to the replaceable product level (replaceable unit) is called fault isolation. Fault diagnosis refers to the process of fault detection and fault isolation.

The performance indicators for evaluating the fault diagnosis system are:

1) Timeliness of fault detection: refers to the ability of fault diagnosis system to detect faults in the shortest time after the system fails. The shorter the time from fault occurrence to fault detection, the better the timeliness of fault detection.

2) Sensitivity of early detection: refers to the ability of fault diagnosis system to detect tiny fault signals. The smaller the fault signal that the fault diagnosis system can detect, the higher the sensitivity of its early detection.

3) False alarm rate and false alarm rate: False alarm means that the system is not out of trouble, but it is wrongly detected to be faulty; Missed report refers to the failure of the system that has not been detected. A reliable fault diagnosis system should reduce the false alarm rate and false alarm rate as much as possible.

4) Fault separation ability: refers to the ability of the diagnosis system to distinguish different faults. The stronger the ability of fault separation, the stronger the ability of diagnosis system to distinguish different faults and the more accurate the fault location.

5) Fault identification ability: refers to the ability of the diagnosis system to identify the fault size and time-varying characteristics. The higher the fault identification ability, the more accurate the fault identification of the diagnosis system, which is more conducive to fault evaluation and maintenance.

6) Robustness: refers to the ability of the diagnosis system to correctly complete the fault diagnosis task in the presence of noise and interference. , while maintaining a low false positive rate and a low false negative rate. The stronger the robustness, the higher the reliability of the diagnosis system.

7) Adaptability: It means that the fault diagnosis system has adaptability to the changed tested object and can make full use of the new information generated by the change to improve itself.

In practical application, it is necessary to analyze and judge the above performance indicators according to the actual situation, and then analyze the diagnosis methods, and get the final diagnosis scheme after appropriate selection.