Time-varying random system event trigger fault detection method based on coordinate transformation

Abstract

The invention belongs to the technical field of event trigger fault detection, and discloses a time-varying random system event trigger fault detection method based on coordinate transformation, and the method comprises the steps that: superposed unknown interference is removed from one subsystem by employing a coordinate transformation method; for each subsystem, an optimal event trigger filter is constructed by utilizing an unbiased and minimum variance optimal criterion; when any system breaks down, a residual error will exceed a preset threshold value, current measurement information willbe immediately sent to a remote filter, and a diagnosis decision triggered by an event does not need to be input. According to the invention, a coordinate transformation method is utilized to realizethe characteristic that the fault detection residual error is only sensitive to system faults and has robustness to additive unknown interference; a considered system can be converted into two subsystems, and interference is removed from one subsystem; and further the optimal gain of each filter is derived, and the optimal gain is applied to an optimization criterion of unbiased and minimum mean square estimation errors.

Description

technical field [0001] The invention belongs to the technical field of event-triggered fault detection, in particular to a time-varying random system event-triggered fault detection method based on coordinate transformation. Background technique [0002] At present, the closest existing technology is: (1) comparative diagnosis method, which is widely used at present, first collects and stores data on various information of various process parameters, such information includes vibration amplitude, frequency, phase, rotational speed , displacement, mode, temperature, pressure and flow, etc. Then establish a database of corresponding information, and at the same time make graphs such as trend analysis. Save a variety of information about the normal operating state of the machine and equipment. When the equipment fails, compare the current data with the data in the normal operating state, and then combine the information characteristics of the fault performance and use the diag...

AI Technical Summary

Problems solved by technology

[0005] To sum up, the existing problems in the existing technology are: the application of coordinate transformation in the existing technology for fault diagnosis detection is vulnerable to system faults, and has not been considered for time-triggered fault detection

Images