Closed-loop system tiny fault detection and estimation method based on data driving

Abstract

The invention discloses a closed-loop system tiny fault detection and estimation method based on data driving, and mainly relates to the technical field of fault diagnosis. The method comprises the following steps: S1, selecting a plurality of working points for a system; S2, calculating P *, T *, b *, [mu]*, [sigma]* and [lambda]*, and selecting a principal component space; S3, determining a fault detection threshold by using an approximate chi-square distribution hypothesis; S4, performing initialization; S5, when a new sampling value is obtained, recording the new sampling value as xk + 1, n, calculating a mean value, and then calculating a score vector value; S6, calculating a mean value and a variance update value of a score vector; S7, calculating the KL distance Kn (tf, t*) of different score vectors of the working point n; S8, estimating the fault amplitude of the working point n; S9, letting k = k+1, and returning to step S5; and S10, judging the fault type of each working point through fuzzy clustering fault diagnosis. The plurality of working points of system operation are fully considered, data in the working range of the system are fully utilized to carry out fault diagnosis on each working point, and the accuracy and robustness of fault diagnosis are improved.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis, in particular to a data-driven closed-loop system micro-fault detection and estimation method. Background technique [0002] With the improvement of equipment performance requirements in industrial systems, a large number of advanced algorithms containing closed-loop control laws, such as proportional integral differential control, optimal control and robust control, have been applied to various control systems to achieve stable operation of the system. However, while the feedback mechanism improves the robustness of the system, it also increases the difficulty of fault diagnosis of the closed-loop system. Because when the fault is in the early stage or the amplitude is small, the impact may be covered by the feedback control amount, resulting in the residual signal of the system when the fault occurs may still fluctuate within a small range, that is, the degree of deviation of the observ...

AI Technical Summary

Problems solved by technology

However, due to non-linearity and multiple control objectives, the actual engineering system often has multiple operating points, and the system switches between various operating points during operation. , while the traditional micro-fault diagnosis for a single working point cannot make full use of the data set of the system operation, and the diagnosis results have certain limitations

Images