Aero-engine pump dispatching system anomaly detection method based on improved DBSCAN algorithm

Abstract

The invention provides an aero-engine pump dispatching system anomaly detection method based on an improved DBSCAN algorithm. Characteristic parameters are selected according to working characteristics of the aero-engine pump dispatching system, and then characteristic extraction is carried out from parameter data to obtain an input characteristic matrix of the algorithm. Meanwhile, according to the characteristics of input data of the algorithm, an innovative anomaly detection algorithm combining dynamic time warping and a density-based clustering algorithm with noise is provided. Compared with a traditional anomaly detection algorithm based on normal distribution, the anomaly detection result of the algorithm provided by the invention is more accurate. According to the method, the defectthat the density-based clustering algorithm with noise cannot mine the relationship among the time sequences in the aspect of processing the time sequence problem is overcome, and the density-based clustering algorithm with noise is combined with the dynamic time warping, so that the capability of the algorithm for processing the time sequence problem is improved.

Description

technical field [0001] The present invention aims at the aero-engine pumping system, and utilizes a density-based clustering algorithm (Density-BasedSpatial Clustering of Applications with Noise, referred to as DBSCAN) combined with a dynamic time warping (Dynamic Time Warping, referred to as DTW) method to control the pumping system. Abnormal detection is carried out during the starting process, which can effectively improve the health monitoring ability of the pump adjustment system. Background technique [0002] As a complex aerodynamic thermodynamic system, aero-engines are of great significance to ensure and improve the safety and reliability of engine work by monitoring their working status and performance parameter change trends, and avoiding failures in a timely and effective manner. Engine data is the basis for avoiding faults. However, with the gradual improvement of system reliability, it is difficult to obtain engine fault data in a short period of time, so the t...

AI Technical Summary

Problems solved by technology

In view of the high sensitivity of the radius parameter ε in the traditional algorithm, it is difficult to select global parameters for multi-layer density data sets, which leads to poor clustering results and other shortcomings.

In 2020, Guo et al. proposed a GS-DBSCAN algorithm based on similarity measurement to solve the problem that the traditional DBSCAN algorithm has poor clustering effect on high-dimensional data sets and the selection of parameters is sensitive.

Images