Rotation machinery early stage fault diagnosis method based on heredity annealing optimization multi-core support vector machine

Abstract

The invention discloses an intelligent diagnosis method targeting the rotation machinery early-stage fault. The intelligent diagnosis method comprises steps of performing time domain, frequency domain and time frequency domain signal processing on the vibration signal of the rotation machinery on the basis of the vibration signal in the operation process of the rotation machinery, constructing a multi-core support machine as a novel intelligent diagnosis model on the basis of a typical local core function and a global core function, constructing a heredity annealing algorithm on the basis of a heredity algorithm and a heredity annealing algorithm, and using the heredity annealing algorithm to optimize the model parameter of the multi-core support vector machine to implement the multiple parameter parallel optimization. The invention fully takes the advantages that the mixing domain characteristic set performs fault gradual characteristic extraction at the early stage of the rotation machinery performance degeneration, the heredity annealing algorithm performs parallel optimization in the parameter and the multi-core support machine can perform early stage fault diagnosis, can effectively perform diagnosis identification on the early stage fault for the rotation machinery device and has a strong interference resistance capability and a capability of wide popularization.

Description

technical field [0001] The invention relates to an integrated intelligent diagnosis method for early fault diagnosis of rotating machinery, in particular to an integrated intelligent diagnosis method for early fault diagnosis of rotating machinery based on genetic annealing optimized multi-core support vector machine. Background technique [0002] The early failure of rotating machinery is the main problem that threatens the safe and reliable operation of rotating machinery. If the early failure of rotating machinery can be reliably diagnosed and the operating status of rotating machinery can be grasped in advance, the early failure of rotating machinery can be prevented from gradually developing into typical failures. , eventually leading to vicious events such as sudden shutdown of rotating machinery. Broadly speaking, early fault diagnosis of rotating machinery equipment is aimed at identifying the current operating status of rotating machinery equipment, finding out the ...

AI Technical Summary

Problems solved by technology

For rotating machinery, the vibration signal reflects its operating state most directly. Therefore, for rotating machinery, vibration signals are generally used for analysis, but the vibration signal strength that reflects the early failure of rotating machinery is relatively small. High-precision sensors can collect early fault characteristic information, but it is difficult to accurately extract the vibration characteristics of early faults of rotating machinery due to interference from environmental factors such as signal propagation paths and propagation media.

In the process of constructing the fault diagnosis model, the two-class classification method, the K-nearest neighbor classification algorithm, the artificial neural network (Artificial Neural Network, ANN) and the Support Vector Machine (Support Vector Machine, SVM) are the most widely used at present. The classification algorithm is simple, but the reliability is low. The reliability of ANN has been improved, but it is prone to over-learning or under-learning and falling into a local minimum. SVM has the best performance among them, but the selection of parameters and the construction of kernel functions are also difficult. It has a certain degree of randomness, so it is not easy to establish a reasonable and reliable early fault intelligent diagnosis model

Images