An SVR-Based Method for Detection and Diagnosis of Coal Mill Efficiency Abnormality

Abstract

The invention discloses a method for detecting and diagnosing abnormal coal mill efficiency based on SVR, which includes the following steps: firstly collect the normal operation data of the coal mill, divide the training set and the verification set according to the proportion, and collect the abnormal / real-time data as the test set . Identify performance variables as model outputs and corresponding operating variables as model inputs, train the model, and use the validation set to determine control limits. Then input the test set and calculate the deviation of the target variable; if the deviation exceeds the control limit, extract the support vector and compare it with the abnormal working conditions to determine the magnitude and direction of the deviation of the variable, and give the order of the variable contribution. Finally, by integrating all efficiency-related models, the causal variables that cause efficiency anomalies and the propagation paths of efficiency anomalies can be determined. By using the method of the invention, the abnormal condition of the efficiency of the coal mill can be detected, and the mining of abnormal variables and the diagnosis of abnormal causes can be realized.

Description

technical field [0001] The invention relates to the field of fault detection and diagnosis of complex industrial equipment, in particular to an SVR-based method for abnormal detection and diagnosis of coal mill efficiency. Background technique [0002] The coal pulverizer is an important auxiliary machine in the fuel system of a thermal power plant. It is used to grind coal into powder and then transport it into the boiler for combustion through primary hot air. If the fineness or humidity of the coal pulverized by the coal mill is not suitable, it will lead to blockage of the combustion pipeline, slagging, scaling, etc., which will affect the combustion efficiency; and if the supply of pulverized coal is insufficient, it may cause boiler failure or shutdown. Therefore, the combustion efficiency of fuel in a thermal power plant largely depends on the production performance and efficiency of the coal mill. The performance of the coal mill may also be affected by factors such...

AI Technical Summary

Problems solved by technology

[0004] The invention provides a method for detecting and diagnosing coal mill efficiency anomalies based on Support Vector Regression (SVR), which overcomes the problems of relying on expert experience and fault samples in the current existing methods, making it difficult to find new anomalies, and difficult to explain the detection results question

Images