On-line fault diagnosis method for sewage treatment based on weighted extreme learning machine based on kernel function

Abstract

The invention discloses a weighted extreme learning machine sewage treatment online fault diagnosis method based on a kernel function, comprising the steps of: 1) eliminating data with incomplete attributes in the sewage data, and then normalizing the data to determine the historical data set and Update the test set; 2) Select the kernel function and weighting scheme, and then determine the model parameters according to the optimal model; 3) Assign weights to each sample of the historical data set according to the selected weighting scheme; 4) Train the model, according to the kernel function Calculate the kernel matrix; 5) Add new samples from the updated test set to the model for testing, and update the historical data set; 6) Return to step 3), retrain the model, and repeat the above process until the online test data is completed, thereby realizing Identification of the online operating status of the sewage treatment process. The method of the invention has short update time and high classification accuracy, and is of great significance for real-time diagnosis of operating faults, safe operation of sewage treatment plants, and improvement of operating efficiency of sewage treatment plants.

Description

technical field [0001] The invention relates to the technical field of sewage treatment, in particular to a kernel function-based weighted extreme learning machine sewage treatment online fault diagnosis method. Background technique [0002] With the rapid development of modern industry and the growth of population, a large amount of domestic sewage and industrial waste water are discharged into water bodies, and the water resources that human beings rely on for survival have been greatly damaged. Water pollution has become one of the main factors restricting human development. In order to prevent further deterioration of the water environment, many countries have taken active measures to build a large number of sewage treatment plants. However, the sewage biochemical treatment process is very complicated, and there are many influencing factors. It is difficult for the sewage treatment plant to maintain long-term stable operation in the actual operation process, and it is ea...

AI Technical Summary

Problems solved by technology

However, the sewage biochemical treatment process is very complicated, and there are many influencing factors. It is difficult for the sewage treatment plant to maintain long-term stable operation in the actual operation process, and it is easy to cause serious problems such as substandard effluent quality, increased operating costs, and secondary environmental pollution. Therefore, it is necessary to Monitor the operation status of the sewage treatment plant, diagnose and deal with the faults in the sewage treatment process in time

[0003] The fault diagnosis of sewage treatment operation status is essentially a pattern classification problem. In the actual state operation classification, the problem of unbalanced distribution of sewage data sets is often encountered. When traditional machine learning methods are used for unbalanced data classification, The correct rate of model classification cannot meet the requirements, which brings great difficulties to the fault diagnosis of sewage biochemical treatment; at the same time, in the actual process, fault diagnosis is actually a continuous learning process, and one of its outstanding features is that learning is not an offline process is carried out, but the data is added one by one, and the process of continuous optimization

Images