Chemical plant power transformer fault prediction method based on particle swarm and neural network

Abstract

The invention discloses a chemical plant power transformer fault prediction method based on a particle swarm and a neural network, and the method comprises the steps: building an Elman neural network used for data prediction and a PNN neural network used for fault type diagnosis according to the type and content data of dissolved gas in transformer oil, and further building a neural network model which takes prediction and diagnosis errors as a target, the number of neurons of an Elman network hidden layer, a PNN network radial basis function distribution coefficient and an allowable maximum error rate are used as constraints, and neural network parameters are optimized by using a particle swarm algorithm, so that an accurate transformer fault prediction and diagnosis model is established, more reference information is provided for maintenance of a transformer, and the service life of the transformer is predicted and evaluated.

Description

technical field [0001] The invention relates to the field of power systems, in particular to a method for predicting faults of power transformers in chemical plants based on particle swarms and neural networks. Background technique [0002] The safe and stable operation of the power grid is the basis of reliable power supply. Once the power grid fails, it will not only damage the power equipment, stop the power supply, affect the normal production and life of the people, but also endanger public safety in severe cases, causing major economic losses and adverse social impacts. The power transformer of chemical plant is not only the pivotal equipment in the power system, but also the key equipment in chemical production. Once it fails, it will have a serious impact on the safe operation of the power grid and chemical production. With the continuous improvement of the technical level and complexity of equipment, the impact of equipment failure on production safety has also incr...

AI Technical Summary

Problems solved by technology

Most of the current transformer fault diagnosis systems use the BP network model, but due to the structural characteristics of the BP network itself, when there are many training samples and the required precision is high, the network often does not converge and easily falls into local optimum, and it is difficult to determine Insufficient number of hidden layers and hidden nodes limits its fault diagnosis ability

At present, the expert system, support vector machine, BP neural network and other methods commonly used in the field of transformer fault diagnosis are all superficial machine learning methods. The learning ability, especially the prediction ability, is limited, and there is a bottleneck in the accuracy of prediction and diagnosis. Further developed deep learning Although it has a strong learning ability, its calculation process is complex

Images