Transformer oil temperature prediction method based on improved particle swarm optimization neural network algorithm

Abstract

The invention provides a transformer oil temperature prediction method based on an improved particle swarm optimization neural network algorithm, and is used for solving the technical problem that an existing transformer oil surface temperature prediction method is low in prediction precision. The method comprises the following steps: firstly, taking one or more variables in related influence factors such as environment temperature, load change and cooler group number as input variables, and taking transformer oil surface temperature as an output variable; secondly, optimizing a weight and a threshold value of the BP neural network model by adopting an improved particle swarm algorithm, and training the optimized BP neural network model to obtain a transformer oil temperature prediction model; and finally, predicting the oil surface temperature of the transformer by using the transformer oil temperature prediction model. The prediction result of the method is highly matched with the real data of the oil surface temperature, the relative error is lower, and the effectiveness of the method for predicting the oil surface temperature of the transformer is verified.

Description

technical field [0001] The invention relates to the technical field of transformer oil surface temperature prediction, in particular to a transformer oil temperature prediction method based on an improved particle swarm optimization neural network algorithm. Background technique [0002] As the core equipment of electric energy conversion and distribution, the health level of the transformer determines the stability of the power system operation. With the increasing scale of the power system, the number of transformers is gradually increasing. Among transformer operating parameters, oil surface temperature is an important parameter for judging its operating state. Therefore, the accurate prediction of transformer oil surface temperature provides a certain reference for the judgment of transformer operating conditions. [0003] At present, scholars at home and abroad have established some prediction models for the prediction of transformer oil surface temperature. Literature...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the technical problem of low prediction accuracy in existing transformer oil surface temperature prediction methods, the present invention combines BP neural network and particle swarm algorithm, and proposes a transformer oil temperature prediction method based on improved particle swarm optimization neural network algorithm, Considering the different working conditions of the three factors of ambient temperature, load change, and the number of cooler groups alone or interacting with each other, the oil surface temperature of the example transformer is predicted by using BP neural network, genetic algorithm optimization neural network and improved particle swarm optimization neural network respectively. And analysis, the analysis results show that the prediction error of the transformer oil surface temperature by the improved algorithm is relatively small, which provides a reliable basis for the judgment of the transformer operating state

Images