Corrosion Prediction Model of Circulating Water in Refinery Unit Based on PCA-PSO-BP

Abstract

The corrosion of circulating water in oil refinery units is prominent due to water quality problems. The establishment of corrosion prediction model based on long-term monitoring data of circulating water quality is of great significance to control the quality of circulating water and identify its corrosion state. In this paper, a prediction model of circulating water corrosion based on optimized back propagation (BP) neural network is established by using 10 kinds of circulating water quality detection indexes and coupon corrosion rate data of a circulating water field in two years. Firstly, the data collection frequency of each index is unified by downsampling, and the data normalization pretreatment is carried out. Then, principal component analysis (PCA) is used to analyze the original water quality data, and 6 new principal components are obtained as the input data of the prediction model; at the same time, in order to improve the prediction accuracy of the model, the parameters of the neural network are optimized by particle swarm optimization algorithm (PSO). Finally, the PCA-PSO-BP prediction model is established and its prediction mean absolute percentage error is 8.32%, which has a better prediction effect and generalization ability than other models.