Prediction of pH Value in Desulfurization System Based on CNN-GRU and Attention Mechanism Model
CNN-GRU model and attention mechanism-based method is proposed to forecast the pH of the slurry in thermal power plant's limestone gypsum flue gas wet desulfurization system,in order to tackle the issue of inaccurate and protracted pH measurements within the absorption tower.Firstly,correlation analysis is conducted on the data stored in the SIS data base.Subsequently,Attention(Attention,ATT)is utilized to adaptively assign weight to the input data re-lated to pH value,distinguishing between strong and weak feature variables based on the weight.Finally,Convolutional Neural Network(CNN)is employed to extract and reduce the dimension of the chosen feature data.The data sent to the Gate Recurrent Unit(Gate Recurrent Unit,GRU)network are optimized so as to improve the training speed and predic-tion accuracy of the neural network.Test of the operation data of 2×350 MW units in a domestic power plant verifies that the pH value prediction model has higher accuracy and stability by comparing it with other mainstream algorithms.Finally,the practicability of the model is verified by combined with the MPC.
attention mechanismconvolutional neural networkgate control loop unitfluid pH valueprediction model
万方数据
维普
