Electrical fire warning algorithm based on improved WOA-BP neural network
Electrical fires constitute a severe threat to both personal safety and property,emphasizing the critical need for enhanced early prediction and warning systems.With the aim of improving the accuracy of electrical fire prediction,a predictive model has been constructed by employing an enhanced whale algorithm to optimize a BP neural network.This model utilizes residual current,working current voltage,and cable temperature as input features for the neural network,in conjunction with the aforementioned enhancement method for optimizing weights and thresholds.The optimized parameters are utilized as initial settings for model training,enabling the prediction of electrical fire probabilities.Experimental data from electrical cabinet circuits are employed,with the predicted probabilities subject to fuzzy processing relative to the duration of abnormal residual current.Consequently,this approach leads to the formulation of fire-related decisions.Research findings indicate that the proposed model achieves a correlation coefficient of 0.97,representing an improvement of 0.08 compared to traditional methods,showcasing higher accuracy and reliability.
electrical fire warningwhale optimization algorithmBP neural networkfuzzification
万方数据
维普
