基于BP神经网络的城市垃圾热值预测

For the complexity and variability of MSW composition, using compiled software of MATLAB, the paper establish BP neural network model to forecast the municipal solid waste heat the value, In this model, the physical composition (plastic, paper, glass, wooden bamboo and fabrics, etc.) of approach waste, of the tiger dump is used as the input, found that when the nodes of hidden layer are 5, the network convergence speed is faster than other nodes, and the forecast bias is the smallest, thus, deciding the structure of predictive models is10-5-1. Using this model to predict the approach calorific value of a garbage dump, the results show that the maximum relative error of the predicted and measured values is less than 3%, indicating that the network prediction model is very practical.