Abstract:
The neural network method to forecast the heating value of Municipal Solid Waste (MSW) is a new way to estimate heating value of MSW by using statistics of economy and natural environment. Indicated by Grey Correlation Coefficient, there are four key factors that affect the heating value of MSW, including rate of gas utilization, annual rainfall, per capita annual living expenditures of urban households and GDP.A social-economic statistics based BP neural network estimation model of the heating value of MSW was established, by using MATLAB with the statistics of those key factors and the historical tracking results of heating value of MSW from Chendu City. The closed set and open set test show that the model has a high precision and stabilization. It can do simulation effectively with data outside of the training sample and can perform fast training.