Ahmad Azari - Chemical Engineering Dept., Engineering School, University of Tehran, I. R, Iran
Mojtaba Shariaty Niasar - Chemical Engineering Dept., Engineering School, University of Tehran, I. R, Iran
Mahmoud Alborzi - Chemical Engineering Dept., Petroleum Industry University, I. R. Iran

n this work, we forecast the gas demand load for Tehran city, based on the most important weather parameters, by using artificial neural network with multilayer back propagation, BP algorithm. At first, the effective daily temperature will be determined and then the data for last days for network training were used. The main advantage of this work is a good agreement of almost 93% with the real data. This network can further be developed to forecast gas load of other cities of Iran.

Gas Load, Forecasting, Neural Network, Multi Layer Perceptron, Back Propagation