Combined economic and emission dispatch by ANN with backprop algorithm using variant learning rate & momentum coefficient

Multi-layered feed-forward artificial neural network (ANN) trained by back-propagation algorithm is used to solve the problem of combined economic and emission dispatch in this paper. The system of generation associates thermal generators and emission involves oxides of nitrogen only. Equality constraints on power balance as well as inequality constraints on generation capacity limits of the generators and transmission loss are also considered. The idea is to minimize total fuel cost of the system and control emission. The problem is first optimized by Lagrange multiplier technique and the result is used to train the ANN wherein tuning parameters eta & alpha are altered to check their effect on convergence rate. The trained ANN is then used to generate test data. It is found that the convergence characteristic of the algorithm is excellent and the results achieved by the proposed method are quite accurate and faster in comparison to the conventional method