Research on Non-Linear Power Coal Blending Model and its Algorithm
Based on the common used coals and designed coal in a power plant in Hunan, a non-linear coal blending model was built. Exhaust algorithm, genetic algorithm and simulated annealing algorithm were used to solve the model respectively using the lowest cost of the blended coal as the objective function. Three kinds of single coals were blended together according to certain proportions based on 15 kinds of common used single coals. The coal characteristics of blended coals were predicted by the model of General Regression Neural Network. The cost, fitness and predicting time obtained by three different algorithms were compared and analyzed. The results show that the solution is ideal in cost and fitness with exhaust algorithm while the predicting time is very long. GA algorithm is good in cost, computing time and reliability.