Research On Coal Blending Model And Optimization Algorithm Of Thermal Power Enterprises

In recent years, because of the serious mismatch between the existing coal and the actualneeds, the various sources of coal, and the relatively large difference in the quality of coal,almost all thermal power enterprises in China are burning a variety of mixed coal. Throughblending the raw coal, choosing the appropriate proportion to blending the coal, making thequality of mixed coal to suitable for the operation of the boiler, finally to enhance theefficiency of combustion, to reduce the wear of boiler, to lower the costs of fuel, and todecrease the emissions of pollutions. By the continuous research of the scientific staff at homeand abroad, we made some progress, but there are some problems, such as using the differentmodel to blending coal, versatility is not strong, expensive and so on.Based on a detailed analysis of current business problems in application model ofblending coal, aim at the two types of current classic models: linear models and nonlinearmodels, though a comparative analysis in the merits of model and the characteristics of coal’squality, we can found that the linear mathematical model is more favorable to the overallprocess of mixing coal. Then using the blending costs as the objective function, fiveindicators of calorific value of a single coal, volatile matter, sulfur, Moisture and Ash asconstraints, to establish a practical blending optimization model with linear relationship inthermal power enterprises. Because PSO algorithm with inertia weight has better globalsearch capability, through a lot of tests to improve PSO algorithm, the idea of majorimprovements including deformation of algorithms, mix of algorithms, discrete binary and soon, the specific implementation methods contain global optimal perturbation, combination ofvariability factor, settings of accuracy and confirm of inertia weight, and ultimately apply theimproved particle swarm algorithm to solve the model.Experimental results show that by choosing the appropriate model for blending, usingthe improved PSO algorithm to solve, we can got a conclusion that the blending proportion ofeach unit of coal is in line with the actual requirements, and the algorithm has a high stabilityand a better advantages.