Pareto Fronts of Mixed Coal Quality and Cost in Power Plant Based on Chance Constraints

Abstract

Coal blending models usually take the composition or properties of coal as constraints to optimize the cost of mixed coal for the blending scheme in power plant. Properties of coal are essentially not optimized with this kind of blending model. Therefore, the maximum likelihood quality of design coal of boiler was defined, when the compositions of raw coal can be regarded as random variables. A multi-objective optimization model of coal properties and cost based on chance constraint was established. Pareto fronts of the multi-objective model were presented by genetic algorithm. Analysis of Pareto fronts shows that the optimal data of coal properties and cost are reasonable, and the chance constraints are met well. The blending model can employ the stability of mixed coal as the optimization objective additionally. According to different characteristics of the generator unit and the actual coal to be blended, the optimization model can have different objectives and constraint combinations, which means the good flexibility and practicability.

Keywords

• thermal power generation
• coal blending
• maximum likelihood principle
• chance constraint
• multi-objective programming
• pareto fronts