In china, many thermal power plants have to burn blended coals forced by the complexity of coal type and market tension and transportation pressure of coal purchasing. As a engineering implementation method of coal blending, “different coals grinding in different mills and then mixed burning in the furnace” has many advantages such as low investment, easy to control milling system parameters and can be optimized online, etc, compared with traditional coal blending methods. But it is limited by the number of mills and cannot achieve high-precision ratio of blending. To remedy this shortcoming, a model of two-level optimization of coal blending for the thermal power plant with direct blowing pulverizing system was established in this paper. The tradional coal blending was regarded as first step of optimization. The secondary optimization was implemented by adjusting the outputs of different mills, then the blend was changed to accurate ratio. Furthermore, since the existence of coal bunker, it made a time lag from coal discharge to combustion, meanwhile, the real-time load was unpredictable and the coal utilization rate was inconsistent of each bunker. The three reasons make it uncertain of the current coal of bunker. To identify each coal in the mill(equivalent to bunker) correctly was the basis of achieving the second blending optimization. Therefore, a soft-sensing model of coal moisture based on the heat balance equation was used to take this work. At last, a intelligent coal blending system by the two-level optimization model was developed for a power plant and achieved good results.

This content is only available via PDF.
You do not currently have access to this content.