配煤是选煤厂商品煤销售的重要环节，但目前普遍存在配煤质量波动大、人工依赖程度高等问题。为了解决动力煤配煤的问题，以顾桥煤矿商品煤配煤过程为研究对象，首先采集商品煤质量数据并建立灰-水-发热量动态数据库，结合滑动窗口方法进行数据划分，并建立了发热量多元线性拟合最优模型；然后基于产品仓料位高度、进仓快灰标准差和发热量，构建了筒仓评分机制并进行等级划分，结合专家经验，制定了筒仓出煤规则库，确定参与配煤的筒仓编号；最后依据质量守恒和能量守恒，求解不同产品仓配煤量。顾桥煤矿商品煤配煤实践表明：发热量多元线性拟合模型准确率最高达到了89.53%；根据评分机制计算出评分函数的各项系数，并结合仓内商品煤的发热量、实际仓存确定了最佳配仓仓号及相应出煤量；系统决策与现场专家决策吻合度超80%，证明了配煤决策的有效性。

Coal blending is an important link in coal preparation plants in sales of their commercial, yet wide fluctuation of quality of blended coal and high reliance on experience of operators are a general problem nowadays. To deal with this situation, the saleable coal blending process applied at Guqiao Mine is taken as the object for case study. The study is made in a procedure as follows: to start off, establish a dynamic database coal ash, moisture and calorific values based on collected saleable coal quality data; after that, partition the data using the sliding-window method for constructing the multivariate linear fitting optimal model of calorific values; then, establish the silo rating and grading mechanism based on coal levels in silos, the standard deviations of express ash analysis and calorific values of the coal loaded into silos; following that, setup a rule library for the coal out loaded from silos based on expert experience and number the silos in which the coal blends are stored; finally, calculate the required quantities of the coal blends from the numbered silo in which different grades of coal are stored, according to the mass and energy conservation principles. Practice shows the accuracy of the multivariate linear fitting model is up to 89.53%; the optimal numbers of the silos and the quantities of coal required to be discharged from them can be determined according to the coefficient of each term of the rating function calculated based on rating mechanism, the calorific values and actual volumes of coal in silos; the consistency between the decision made by the system and by onsite experts exceeds 80%, fully proving the effectiveness of the power coal blending technology.