In order to effectively deal with the complex electricity and coal market

strengthening the smart fuel management has become an important part of thermal power plant management. Aiming at solving the problems that the coal yard of a coal-fired power station occupies small area

the types of incoming coal are complex

and the coal-fired coal stacking is chaotic

by extracting the coal quality information of historical incoming coal

K-means and DBSCAN clustering algorithms are used to analyze the low-level coal. The calorific value

volatile matter and sulfur content are clustered and analyzed

and the two clustering algorithms are compared from the perspective of silhouette coefficient

cluster stability and sample division fineness

and finally K-means with better clustering effect is selected as the calculation method for coal quality division. The K-means algorithm divides the selected historical coal quality information data set into four categories

the contour coefficient is 0.587

and the coal quality components in each category are similar. The incoming coal frequency and the incoming coal weight ratio under different cluster labels are counted

and the coal yard is divided into corresponding proportions. The incoming coal of the same classification is stacked in each partition

and on this basis

the incoming coal in the digital coal yard platform is designed. Coal stacking guidance and information storage process are of great significance to improving the utilization of storage yard space and the efficiency of coal yard management.