Operating condition classification and identification of grate coolers based on D-BIRCH algorithm

Abstract The smooth operation of the grate cooler is directly related to the energy efficiency of cement production, and the accurate control of its working condition is the key to improving the quality of cement and improving the cooling efficiency. In this paper, a method of classification and identification of grate coolers based on D-BIRCH is proposed. Firstly, by analyzing the whole cement firing process, grate pressure difference, raw feed amount, and kiln main motor flow were selected as the test parameters. Then, the D-BIRCH algorithm is used to obtain the operating condition label of each operating condition by clustering. Finally, the GBDT algorithm is used to train the condition label to get the corresponding condition recognition model. In this paper, the field operation data of a cement enterprise is selected for simulation, and the simulation results verify the feasibility of the proposed method.