Grate cooler working state monitoring method based
on image recognition

(22) 07.05.2020

(43) 15.04.2021

(57) A grate cooler working state monitoring method based on image recognition, comprising: a hardware adjustment step (S1); a camera calibration step (S2); an image data pre-processing step (S3); an edge thickness calculation step (S4); and a fluidized surface area calculation step (S5). Thus, errors caused by using grate speed of a grate cooler to indirectly estimate material layer thickness are resolved, and the function of fluidized surface area dynamic measurement is made possible, allowing the thickness and fluidized state of a material layer in a grate cooler to be measured directly. A measurement value is transmitted to an expert smart control system, and the expert smart control system, in combination with a machine learning algorithm, analyzes, calculates, and implements grate speed control, allowing the problem of unstable grate cooler operation to be avoided completely, improving the utilization rate of waste heat from cement clinker, reducing energy waste, and improving system yield and quality.

(71) Nanjing Kisen International Engineering Co., Ltd., Jiangsu 210036 (CN)