Expert control system for material layer depth of chain grate based on Q_learning algorithm

Due to the big fluctuation of raw ball quantity in the material layer depth control system of chain grate, the instability of return fines quantity generated by palletizing qualification rate and the delay of material layer depth detection, a automatic control method of material layer depth is proposed based on adaptive expert control. Material layer depth is taken as a main control objective for machine speed of chain grate. The machine speed of chain grate determined by raw ball quantity is realized by expert control rule, and the fine adjustment of machine speed is performed by material layer depth. During the control process, expert control rule has self-learning and self-tuning ability to reduce the influence on material layer depth control caused by the instability of return fines quantity, and Q_learning algorithm is adopted for expert self-learning parameters to update expert database quickly. The effectiveness of control strategy is verified by the practical operation results, and the control error of material layer depth is in the control range of set values.