基于RBF神经网络与强化学习的烧结机电除尘系统智能控制研究OA

Aiming at the contradiction between stable dust mass concentration control and reduced operational energy consumption in traditional 360 m2 sintering machine electrostatic precipitator systems,this paper proposes a multi-objective intelligent control scheme based on a Radial Basis Function(RBF)neural network and reinforcement learning.The system collects real-time comprehensive electric field parameters via the Modbus/TCP protocol,utilizes the RBF neural network to accurately predict dust mass concentration trends,and designs a reinforcement learning algorithm with a"concentration-energy consumption"core reward function to dynamically coordinate and optimize the output strategy of each electric field power supply.Industrial application results show that the proposed control system can not only stabilize the outlet dust mass concentration below 50 mg/m3 but also reduce the comprehensive system energy consumption by approximately 15%under the same operating conditions,achieving a win-win outcome of environmental compliance and energy savings.