Furnace Temperature Control for MSWI Process Based on Neural Network Predictive Controller

To tackle the challenge of controlling furnace temperature in municipal solid waste incineration (MSWI), which is critical for incineration efficiency and pollutant reduction, this paper introduces a model predictive control (MPC) strategy utilizing a back propagation neural network (BPNN). Traditional PID struggles with the nonlinearity and disturbances of process. Our approach involves constructing a BPNN-based prediction model to forecast future furnace temperatures. Using gradient descent, we optimize the objective function within a predetermined period, allowing for real-time adjustment of control variables. This effectiveness of method is validated through experiments with data from a Beijing MSWI plant, demonstrating enhanced tracking and disturbance management capabilities.