# Research on Temperature Control Method of Rice Noodles Extruder Based on APSO-MPC

## TL;DR

This paper introduces a new temperature control method for rice noodle extruders using APSO-MPC to improve accuracy and reduce disturbances.

## Contribution

The novel approach combines adaptive particle swarm optimization with model predictive control for enhanced temperature regulation in rice noodle extruders.

## Key findings

- The proposed method achieves a maximum temperature overshoot of 7.75% and a steady-state error within ±1 °C.
- Compared to fuzzy PID control, the APSO-MPC method shows stronger adaptability and higher control accuracy.
- The longest adjustment time recorded is 1228 seconds under various parameter combinations.

## Abstract

Aiming to address the problems of many temperature control disturbances and the hysteresis of control output in existing rice noodle extruders, a temperature control method for a rice noodle extruder based on adaptive particle swarm optimization (APSO) optimization model predictive control (MPC) was designed. Firstly, the temperature control principle of the rice noodle extruder is analyzed by combining the structure of the rice noodle extruder. The temperature balance equation of the barrel is constructed by thermodynamic analysis, and the temperature prediction model is established. The APSO algorithm is further selected to perform the adaptive parameter identification of the model based on the collected input/output data. Then, aiming at high-precision temperature control, the objective function is constructed by combining the temperature prediction value and the reference trajectory, and the objective function is optimized to obtain the optimal control sequence. At the same time, the feed rate is selected as feedforward, the feed rate change is monitored by detecting the feed screw speed, and the optimal control sequence is corrected to eliminate the interference caused by the fluctuation of the feed rate. The experimental results show that the maximum temperature overshoot under different parameter combinations is 7.75%, the steady-state error is within ±1 °C, and the longest adjustment time is 1228 s. Compared with fuzzy PID control, it has stronger adaptability and higher control accuracy.