Objective The environmental parameters of fruit and vegetable preservation are characterized by time-varying, non-linear, strong hysteresis and large inertia, which leads to the problems of low control accuracy and weak robustness of traditional control methods. The goal was to develop a new control strategy to address these problems.

Method We combined conventional proportional-integral-derivative (PID) and back-propagation neural network (BPNN) algorithms to develop a control strategy based on BPNN-PID. We studied the effect of using different control strategies on the regulation of freshness environment parameters through an independently built test platform for environmental regulation in fruit and vegetable preservation and an independently designed control system.

Result The experimental results showed that the temperature overshoot of the environment control system based on BPNN-PID control strategy was 1.7 ℃, the stable time was 80 min, and the steady-state error was 0.2 ℃. The overshoot of environmental relative humidity was 2.8%, the stable time was 55 min, and it was stable in the range of 80%–90%. Compared with conventional PID control strategy, the ambient temperature overshoot of BPNN-PID control strategy was reduced by 2.1℃, the steady-state error was reduced by 0.3 ℃, and the steady-state time was shortened by 25 min. The environmental relative humidity overshoot was reduced by 2.2%, the stabilization time was shortened by 25 min, and the fluctuation ranges of environmental parameters were reduced.

Conclusion The system shows good dynamic adjustment ability, strong robustness and obvious improvement in control performance, which enables accurate control of environmental parameters of fruit and vegetable preservation and meets the requirements of fruit and vegetable preservation and storage. The research results can provide references for the regulation of environmental parameters of fruit and vegetable preservation.