Design and Simulation of an Adaptive Fuzzy-PID Cascade Controller for Greenhouse Temperature Regulation under External Climatic Disturbances

Abstract

Performance deterioration under sun radiation fluctuations, wind-induced heat outflow and intermittent vent opening. This paper proposes an AFLC where the Mamdani fuzzy supervisor schedules an inner MPID loop gains and solar-feedforward setpoint offset while,  a mini-Particle-Swarm-a conventional PID suffersOptimization (PSO) iteratively re-tunes the rule-base-weights every ten minutes according to the last 60-minutes Integral-Absolute-Error (IAE). The multi-layered architecture was successfully tested against a single-glass, 480m2, lumped-capacitance thermal model of a greenhouse under two 48-hours 80%-monthly-day-average sun-radiation drops and random side-vent cycling scenarios. Compared to conventional-PID,  static fuzzy-PID, the AFLC reduced IAE 16.7%-20.0%, ISE 13.4%-15.8% and consumption 17.4%-19.0%, maximum overshoot was below 21% and PSO adaptation converged in less than ten iterations per cycle. The full Python simulation code and datasets are made available as open-access material.