Controlling and Monitoring Milk Pasteurization using Fuzzy Logic integrated with the Internet of Things (IoT)

Abstract

Precise temperature control in nonlinear thermal processes with fluctuating conditions remains a challenge for traditional control methods. This research presents a Mamdani-type fuzzy logic control system integrated with an IoT architecture for milk pasteurization using the Low-Temperature Long-Time (LTLT) method, which is widely used in Indonesia for small-scale dairy production. The controller takes milk temperature and volume as linguistic inputs and outputs a continuous PWM signal to regulate heater power. Unlike traditional on–off systems or model-based PID controllers, this fuzzy logic approach does not rely on an explicit mathematical model and remains effective across different milk volumes without retuning. Experiments with 3 L, 5 L, and 8 L of milk show that the controller keeps temperatures close to the 64°C target, with averages of 64.11°C, 64.07°C, and 64.03°C, respectively. Max overshoot is limited to 1.56%, 0.29%, and 0.19%, while high-temperature stability is demonstrated by standard deviations of 0.26, 0.08, and 0.12, indicating robustness. Furthermore, it functions at a lower average PWM duty cycle compared to on–off control, leading to smoother operation and improved efficiency. This system can handle nonlinear thermal processes with varying loads and is supported by real-time IoT connectivity monitoring.