Distributed model-based predictive control for peak load management using Arduino-based LoRa communication

초록

This study proposes a multi-zone distributed model-based predictive control (MZ-DMPC) method for real-time data collection and peak load reduction in multi-zone buildings. The method utilizes Arduino-based long range (LoRa) communication and a distributed optimization algorithm. First, grey-box modeling was performed using historical data collected via LoRa communication. The target building is a six-story campus building with a two-zone experimental chamber on the first floor. MZ-DMPC simulations achieve about 47% peak load reduction while maintaining electricity cost savings similar to those of basic MPC. Future implementation research will focus on real-time data collection using LoRa communication and integrating infrared communication for automatic control. The proposed LoRa-based MZ-DMPC offers a cost-effective and scalable solution for smart city applications, with the potential to reduce peak loads in both individual buildings and large complexes. © 2025 Building Simulation Conference Proceedings. All rights reserved.