Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An IoT Framework for Building Energy Optimization Using Machine Learning-based MPC

Aryan Morteza, Hosein K. Nazari, Peyman Pahlevani

TL;DR

This work presents an IoT framework that enables machine-learning based MPC for AHU control in buildings with limited building information, targeting legacy HVAC systems. It uses an ANN to learn daily Environmental Dynamic Functions and produce an adaptive linear thermal model that feeds a model predictive controller with a horizon of $p=48$ and a sampling interval of $t_{sampling}=30$ min, accommodating both continuous and binary actuator mappings. Experimental deployment over 126 days with 24 sensor nodes demonstrates a 57.59% reduction in electricity consumption compared with a clock-based manual controller while maintaining occupant comfort, highlighting practical benefits for upgrading existing structures. The study also identifies limitations in hot/humid conditions and suggests future enhancements such as improved sensing, energy harvesting, expansion to remaining AHUs, and cloud-based deployment to broaden impact.

Abstract

This study proposes a machine learning-based Model Predictive Control (MPC) approach for controlling Air Handling Unit (AHU) systems by employing an Internet of Things (IoT) framework. The proposed framework utilizes an Artificial Neural Network (ANN) to provide dynamic-linear thermal model parameters considering building information and disturbances in real time, thereby facilitating the practical MPC of the AHU system. The proposed framework allows users to establish new setpoints for a closed-loop control system, enabling customization of the thermal environment to meet individual needs with minimal use of the AHU. The experimental results demonstrate the cost benefits of the proposed machine-learning-based MPC-IoT framework, achieving a 57.59\% reduction in electricity consumption compared with a clock-based manual controller while maintaining a high level of user satisfaction. The proposed framework offers remarkable flexibility and effectiveness, even in legacy systems with limited building information, making it a pragmatic and valuable solution for enhancing the energy efficiency and user comfort in pre-existing structures.

An IoT Framework for Building Energy Optimization Using Machine Learning-based MPC

TL;DR

This work presents an IoT framework that enables machine-learning based MPC for AHU control in buildings with limited building information, targeting legacy HVAC systems. It uses an ANN to learn daily Environmental Dynamic Functions and produce an adaptive linear thermal model that feeds a model predictive controller with a horizon of and a sampling interval of min, accommodating both continuous and binary actuator mappings. Experimental deployment over 126 days with 24 sensor nodes demonstrates a 57.59% reduction in electricity consumption compared with a clock-based manual controller while maintaining occupant comfort, highlighting practical benefits for upgrading existing structures. The study also identifies limitations in hot/humid conditions and suggests future enhancements such as improved sensing, energy harvesting, expansion to remaining AHUs, and cloud-based deployment to broaden impact.

Abstract

This study proposes a machine learning-based Model Predictive Control (MPC) approach for controlling Air Handling Unit (AHU) systems by employing an Internet of Things (IoT) framework. The proposed framework utilizes an Artificial Neural Network (ANN) to provide dynamic-linear thermal model parameters considering building information and disturbances in real time, thereby facilitating the practical MPC of the AHU system. The proposed framework allows users to establish new setpoints for a closed-loop control system, enabling customization of the thermal environment to meet individual needs with minimal use of the AHU. The experimental results demonstrate the cost benefits of the proposed machine-learning-based MPC-IoT framework, achieving a 57.59\% reduction in electricity consumption compared with a clock-based manual controller while maintaining a high level of user satisfaction. The proposed framework offers remarkable flexibility and effectiveness, even in legacy systems with limited building information, making it a pragmatic and valuable solution for enhancing the energy efficiency and user comfort in pre-existing structures.
Paper Structure (13 sections, 4 equations, 17 figures, 2 tables, 1 algorithm)

This paper contains 13 sections, 4 equations, 17 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Review of the Experimental Related Works
  3. IoT framework
  4. Sensor nodes and gateways
  5. Central Server
  6. Actuator and testbed building
  7. Machine learning for thermal model
  8. MPC Formulation
  9. Results
  10. Temperature analyzing and occupants' comfort
  11. Energy consumption
  12. Performance of the AHU in the warm season
  13. Conclusion and Future Work

Figures (17)

  • Figure 1: The experimental setup involved collecting stream data at five-minute intervals using Raspberry Pis. The collected data were then transmitted to the central server via the MQTT protocol, which enabled the AHU system to make decisions based on the occupants/administrators given a setpoint.
  • Figure 2: The long-range sensor node was designed using a microcontroller and manufactured in this study.
  • Figure 3: Test environment overview (left) Blueprint of the building map indicating the locations of AHUs and rooms. (Right) Photograph of the building where the sunbeam shines on the southern side.
  • Figure 4: The AHU equipment diagram displays the direction of the hot and cold water flow during the heating and cooling processes using blue and red arrows. Manual valves are used to prevent hot water from entering the system when the cooling tower is active and vice versa.
  • Figure 5: Increasing (top) and decreasing (bottom) EDFs generated by ML and measured data for two random days between the previous year, mid-February, and mid-March.
  • ...and 12 more figures