Predicting HVAC Energy Consumption in Commercial Buildings Using Multiagent Systems

Nan Li, Jun-young Kwak, Burçin Becerik-Gerber, Milind TambeProceedings of the ... ISARC 2013 · DOI

TL;DR

This paper introduces a novel multiagent system (MAS) model for predicting HVAC energy consumption in commercial buildings. The model achieves high accuracy (up to 98.2%) in predicting total energy use, with ventilation energy being particularly well-predicted, while cooling energy presents a greater challenge.

What Problem Does This Paper Address?

Commercial buildings are significant energy consumers, and understanding their future energy needs is crucial for effective energy management, conservation, and cost optimization. Traditional methods for predicting energy consumption often lack the temporal and spatial resolution needed for detailed building management, or they require extensive and complex input data. This paper aims to address this gap by developing a more granular and customizable prediction model using a multiagent systems approach.

Methods

The core of the proposed method is a multiagent system (MAS) framework called SAVES (Sustainable multiAgent systems for optimizing Variable objectives including Energy and Satisfaction). SAVES simulates the energy behaviors of HVAC systems within commercial buildings by modeling interactions between "room agents" (managing energy consumption within a room) and "proxy agents" (representing occupants). This MAS framework is integrated with a simulation model that calculates HVAC energy consumption based on physical and mechanical characteristics of the building and its systems. The prediction is performed at daily, weekly, and monthly intervals.

Datasets

The study utilizes ground truth energy consumption data collected from a real test-bed office building over 267 consecutive days. This data was used for both training and validation of the proposed model.

Key Insights and Ideas

  • MAS for Granular Prediction: The use of a multiagent system allows for a more nuanced simulation of building energy dynamics, moving beyond aggregate building-level predictions to potentially finer resolutions.
  • Decomposition Reveals Challenges: Breaking down HVAC energy consumption into cooling and ventilation components highlights that while ventilation is highly predictable, cooling energy consumption is the primary source of error and variation in the overall prediction.
  • Importance of Environmental Factors: The paper suggests that the variability in actual cooling energy usage, driven by factors like outside air temperature and humidity, needs to be better captured by the model.

Key Results

The proposed MAS model demonstrated strong performance in predicting total HVAC energy consumption, achieving accuracies between 92.6% and 98.2%. The Coefficient of Variation of the Root Mean Square Error (CV-RMSE) ranged from 7.8% for monthly predictions to 22.2% for daily predictions. Ventilation energy consumption was predicted with very high accuracy (over 99%) and low variation (CV-RMSE of 3.1% to 16.3%). Cooling energy consumption, however, showed higher inaccuracies and variations.

Limitations and Future Work

The authors acknowledge that the prediction of cooling energy consumption needs improvement. They plan to address this by:

  • Integrating actual daily temperature data and other weather variables (humidity, wind speed) into the simulation.
  • Installing meters to measure HVAC energy consumption in different areas of the building to enable validation at finer spatial resolutions.
  • Adding heating energy consumption to the prediction model.
  • Applying the improved model to evaluate the energy implications of various facility management policies.

Summary updated

Abstract

Predicting HVAC Energy Consumption in Commercial Buildings Using Multiagent Systems Nan Li, Jun-young Kwak, Burcin Becerik-Gerber, Milind Tambe Pages 987-994 (2013 Proceedings of the 30th ISARC, Montréal, Canada, ISBN 978-1-62993-294-1, ISSN 2413-5844) Abstract: Energy consumption in commercial buildings has been increasing rapidly in the past decade. The knowledge of future energy consumption can bring significant value to commercial building energy management. For example, prediction of energy consumption decomposition helps analyze the energy consumption patterns and efficiencies as well as waste, and identify the prime targets for energy conservation. Moreover, prediction of temporal energy consumption enables building managers to plan out the energy usage over time, shift energy usage to off-peak periods, and make more effective energy purchase plans. This paper proposes a novel model for predicting heating, ventilation and air conditioning (HVAC) energy consumption in commercial buildings. The model simulates energy behaviors of HVAC systems in commercial buildings, and interacts with a multiagent systems (MAS) based framework for energy consumption prediction. Prediction is done on a daily, weekly and monthly basis. Ground truth energy consumption data is collected from a test bed office building over 267 consecutive days, and is compared to predicted energy consumption for the same period. Results show that the prediction can match 92.6 to 98.2% of total HVAC energy consumption with coefficient of variation of the root mean square error (CV-RMSE) values of 7.8 to 22.2%. Ventilation energy consumption can be predicted at high accuracies (over 99%) and low variations (CV-RMSE values of 3.1 to 16.3%), while cooling energy consumption accounts for majority of inaccuracies and variations in total energy consumption prediction. Keywords: Heating, ventilation, and air conditioning (HVAC), Energy consumption prediction, Commercial building, Multiagent systems (MAS) DOI: https://doi.org/10.22260/ISARC2013/0108 Download fulltext Download BibTex Download Endnote (RIS) TeX Import to Mendeley
BibTeX
@inproceedings{li2013predicting,
  author = {Nan Li and Jun-young Kwak and Burçin Becerik-Gerber and Milind Tambe},
  title = {{Predicting HVAC Energy Consumption in Commercial Buildings Using Multiagent Systems}},
  booktitle = {Proceedings of the ... ISARC},
  year = {2013},
  doi = {10.22260/isarc2013/0108},
}
Figure