skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Mining Building Energy Management System Data Using Fuzzy Anomaly Detection and Linguistic Descriptions

Abstract

Building Energy Management Systems (BEMSs) are essential components of modern buildings that utilize digital control technologies to minimize energy consumption while maintaining high levels of occupant comfort. However, BEMSs can only achieve these energy savings when properly tuned and controlled. Since indoor environment is dependent on uncertain criteria such as weather, occupancy, and thermal state, performance of BEMS can be sub-optimal at times. Unfortunately, the complexity of BEMS control mechanism, the large amount of data available and inter-relations between the data can make identifying these sub-optimal behaviors difficult. This paper proposes a novel Fuzzy Anomaly Detection and Linguistic Description (Fuzzy-ADLD) based method for improving the understandability of BEMS behavior for improved state-awareness. The presented method is composed of two main parts: 1) detection of anomalous BEMS behavior and 2) linguistic representation of BEMS behavior. The first part utilizes modified nearest neighbor clustering algorithm and fuzzy logic rule extraction technique to build a model of normal BEMS behavior. The second part of the presented method computes the most relevant linguistic description of the identified anomalies. The presented Fuzzy-ADLD method was applied to real-world BEMS system and compared against a traditional alarm based BEMS. In six different scenarios, the Fuzzy-ADLD method identifiedmore » anomalous behavior either as fast as or faster (an hour or more), that the alarm based BEMS. In addition, the Fuzzy-ADLD method identified cases that were missed by the alarm based system, demonstrating potential for increased state-awareness of abnormal building behavior.« less

Authors:
; ; ;
Publication Date:
Research Org.:
Idaho National Laboratory (INL)
Sponsoring Org.:
USDOE
OSTI Identifier:
1156912
Report Number(s):
INL/JOU-14-31800
DOE Contract Number:
DE-AC07-05ID14517
Resource Type:
Journal Article
Resource Relation:
Journal Name: Transactions on Industrial Informatics; Journal Volume: 10; Journal Issue: 3
Country of Publication:
United States
Language:
English
Subject:
32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATION; 42 ENGINEERING; 47 OTHER INSTRUMENTATION; Anomaly Detection, Building Energy Management Sys

Citation Formats

Dumidu Wijayasekara, Ondrej Linda, Milos Manic, and Craig Rieger. Mining Building Energy Management System Data Using Fuzzy Anomaly Detection and Linguistic Descriptions. United States: N. p., 2014. Web. doi:10.1109/TII.2014.2328291.
Dumidu Wijayasekara, Ondrej Linda, Milos Manic, & Craig Rieger. Mining Building Energy Management System Data Using Fuzzy Anomaly Detection and Linguistic Descriptions. United States. doi:10.1109/TII.2014.2328291.
Dumidu Wijayasekara, Ondrej Linda, Milos Manic, and Craig Rieger. Fri . "Mining Building Energy Management System Data Using Fuzzy Anomaly Detection and Linguistic Descriptions". United States. doi:10.1109/TII.2014.2328291.
@article{osti_1156912,
title = {Mining Building Energy Management System Data Using Fuzzy Anomaly Detection and Linguistic Descriptions},
author = {Dumidu Wijayasekara and Ondrej Linda and Milos Manic and Craig Rieger},
abstractNote = {Building Energy Management Systems (BEMSs) are essential components of modern buildings that utilize digital control technologies to minimize energy consumption while maintaining high levels of occupant comfort. However, BEMSs can only achieve these energy savings when properly tuned and controlled. Since indoor environment is dependent on uncertain criteria such as weather, occupancy, and thermal state, performance of BEMS can be sub-optimal at times. Unfortunately, the complexity of BEMS control mechanism, the large amount of data available and inter-relations between the data can make identifying these sub-optimal behaviors difficult. This paper proposes a novel Fuzzy Anomaly Detection and Linguistic Description (Fuzzy-ADLD) based method for improving the understandability of BEMS behavior for improved state-awareness. The presented method is composed of two main parts: 1) detection of anomalous BEMS behavior and 2) linguistic representation of BEMS behavior. The first part utilizes modified nearest neighbor clustering algorithm and fuzzy logic rule extraction technique to build a model of normal BEMS behavior. The second part of the presented method computes the most relevant linguistic description of the identified anomalies. The presented Fuzzy-ADLD method was applied to real-world BEMS system and compared against a traditional alarm based BEMS. In six different scenarios, the Fuzzy-ADLD method identified anomalous behavior either as fast as or faster (an hour or more), that the alarm based BEMS. In addition, the Fuzzy-ADLD method identified cases that were missed by the alarm based system, demonstrating potential for increased state-awareness of abnormal building behavior.},
doi = {10.1109/TII.2014.2328291},
journal = {Transactions on Industrial Informatics},
number = 3,
volume = 10,
place = {United States},
year = {Fri Aug 01 00:00:00 EDT 2014},
month = {Fri Aug 01 00:00:00 EDT 2014}
}
  • Significant portion of world energy production is consumed by building Heating, Ventilation and Air Conditioning (HVAC) units. Thus along with occupant comfort, energy efficiency is also an important factor in HVAC control. Modern buildings use advanced Multiple Input Multiple Output (MIMO) control schemes to realize these goals. However, since the performance of HVAC units is dependent on many criteria including uncertainties in weather, number of occupants, and thermal state, the performance of current state of the art systems are sub-optimal. Furthermore, because of the large number of sensors in buildings, and the high frequency of data collection, large amount ofmore » information is available. Therefore, important behavior of buildings that compromise energy efficiency or occupant comfort is difficult to identify. This paper presents an easy to use and understandable framework for identifying such behavior. The presented framework uses human understandable knowledge-base to extract important behavior of buildings and present it to users via a graphical user interface. The presented framework was tested on a building in the Pacific Northwest and was shown to be able to identify important behavior that relates to energy efficiency and occupant comfort.« less
  • EMS-derived data may offer important advantages, in terms of cost and effort vs data quality, as compared with conventional approaches now being used to analyze end-use performance in large buildings. Of course, using EMS's for performance monitoring will not be equally promising in all commercial buildings nor for all data needs. There must be a computerized energy management system already in place, with appropriate sensing, data storage, and communication capabilities. However, these systems are already becoming standard equipment, especially in buildings larger than 100,000 square feet. It is possible to take advantage of the data available from even a simplemore » EMS to determine some operating characteristics of a commercial building, especially to understand the way the electricity is used. Hourly EMS data help to indicate the weather dependency of the energy use in the building and can be used as constraint in load profile estimations where they are invaluable in establishing an accurate end-use breakdown.« less
  • Resilient control systems in critical infrastructures require increased cyber-security and state-awareness. One of the necessary conditions for achieving the desired high level of resiliency is timely reporting and understanding of the status and behavioral trends of the control system. This paper describes the design and development of a neural-network based data-fusion system for increased state-awareness of resilient control systems. The proposed system consists of a dedicated data-fusion engine for each component of the control system. Each data-fusion engine implements three-layered alarm system consisting of: (1) conventional threshold-based alarms, (2) anomalous behavior detector using self-organizing maps, and (3) prediction error basedmore » alarms using neural network based signal forecasting. The proposed system was integrated with a model of the Idaho National Laboratory Hytest facility, which is a testing facility for hybrid energy systems. Experimental results demonstrate that the implemented data fusion system provides timely plant performance monitoring and cyber-state reporting.« less
  • Lignite quality parameters have had central importance for power plants. This article addresses a comparative study on fuzzy and regression modeling for estimating the calorific value of lignite, which is one of the quality parameters from the other parameters: moisture, ash, volatile matter, and sulphur content. For the estimations, data driven models were designed based on linguistic fuzzy modeling structures. In addition, estimations of the fuzzy models were compared with linear regression estimations. The great majority of performance evaluations showed that the fuzzy estimations are very satisfactory in estimating calorific value of lignite.
  • Anomaly Detection; Building Energy Management Systems; Computational Intelligence