Fault detection analysis using data mining techniques for a cluster of smart office buildings

https://doi.org/10.1016/j.eswa.2015.01.010Get rights and content

Highlights

  • An energy fault detection analysis was performed for a cluster of buildings.

  • Pattern recognition techniques coupled with outliers detection methods were used.

  • Anomalies are detected during early morning, lunch break, and end of working hours.

  • The methodology can be easily implemented in BEMS.

Abstract

There is an increasing need for automated fault detection tools in buildings. The total energy request in buildings can be significantly reduced by detecting abnormal consumption effectively. Numerous models are used to tackle this problem but either they are very complex and mostly applicable to components level, or they cannot be adopted for different buildings and equipment. In this study a simplified approach to automatically detect anomalies in building energy consumption based on actual recorded data of active electrical power for lighting and total active electrical power of a cluster of eight buildings is presented. The proposed methodology uses statistical pattern recognition techniques and artificial neural ensembling networks coupled with outliers detection methods for fault detection. The results show the usefulness of this data analysis approach in automatic fault detection by reducing the number of false anomalies. The method allows to identify patterns of faults occurring in a cluster of bindings; in this way the energy consumption can be further optimized also through the building management staff by informing occupants of their energy usage and educating them to be proactive in their energy consumption. Finally, in the context of smart buildings, the common detected outliers in the cluster of buildings demonstrate that the management of a smart district can be operated with the whole buildings cluster approach.

Section snippets

Introduction and background knowledge

Buildings are becoming more and more complex energy systems consisting of several elements i.e. heating/cooling systems, ventilation systems, lighting and control systems etc. In addition, buildings have multifarious activities and the occupants may have different demands from a building. Even though building ramification is growing, communication between the participants and the building elements during the building life is poor (Djuric & Novakovic, 2009). The building energy system and the

Buildings cluster and data description

The cluster of buildings is located within the Italian National Agency for New technologies, Energy and Sustainable Economic Development (ENEA) Casaccia research center and includes buildings F66–F67–F68–F69–F70–F71–F72–F73, positioned in two different blocks. A first block, consisting of three contiguous buildings is oriented along the axis NW–SE, while the second block consists of 5 buildings and its main orientation is NE–SW as shown in Fig. 1. The eight buildings have similar

Proposed methods

One of the effective ways of analyzing large data is to identify recurring patterns in the raw data. Clustering and classification are two common techniques used for finding hidden patterns in data sets. Discovering the patterns in data before applying the outliers detection algorithm is very useful to find anomalies in the building energy consumption. Outliers are cases that have data values very different from the data values for the majority of cases in the data set. Statistical-based (

Methodology

The proposed approach was applied to each building of the cluster for both winter and summer data. This section contains the application of each method described in Section 3.

For all simulations performed in this work, the values of the active electrical power for lighting and the total active electrical power of each building are considered as dependent variables. The independent variables considered are: date, day of the week, time of the day, average indoor temperature, average outdoor

Results and discussion

The results obtained from the analysis of fault detection conducted separately for each building, for the two dependent variables, and for the two time periods are summarized and presented in this section.

Conclusions

The research was aimed at testing the potential of using data mining techniques and ANN BEM in combination with outliers detection for an automated fault detection process. The application of the proposed approach can improve fault detection process in building application by reducing the number of false anomalies.

The methods proposed and implemented have proven adequate for detection purposes with different potentials and limitations. In particular the neural ensemble method has always proven

Glossary

ANN
artificial neural network
ANNE
artificial neural networks ensemble
BEM
Basic Ensemble Method
BEMS
building energy management system
CART
classification and regression tree
DBSCAN
Density-Based Spatial Clustering of Applications with Noise
FDD
fault detection and diagnosis
GESD
generalized extreme studentized deviate
h
user-specified constant in peak detection method
HVAC
heating, ventilation and air conditioning
k
number of right/left neighbors of xi
m
mean of all positive values of the peak function
minPts

References (59)

  • R. Razavi-Far et al.

    Efficient residuals pre-processing for diagnosing multi-class faults in a doubly fed induction generator, under missing data scenarios

    Expert Systems with Applications

    (2014)
  • F. Rossi et al.

    Artificial neural networks and physical modeling for determination of baseline consumption of CHP plants

    Expert Systems with Applications

    (2014)
  • J.E. Seem

    Pattern recognition algorithm for determining days of the week with similar energy consumption profiles

    Energy and Buildings

    (2005)
  • J.E. Seem

    Using intelligent data analysis to detect abnormal energy consumption in buildings

    Energy and Buildings

    (2007)
  • W. Shang et al.

    An intelligent fault diagnosis system for newly assembled transmission

    Expert Systems with Applications

    (2014)
  • S. Wang et al.

    AHU sensor fault diagnosis using principal component analysis method

    Energy and Buildings

    (2004)
  • S. Wu et al.

    Cross-level fault detection and diagnosis of building HVAC systems

    Building and Environment

    (2011)
  • Y. Zhao et al.

    Pattern recognition-based chillers fault detection method using support vector data description (SVDD)

    Applied Energy

    (2013)
  • O. Akin, & et al. (Eds.), (2010). Flow charts and data models for initial commissioning of advanced and low energy...
  • M.A. Arbib

    The handbook of brain theory and neural networks

    (1995)
  • A. Arning et al.

    A linear method for deviation detection in large databases

    KDD

    (1996)
  • D.G. Arseniev et al.

    Intelligent fault detection and diagnostics system on rule-based neural network approach

  • C.M. Bishop

    Neural Networks for Pattern Recognition

    (1995)
  • L. Breiman

    Combining predictors. Combining artificial neural nets: Ensemble and modular multi-net systems

    (1999)
  • V.J. Carey et al.

    Resistant and test-based outlier rejection: Effects on Gaussian one-and two-sample inference

    Technometrics

    (1997)
  • D.E. Claridge et al.

    Monitored commercial building energy data: Reporting the results

    Ashrae Transactions

    (1992)
  • Dodier, R. H., & Kreider, J. F. (1999). Detecting whole building energy problems (No. CONF-990102–). University of...
  • M. Ester et al.

    A density-based algorithm for discovering clusters in large spatial databases with noise

    KDD

    (1996)
  • R. Fontugne et al.

    Strip, bind, and search: A method for identifying abnormal energy consumption in buildings

  • Cited by (145)

    • An outlier management framework for building performance data and its application to the power consumption data of building energy systems in non-residential buildings

      2023, Journal of Building Engineering
      Citation Excerpt :

      The lack of statistical supports would limit their further application to data with larger scale. The building performance data is valuable for various research interests, such as data-driven fault detection, diagnosis, and estimation (FDD&E) [7,17–27], performance optimization, predictive modeling, building power consumption pattern recognition [8,13–16,23,28–33], etc. This section will focus to review the outlier management progress in the building power data.

    View all citing articles on Scopus
    View full text