Analytical investigation of autoencoder-based methods for unsupervised anomaly detection in building energy data

Cheng Fan, Fu Xiao, Yang Zhao, Jiayuan Wang

Research output: Journal article publicationJournal articleAcademic researchpeer-review

169 Citations (Scopus)


Practical building operations usually deviate from the designed building operational performance due to the wide existence of operating faults and improper control strategies. Great energy saving potential can be realized if inefficient or faulty operations are detected and amended in time. The vast amounts of building operational data collected by the Building Automation System have made it feasible to develop data-driven approaches to anomaly detection. Compared with supervised analytics, unsupervised anomaly detection is more practical in analyzing real-world building operational data, as anomaly labels are typically not available. Autoencoder is a very powerful method for the unsupervised learning of high-level data representations. Recent development in deep learning has endowed autoencoders with even greater capability in analyzing complex, high-dimensional and large-scale data. This study investigates the potential of autoencoders in detecting anomalies in building energy data. An autoencoder-based ensemble method is proposed while providing a comprehensive comparison on different autoencoder types and training schemes. Considering the unique learning mechanism of autoencoders, specific methods have been designed to evaluate the autoencoder performance. The research results can be used as foundation for building professionals to develop advanced tools for anomaly detection and performance benchmarking.

Original languageEnglish
Pages (from-to)1123-1135
Number of pages13
JournalApplied Energy
Publication statusPublished - 1 Feb 2018


  • Anomaly detection
  • Autoencoder
  • Building energy management
  • Building operational performance
  • Unsupervised data analytics

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Building and Construction
  • Energy(all)
  • Mechanical Engineering
  • Management, Monitoring, Policy and Law


Dive into the research topics of 'Analytical investigation of autoencoder-based methods for unsupervised anomaly detection in building energy data'. Together they form a unique fingerprint.

Cite this