Skip to content

2020

  • in Thesis
  • 15 min read

Detection and diagnosis of anomalous energy consumption patterns in buildings through a data analytics based approach: the case of Politecnico di Torino.

Abstract

In recent years, Smart Metering Infrastructure (SMI) has enabled the collection of huge amounts of building-related data. However, very often, only time series of a few aggregated variables associated with building energy consumption are available. Therefore, it becomes necessary to extract from meter level data as much information as possible in order to optimize building energy management, by reducing losses due to inefficiencies or anomalous behaviour of sub-systems and equipment. This paper proposes an innovative top-down Fault Detection and Diagnosis (FDD) methodology able to automatically detect at whole build- ing meter-level anomalous energy consumption and then diagnosticate which sub-load could be responsible. The process consists of a multi-step procedure combining various data mining techniques. An evolutionary classification tree is firstly implemented to discover frequent and infrequent daily aggregated energy patterns opportunely abstracted through a symbolic approximation pro- cess. Then a post-mining analysis based on Association Rule Mining (ARM) is performed to discover the main sub-loads affecting the detected anomalous energy patterns. The methodology is tested on metering data related to the electrical load of a transformer substation of a university campus, leading to the development of a tool useful to support the energy management with a complete characterization and diagnosis of energy demand at a daily scale.

References

  • Maria Teresa Zitelli Application of Data Analytics techniques for the analysis of building energy performance during operation : the case of Politecnico di Torino. Rel. Alfonso Capozzoli, Marco Savino Piscitelli. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Energetica E Nucleare, 2022

A data analytics-based tool for the detection and diagnosis of anomalous daily energy patterns in buildings

Abstract

In this paper, a tool for the detection and diagnosis of anomalous electrical daily energy patterns relative to a transformer substation of a university campus was developed and tested. Through an innovative pattern recognition analysis consisting in a multi-step clustering process, six clusters of anomalous daily load profiles were identified and isolated in two-year historical data of total electrical energy consumption. The infrequent electrical load profiles were found to be strongly affected, in terms of both shape and magnitude, by the energy consumption behaviour related to the heating/cooling mechanical room. Then, a fault-free predictive model, which uses artificial neural network (ANN) in combination with a Regression Tree, was developed to detect anomalous trends of the electrical energy consumption. The model was able to detect the 93.7% of the anomalous profiles and only the 5% of fault-free days were wrongly predicted as anomalous. Eventually, a diagnosis phase was conceived and validated with a testing data set. A number of daily abnormal load profiles were detected and compared with the centroids of the anomalous clusters identified in the pattern-recognition stage. The work led to the development of a flexible intelligent tool useful for operating a continuous commissioning of the campus facilities.

References

  • Piscitelli, M. S., Brandi, S., Capozzoli, A., & Xiao, F. (2021, February). A data analytics-based tool for the detection and diagnosis of anomalous daily energy patterns in buildings. In Building Simulation (Vol. 14, pp. 131-147). Tsinghua University Press.