Detecting and Diagnosing Incipient Building Faults Using Uncertainty Information from Deep Neural Networks

Abstract

Early detection of incipient faults is of vital importance to reducing maintenance costs, saving energy, and enhancing occupant comfort in buildings. Popular supervised learning models such as deep neural networks are considered promising due to their ability to directly learn from labeled fault data; however, it is known that the performance of supervised learning approaches highly relies on the availability and quality of labeled training data. In Fault Detection and Diagnosis (FDD) applications, the lack of labeled incipient fault data has posed a major challenge to applying these supervised learning techniques to commercial buildings. To overcome this challenge, this paper proposes using Monte Carlo dropout (MC-dropout) to enhance the supervised learning pipeline, so that the resulting neural network is able to detect and diagnose unseen incipient fault examples. We also examine the proposed MC-dropout method on the RP-1043 dataset to demonstrate its effectiveness in indicating the most likely incipient fault types.

Consistency Index-Based Sensor Fault Detection System for Nuclear Power Plant Emergency Situations Using an LSTM Network

“…If physically redundant sensors are not installed, analytical redundancy approaches are applied that estimate sensor error from the relation between a range of different target sensors. These approaches can be classified into model-based [9,10], knowledge-based [11][12][13], and data-driven methods [14][15][16][17] depending on the understanding of the depth of the system and the available amount of data. Jiang et al (2011) demonstrated this classification of general sensor fault detection [18].…”

An Encoder-Decoder Based Approach for Anomaly Detection with Application in Additive Manufacturing

“…A review of the literature reveals that data-driven approaches relying on supervised learning have demonstrated promising results in various applications, e.g. Fault Detection and Diagnosis (FDD) in air conditioning systems [9], [13], [14].…”

“…This is a limitation of supervised methods because it is almost impossible to obtain every possible type of anomaly that could happen on a system. To address this problem, Jin et al recently proposed a FDD method that uses Monte-Carlo dropout [9] to estimate the prediction uncertainty of deep neural networks. The method was applied to the identification of incipient faults that are not represented in the training data that only consists of labeled data of normal and severe faults.…”

Incipient fault detection of sensors used in wastewater treatment plants based on deep dropout neural network

“…Fault detection and diagnosis (FDD) in general is classified into three categories: model based, signal based and data based [13]. The first and the second categories require system experts and at the same time is costly because of which these methods become inconvenient for complex systems.…”

“…The deep neural network is one of the machine learning techniques which is reliable and effective in indicating incipient faults [13]. It is a supervised learning method which requires labelled data for training.…”