Multivariate event detection methods for non-intrusive load monitoring in smart homes and residential buildings

doi:10.1016/j.enbuild.2019.109624

Energy and Buildings

Volume 208, 1 February 2020, 109624

https://doi.org/10.1016/j.enbuild.2019.109624 Get rights and content

Highlights

•
Peformance comparison of four detectors in their multivariate and univariate versions using specific detection metrics.
•
Setting up of a new Feature Selection Algorithm for Detection Purposes.
•
Proposal of a new dataset for Non Intrusive Load Monitoring event-based approaches containing labeled and time-stamped power time series of appliances turned on and off.
•
Application of the feature selection algorithm to a publicily available dataset and to our own dataset.
•
Multivariate time series help to improve the detection performances with electrical features selected by the proposed algorithm.

Abstract

Non Intrusive Load Monitoring (NILM) approaches refer to the analysis of the aggregated electrical signals of Home Electrical Appliances (HEAs) in order to identify their operating schedules. It has emerged as a promising solution to help residential consumers to reduce their electricity bills through a breakdown of energy consumption. NILM methods are either event-based or non event-based. This categorization depends on whether or not they rely on the detection of HEAs’ significant state transitions (e.g., On/Off or state change) in power consumption signals. This paper focuses on event-based approaches and especially in multivariate change detection algorithms. It aims at highlighting the benefits brought by a multivariate approach for change detection using the appropriate electrical features. We first suggest to extend four existing change detection algorithms in the multidimensional case. The studied detection algorithms are first detailed and compared to each other and to their existing scalar versions through numerical simulations. Then, a new feature selection algorithm for change detection is presented and assessed when combined with the most efficient detector among the four investigated ones. Finally, the feature selection method for detection purposes is applied to two different NILM case studies. The first one uses power features derived from the BLUED current and voltage measurements and the second one is based on current and voltage measurements acquired using our own acquisition system. Compared to the classical scalar approach, the results show that the multivariate approach brings a significant performance improvement when the features selected by the proposed algorithm are used.

Introduction

In the current context of energy transition, it is essential to better monitor the domestic consumption. Feedback on electricity consumption appears to be a major tool to save energy through more conscious user behaviour [1]. NILM has emerged as a promising solution to provide a breakdown of the residential energy consumption without instrumenting HEAs [2], [3], [4]. The appliances identification is made through the analysis of the current and voltage signals using a single measuring device connected to the house electrical panel [5]. This information is relayed back to the consumers, who can therefore make an informed decision about energy savings. Throughout literature, a large number of NILM approaches are reported [2], [6]. Most of the methods can be classified into supervised or unsupervised methods. The former rely on classification algorithms that require a sufficient amount of labeled data for a training process [6], [7]. On the opposite, the latter methods are based on clustering procedures which do not need a labeled training dataset [6]. Finally, NILM approaches can be categorized into non event-based and event-based approaches [8]. Non event-based methods disaggregate the power consumption using techniques such as the Hidden Markov models (HMM) [9], [10] which identify repetitive HEAs’ patterns in the load curve based on HEAs previous states. The event-based approaches assume that each event in the total household power consumption is a response to a HEA state change [8]. Fig. 1 shows an overview of event-based NILM approaches. In this research work, a specific interest is given to HEAs’ event detection which is an elementary step in the NILM pipeline [11]. According to the existing literature on detection algorithms for event-based NILM methods, two main approaches are considered: the heuristic and the probabilistic methods [12]. Algorithms under the first category analyze a time series of data, looking for changes above a given threshold. In [13], the events are detected in the total active power signal by computing the absolute value of the difference between two consecutive samples, which is then compared to a pre-defined threshold. M.N. Meziane et al. in [14], propose an event detector called High Accuracy NILM Detector (HAND) that tracks the standard deviation variation of the current signal envelope using a sliding window. A threshold separates the events characterized by high standard deviations from the steady states defined by much smaller standard deviations. They obtained a probability of detection of 96.7% on simulated data.

Algorithms in the second category of probabilistic methods comprise three entities: A stochastic process under observation, a change point at which the statistical properties of the process undergo a change, and a decision maker that observes and detects the change of the process statistical properties [15], [16]. In the NILM context, three algorithms are mainly used to detect changes in active power signals: the Generalized Likelihood Ratio (GLR) test, the Chi-squared Goodness Of-Fit (χ² GOF) and the CUSUM detectors. Authors in [17], [18] use the GLR approach to test if two consecutive time frames share a common distribution by deriving a decision function from the log-probability distribution ratio before and after a potential change of the mean value. In [18], [19], a χ² test statistic is used to assess if two neighboring windows share a common distribution. If this is not the case, an event is then detected. Researchers in [20], [21] apply a CUSUM algorithm for the detection of both the beginning and the end of a HEA transient-state active power signal. Zhu et al. [21] reached a probability of detection of 90% by applying their approach on real data comprising 200 events of eight different HEAs switched on and off.

The common thread in all these research works is that the detection is done in a univariate context by only considering the active power signal, whereas prior literature on NILM focuses on identifying an effective features’ set that defines a unique HEA signature [22], [23], [24]. Fig. 2 illustrates three power time series related to the same scenario of several HEAs switched on and off using our own acquisition system [25], and three power time series derived from the Building-Level fully-labeled Dataset for Electricity Disaggregation (BLUED) [26]. It can be observed that changes are present in the active power time series, the reactive power time series and their 4th and 5th order harmonic. The information provided by these time series could be used together to improve the detection robustness. To the best of our knowledge, there is no existing literature on multivariate change detectors designed for NILM event-based methods. Consequently, the goal of this paper is to demonstrate the benefit brought by using a multivariate approach for change detection with the most relevant features. Reference [27] presents our preliminary results related to the application of univariate change detectors on active power signals. In this paper, the pool of studied change detectors is enlarged and extended to the multivariate case by considering several power features such as the active power, the reactive power and their respective harmonics. Four algorithms suited for mean change detection are considered in both the univariate and the multivariate cases. These change detectors are the Bayesian Information Criterion (BIC), the Cumulative Sum (CUSUM), the Hotelling T² test (equivalent to the GLR statistic test for normal distributions) and an updated version of the Effective Residual algorithm described in [27].

This paper is organized as follows. In Section 2, the four change detectors are presented. Their performances are assessed through Monte Carlo experiments using detection performance evaluation metrics, considering both univariate and multivariate cases. In Section 3, the proposed Feature Selection Algorithm for Detection Purposes (FSADP) is introduced. Section 4 focuses on the experimental results when applying the FSADP on real-world power signals corresponding to a scenario of HEAs switched on and off. Firstly, power signals are obtained from the BLUED dataset current and voltage measurements [26], and secondly, power signals are computed from current and voltage measurements using our own acquisition system. The subsequent discussion addresses the selection of the power features that are best suited for the detection task. Finally, the main conclusions and contributions of the current work are provided in Section 5.

Section snippets

Hypothesis testing framework

In this subsection, the mathematical formulation of the detection problem is introduced. Let x_m ∈ lR^p be a multidimensional time series such as $x_{m} = {(x_{m, 1}, \dots, x_{m, j}, \dots, x_{m, p})}^{T},$ where x_m,j is the value of feature $j \in {1, \dots, p}$ at time instant m. Let $X_{n, p} = (x_{n_{a}}, x_{n_{a} + 1}, \dots, x_{m}, \dots, x_{n})$ with $m \in {n_{a}, \dots, n},$ be the $p \times (w + u)$ matrix (w being an even integer) of the last $n - n_{a} + 1 = w + u$ samples of x_m until the current time instant n. Since a HEA transient state is not necessarily an abrupt change, we use the “Window with Margins”

Motivations

In NILM approaches, the features are extracted attributes from current and voltage measurements. In Fig. 2, it can be well observed that an event induces a common change for different electrical features at different scales. In other words, some features might be informative and useful whereas others are not. FS is a process commonly used for classification tasks to reduce overfitting and to avoid the curse of dimensionality problem [39], [40]. The selected subset contains the minimal number of

Application to NILM

In this section, the Hotelling T² detector, which showed the best performances, is combined to the proposed FSADP and applied to real data streams in order to find the most relevant features for the detection task. We first introduce the power features computed using current and voltage measurements. Then we present the publicly available BLUED dataset dedicated to NILM event-based approaches and a new proposed dataset built using our acquisition system. Both datasets contain several events of

Conclusion

This paper explores and extends to the multivariate case four methods for detecting HEAs state changes in a NILM context. Furthermore, a new labeled dataset is introduced and made available for the sake of reproducible research. We also propose a new method for properly selecting power features that improve detection performances over the existing univariate methods based on active power signal only. Our study further supports prior NILM studies that also emphasized the need of an effective set

Declaration of Competing Interest

All authors of the submitted manuscript entitled “Multivariate Abrupt Change Detection for Non Intrusive Load Monitoring in Residential Buildings” confirm that they have contributed significantly for this research work, revised it critically and approved the manuscript. We also confirm that it has not been submitted earlier in any journal and is not being considered for publication elsewhere.

As scientists, we certify that the paper has been submitted with full responsibility and research

References (45)

A. Nilsson et al.
Smart homes, home energy management systems and real-time feedback: lessons for influencing household energy consumption from a swedish field study
Energy Build.
(2018)
L. Farinaccio et al.
Using pattern recognition approach to disaggregate the total electricity consumption in a house into the major end-uses
Energy Build.
(1999)
R. Bonfigli et al.
Non-intrusive load monitoring by using active and reactive power in additive factorial hidden markov models
Appl. Energy
(2017)
L.K. Norford et al.
Non-intrusive electrical load monitoring in commercial buildings based on steady-state and transient load-detection algorithms
Energy Build.
(1996)
C.C. Yang et al.
Comparative study of event detection methods for non-intrusive appliance load monitoring
Energy Procedia
(2014)
L. De Baets et al.
On the bayesian optimization and robustness of event detection methods in NILM
Energy Build.
(2017)
Z. Zhu et al.
A novel CUSUM-based approach for event detection in smart metering
IOP Conf. Series: Mater. Sci. Eng.
(2018)
N. Sadeghianpourhamami et al.
Comprehensive feature selection for appliance classification in NILM
Energy Build.
(2017)
S.M. Kay
Fundamentals of Statistical Signal Processing: Estimation Theory
(1997)
M. Rong et al.
Feature selection and its use in big data: challenges, methods, and trends
IEEE Access
(2019)

Z. Zheng et al.

A supervised event-based non-intrusive load monitoring for non-linear appliances

Sustainability

(2018)

L. Pereira et al.

Performance evaluation in non-intrusive load monitoring: datasets, metrics, and tools: a review

Wiley Interdiscip. Rev.: Data Min. Knowl. Discov.

(2018)

S. Henriet et al.

A generative model for non-intrusive load monitoring in commercial buildings

Energy Build.

(2018)

D. Fischer et al.

Model for electric load profiles with high time resolution for german households

Energy Build.

(2015)

A. Ruano et al.

NILM Techniques for intelligent home energy management and ambient assisted living: a review

Energies

(2019)

C. Nalmpantis et al.

Machine learning approaches for non-intrusive load monitoring: from qualitative to quantitative comparation

Artif. Intell. Rev.

(2019)

W. Dan et al.

Review of non-intrusive load appliance monitoring

Ieee 3^rd advanced information technology, electronic and automation control conf. (iaeac)

(2018)

K. Basu et al.

Time series distance-based methods for non-intrusive load monitoring in residential buildings

Energy Build.

(2015)

K.D. Anderson et al.

Event detection for non intrusive load monitoring

Iecon 2012 - 38th annual conf. on ieee industrial electronics society

(2012)

M. Weiss et al.

Leveraging smart meter data to recognize home appliances

Ieee int. conf. on pervasive computing and communications

(2012)

M. Nait Meziane et al.

High accuracy event detection for non-intrusive load monitoring

Ieee int. conf. on acoustics, speech and signal processing (icassp)

(2017)

M. Basseville et al.

Detection of Abrupt Changes: Theory and Application

(1993)

Cited by (31)

Dynamic adaptive event detection strategy based on power change-point weighting model
2024, Applied Energy
Event detection is a prerequisite and key component of NILM (Non-Intrusive Load Monitoring) by monitoring transient changes in residential loads to discern whether a transient event has occurred in an appliance. However, the event detection performance of existing algorithms is affected by the operating environment, and it isn't easy to maintain high accuracy. For this reason, this paper proposes an adaptive event detection method based on the PCW (power change-point weights) model. Specifically, the DACUSUM (Dynamic Adaptive Cumulative Sum) algorithm with dynamic updating of parameters is first proposed, which effectively avoids the miss and false detection of CUSUM in the process of event detection. Secondly, the PCW model is proposed, which is capable of evaluating the effect of event detection of thresholds through the transient information entropy without prior knowledge. Lastly, based on the DACUSUM and PCW model, the threshold-adaptive event detection method is proposed, which takes the transient information entropy as the objective function and utilizes the genetic algorithm to dynamically adjust the thresholds to improve the performance of event detection under different operating environments. Taking eight typical appliances as an example, on the one hand, the proposed DACUSUM reduces the leakage and false detection phenomena compared with CUSUM and improves the event detection performance. On the other hand, the PCW model-based event detection strategy doesn't need human intervention or prior knowledge and is adaptable to different operating environments. The experimental results show that the proposed strategy achieves F1 scores of over 90% for the event detection of eight types of home appliances.
A low complexity binary-weighted energy disaggregation framework for residential electricity consumption
2023, Energy and Buildings
The discipline of Non-Intrusive Load Monitoring (NILM) has witnessed a surge in the application of machine learning and pattern recognition approaches, enabling researchers to investigate NILM problems. This paper introduces a novel energy disaggregation system that employs a binary-weight matrix to separate the power consumption into distinct signal patterns. The framework includes filtering techniques, followed by event detection and energy disaggregation. To address the challenges of event detection, a regional threshold-based algorithm is developed, eliminating the need for predefined thresholds. A comprehensive complexity analysis of the developed algorithms reveals a reduced computational complexity, making the framework suitable for real-time deployment. For performance assessment, the Reference Energy Disaggregation dataset (REDD) and Energy Monitoring via Building Electricity Disaggregation dataset (EMBED) are utilized. A frequency of 1 Hz is maintained to ensure accurate evaluation. The proposed event detection algorithm achieves a precision of 92.5% and an f1-score of 73.6% on EMBED data, improving average precision by 23.5% and a substantial reduction in false positives compared to an existing method. The energy disaggregation algorithm separates the power consumption of four devices in 15.6 s and the entire framework (filtering, event detection, and energy disaggregation) takes 15.9 s for execution, all achieved using a semi-supervised training-less approach.
Non-intrusive condition monitoring based on event detection and functional data clustering
2023, Engineering Applications of Artificial Intelligence
Implementing monitoring electricity consumption strategies in industrial environments provides improvements in both the maintenance process and energy efficiency. The contribution of this work is an industry-oriented non-intrusive load monitoring approach based on an unsupervised algorithm, encompassing a method of event detection, functional data clustering, and condition monitoring. With this method, multiple devices can be monitored by only one electric meter in industrial environments, enhancing the early detection of anomalies and energy inefficiencies. The proposed approach presents a robust event detection to deal with different industrial contexts due to its intuitive parameters, which allows adapting the detection to be more sensitive or to filter out higher noise variations. Unlike other feature-based clusterings, the proposed functional data clustering enables high-precision identification of transient state patterns, characterizing specific shapes for each pattern to properly cluster them, and provides higher reliability during load detection. Thus, this load detection segments the power consumption to extract transient states and identify which load acts on each event based on functional data clusters. By detecting when loads start to consume, the proposed energy disaggregation extracts the frequency spectrum of each device from the aggregate current consumption, which is used in a condition monitoring strategy to track the spectrum behavior of each device. In this way, the condition of multiple loads can be monitored using a single electric meter, whose information can be relevant to accurately schedule maintenance interventions and detect anomalies or inefficiencies early. The proposed approach was validated in three industrial contexts: The load detection accuracy was verified in an industrial testbed, giving a precision higher than 99% for monitoring five devices. The second and third industrial scenarios validate the accuracy of the proposed condition monitoring method. The last scenario was carried out at Bilbao airport to track the condition of multiple conveyor belts of a baggage handling system located at check-in. As a result, the degradation trend of three sets of conveyor belts was monitored.
Robust event detection for residential load disaggregation
2023, Applied Energy
Nonintrusive load monitoring (NILM) can facilate the transition to energy-efficient and low-carbon buildings. Event detection is the first and most critical step in event-based NILM and can improve the efficiency and performance of NILM by decreasing inference times to the number of events with transient features extracted from events. However, existing event detection methods with fixed parameters may fail to achieve high accuracy in the case of uncertain and complicated residential load changes such as high fluctuation, long transition, and near simultaneity in both power and time dimensions. Besides, it is difficult to transfer the fixed parameter to new households with different load profiles. Furthermore, most of these methods prove that they are able to detect events but not able to extract features for load disaggregation. This paper proposes a robust event detection method with adaptive parameters to deal with such issues. Specifically, a window with adaptive margins, multi-window screening, and adaptive threshold method is proposed to detect events in aggregated load data with high sampling rate (>1 Hz). The proposed method captures the transitions by adaptively tuning parameters including window width, margin width, and thresholds. It can also achieve good performance with blind parameter setting so that it is suitable for unknown households or datasets. Furthermore, it captures complete transitions that are indispensable for transient feature extraction. Case studies on a 20 Hz dataset, the 50 Hz LIFTED dataset, and the 60 Hz BLUED dataset show that the proposed method can robustly outperform other state-of-the-art event detection methods. The robust performance of the proposed method is also verified by a cross validation of parameters among different datasets. Lastly, the proposed event detection method is demonstrated to have the merits of improving the performance of load disaggregation.
Electric energy disaggregation via non-intrusive load monitoring: A state-of-the-art systematic review
2022, Electric Power Systems Research
Citation Excerpt :
The event detection module of NILM looks for events in the aggregate signal using either an expert heuristics model, a probabilistic model, or a matched filter model [34]. Few examples of these models include a change in amplitude, cumulative sum, generalized likelihood ratio, goodness-of-fit [35], Dynamic Time Warping, clustering [36], multivariate event detection [37], and maximum and minimum points based [38] event detectors. Following confirmation of an event, the signature extraction module comes into action.
Appliance energy consumption tracking in a building is one of the vital enablers of energy and cost saving. An economical and viable solution would be to estimate individual appliance consumption from a single-point measurement without using dedicated appliance-wise sensors or sub-meters. Such solutions are called non-intrusive load monitoring or energy disaggregation. Though the technology was introduced around three decades ago, the recent deployments of smart meters and machine learning-based algorithms attract the researcher to develop many novel solutions to the technology. To this end, this paper aims to provide a comprehensive, unbiased, and systematic state-of-the-art review of the energy disaggregation technology by evaluating two research questions. These are, first, how did the research community approach to solve the NILM problem and how has the technology evolved, and second, how do various NILM solutions stand compared to others in terms of a few key attributes. Further, some key challenges and future research directions of the technology are highlighted.
Incorporating coincidental water data into non-intrusive load monitoring
2022, Sustainable Energy, Grids and Networks
Citation Excerpt :
A fundamental step in any event-based NILM method is detecting the events accurately, where an event in general refers to a mode transition of any appliance. Due to the presence of noise and fluctuations in the power signal, it is essential to establish a reliable event detection procedure [39]. In a majority of research articles on event-based NILM, an event is detected based on the difference of two consecutive samples considering a predefined threshold.
Non-intrusive load monitoring (NILM) is among successful approaches aiding residential energy management. However, the presence of multi-mode appliances and appliances with close power values and lack of a proper volume of training dataset have remained influential in worsening the computational complexity and diminishing the accuracy of classification-based NILM algorithms. To tackle these challenges, we propose a novel classification process, which considers the correlation of water and electricity consumption of some appliances as a novel signature in the network to improve the accuracy of disaggregation process in overlapping modes and tackle the lack of proper volume of training dataset. In the first phase of the proposed method, the $K$ -nearest neighbors method, as a fast classification technique, is employed to extract power signals of appliances with exclusive non-close power values. Then, two different deep learning-based methods are proposed to disaggregate the consumption of appliances with close consumption values considering the correlation of electricity and water consumption of some appliances. Throughout these methods the water consumption of these appliances are also disaggregated. The main objectives of the proposed methods are increasing accuracy in the close modes of power consumption, and informing consumers about the water consumption pattern of some appliances. To illustrate the proposed processes and validate its effectiveness, the Almanac Minutely Power Dataset as a real dataset with a sampling rate of 1-minute is considered. The numerical results show marked improvement with respect to the existing classification-based NILM techniques. Moreover, it shows the applicability of proposed methods in dealing with low-frequency readings carried out by existing smart meters.

View all citing articles on Scopus

View full text

Multivariate event detection methods for non-intrusive load monitoring in smart homes and residential buildings

Highlights

Abstract

Introduction

Section snippets

Hypothesis testing framework

Motivations

Application to NILM

Conclusion

Declaration of Competing Interest

Energy Build.

Energy Build.

Appl. Energy

Energy Build.

Energy Procedia

Energy Build.

IOP Conf. Series: Mater. Sci. Eng.

Energy Build.

IEEE Access

Sustainability

Performance evaluation in non-intrusive load monitoring: datasets, metrics, and tools: a review

Wiley Interdiscip. Rev.: Data Min. Knowl. Discov.

A generative model for non-intrusive load monitoring in commercial buildings

Energy Build.

Model for electric load profiles with high time resolution for german households

Energy Build.

NILM Techniques for intelligent home energy management and ambient assisted living: a review

Energies

Machine learning approaches for non-intrusive load monitoring: from qualitative to quantitative comparation

Artif. Intell. Rev.

Review of non-intrusive load appliance monitoring

Ieee 3rd advanced information technology, electronic and automation control conf. (iaeac)

Time series distance-based methods for non-intrusive load monitoring in residential buildings

Energy Build.

Event detection for non intrusive load monitoring

Iecon 2012 - 38th annual conf. on ieee industrial electronics society

Leveraging smart meter data to recognize home appliances

Ieee int. conf. on pervasive computing and communications

High accuracy event detection for non-intrusive load monitoring

Ieee int. conf. on acoustics, speech and signal processing (icassp)

Detection of Abrupt Changes: Theory and Application

Ieee 3^rd advanced information technology, electronic and automation control conf. (iaeac)