Residential Smart Meters Research Articles

Unknown appliances detection for non-intrusive load monitoring based on vision transformer with an additional detection head

Non-intrusive load monitoring (NILM) offers precise insights into equipment-level energy consumption by analyzing current and voltage data from residential smart meters, thus emerging as a potential strategy for demand-side management in power systems. However, a prevalent limitation in current NILM techniques is the presupposition of a known inventory of household appliances, an assumption that often becomes impractical due to the regular introduction of new appliances by consumers. To address this challenge, our approach integrates a vision transformer network with an additional detection head (ViTD), utilizing V-I trajectory images. Initially, the ViT model is trained to classify known appliances. Subsequently, an additional detection head is incorporated to manipulate the embedded features, encouraging the formation of distinct, compact class centers for the known appliance categories. During testing, samples are identified as either known or unknown appliances based on their proximity to these class centers. We utilize two public datasets, PLAID and WHITED, to demonstrate the effectiveness and superiority of our proposed method.

Distribution Transformer Remaining Useful Life Estimation Considering Electric Vehicle Penetration

A major portion of a power system's asset portfolio comprises distribution transformers on residential premises. The rapid and massive acceptance of electric vehicles is posing challenges for distribution transformers to operate over their expected lifespan. This work proposes a four-layer framework to assess the real-time and anticipated aging of a distribution transformer and estimate the remaining useful life of a distribution transformer. The first layer stores residential smart meter data to be utilized for the kVA load estimation of a distribution transformer in the second layer. The performance of two powerful forecasting tools, i.e., Time Series Decomposition and Hidden Markov Model, is compared in the third layer. The historical and forecast data, along with the distribution transformer's thermal parameters, are used for its remaining useful life assessment. Numerical validation is conducted on real-world data utilizing electricity consumption and ambient temperature of fifteen households in London, Ontario, Canada. This work also includes the penetration of the most popular electric vehicles in Canada, along with service drop cable data and practical secondary distribution circuit configuration.

Stochastic generation of residential load profiles with realistic variability based on wavelet-decomposed smart meter data

Residential smart meter data with high time resolution are integral to many data-driven applications, ranging from hosting capacity studies to R&D activities of private enterprises. However, privacy legislation restricts public availability of large-scale datasets. Furthermore, existing datasets may suffer from imbalances in terms of underrepresented classes. To address these concerns, this study presents a novel decomposition–recombination approach for generating synthetic load profiles that exhibit realistic variability and demand peaks. High-frequency load profiles are decomposed into a low-frequency base load and high-frequency variability at the daily level through a discrete wavelet transformation. Components from different households are subsequently rescaled, shifted and recombined in a stochastic load profile generator to obtain new daily load profiles with high-fidelity behavior. The performance of this generator is evaluated through benchmarking, resulting in a mean average error of 0.09 kW on an average value of less than 3 kW for the daily peaks, whilst preserving their seasonality. The introduced load profile generator is validated as an alternative to privacy-sensitive residential smart meter data in a hosting capacity case study. The analysis focuses on the voltage drop caused by residential electric vehicle charging, considering both real and synthetic data. The synthetic data demonstrated voltage drops with a mean average error less than 0.2 V for the 10th and 90th percentile when benchmarked with respect to the real voltage level distribution. The introduced decomposition–recombination method is shown to accurately capture the high-frequency variability and peak behavior, and is suitable for practical applications at the daily level.

Exploring regression models to enable monitoring capability of local energy communities for self‐management in low‐voltage distribution networks

This study proposes a data-driven approach to enable the self-management capability of local energy communities (LECs) via transformer congestion monitoring in low-voltage distribution networks. A set of regression models is adopted in this approach, while the data from residential smart meters (SMs) is leveraged. Four machine learning algorithms, namely ridge regression, support vector regression, random forest regression (RFR) and eXtreme gradient boosting regression (XGBR), are compared to select the best-performing regression model using a cross-validation method. A comprehensive framework is provided to facilitate comparison of the algorithm, consisting of data pre-processing, model fitting and validation, and model deployment. A thorough analysis is also SMs' measurements. The obtained results highlight that the regression-based method can effectively estimate the transformer loading, that is with the Pearson correlation coefficient R and root mean square error calculated for the real values and the estimated values of around 0.98 and 0.87, respectively, by using only a limited set of SM measurements (5 out of 21 SMs used) provided by the LECs while preserving customers' privacy rights. Among the examined algorithms, the XGBR algorithm appears the best method as it achieves adequate accuracy at significantly less simulation time (i.e. one-third of the simulation time of the RFR). By applying the proposed approach, the monitoring and self-management capability of the LECs can be realised.

Charging Load Pattern Extraction for Residential Electric Vehicles: A Training-Free Nonintrusive Method

Extracting the charging load pattern of residential electric vehicle (REV) will help grid operators make informed decisions in terms of scheduling and demand-side response management. Due to the multistate and high-frequency characteristics of integrated residential appliances from the residential perspective, it is difficult to achieve accurate extraction of the charging load pattern. To deal with that, this article presents a novel charging load extraction method based on residential smart meter data to noninvasively extract REV charging load pattern. The proposed algorithm harnesses the low-frequency characteristics of the charging load pattern and applies a two-stage decomposition technique to extract the characteristics of the charging load. The two-stage decomposition technique mainly includes: the trend component of the charging load being decomposed by seasonal and trend decomposition using loess method, and the low-frequency approximate component being decomposed by discrete wavelet technology. Furthermore, based on the extracted characteristics, event monitoring, and dynamic time warping is applied to estimate the closest charging interval and amplitude. The key features of the proposed algorithm include 1) significant improvement in extraction accuracy; 2) strong noise immunity; 3) online implementation of extraction. Experiments based on ground truth data validate the superiority of the proposed method compared to the existing ones.

Smart meter data classification using optimized random forest algorithm

Implementing a proper clustering algorithm and a high accuracy classifier for applying on electricity smart meter data is the first stage in analyzing and managing electricity consumption. In this paper, Random Forest (RF) classifier optimized by Artificial Bee Colony (ABC) which is called Artificial Bee Colony-based Random Forest (ABC-RF) is proposed. Also, in order to determine the representative load curves, the Convex Clustering (CC) is used. The solution paths generated by convex clustering show relationships among clusters that were hidden by static methods such as k-means clustering. To validate the proposed method, a case study that includes a real dataset of residential smart meters is implemented. The results evidence that the proposed ABC-RF method provides a higher accuracy if compared to other classification methods.

How would real-time electricity pricing affect the Saudi power sector in the long-run?

This paper shows the effects of real-time electricity pricing (RTP) on the long-run marginal costs of power generation in Saudi Arabia. To do this, it links a multi-sector energy system model with a residential electricity use model. The energy system model contains an economic power dispatch optimization component. The residential framework embeds households, whose decisions are governed by microeconomic principles, in a physical building energy model. The analysis entails liberalizing fuel prices for the power utilities and setting the dynamic prices of electricity equal to the long-run marginal electricity supply costs. The electricity prices are solely offered to households. The key takeaways from this analysis are:•RTP, a form of dynamic electricity pricing, reduces the variability of the marginal costs for Saudi power utilities throughout the day.•Lowered capital spending by the Saudi power sector results from RTP and consequently lower power loads. Moreover, the curtailed investment in power plants would more than cover the costs of residential smart meter replacements.

Preserving Privacy of Smart Meter Data in a Smart Grid Environment

The use of data from residential smart meters can help in the management and control of distribution grids. This provides significant benefits to electricity retailers as well as distribution system operators but raises important questions related to the privacy of consumers' information. In this article, an innovative differential privacy (DP) compliant algorithm is developed to ensure that the data from consumer's smart meters are protected. The effects of this novel algorithm on the operation of the distribution grid are thoroughly investigated not only from a consumer's electricity bill point of view but also from a power systems point of view. This method allows for an empirical investigation into the losses, power quality issues, and extra costs that such a privacy-preserving mechanism may introduce to the system. In addition, severalcost allocation mechanisms based on the cooperative game theory are used to ensure that the extra costs are divided among the participants in a fair, efficient, and equitable manner. Overall, the comprehensive results show that the approach provides privacy preservation in line with the consumer's preferences and does not lead to significant cost or loss increases for the energy retailer. In addition, the novel algorithm is computationally efficient and performs very well with a large number of consumers, thus demonstrating its scalability.

Data privacy and residential smart meters: Comparative analysis and harmonization potential

Building on privacy principles of the Fair Information Practice Principles and the European Union's General Data Protection Regulation, the study compares national policies and programs in Europe and North America and identifies prevailing practices for implementing privacy goals for residential energy customers: customer opt-out policies, sampling and sharing guidelines, independent data storage, and governmental enforcement authority. The analysis provides the basis for privacy standards that could apply to advanced-metering customer data across countries, even with rapidly evolving technology.

Uncertainty Matters: Bayesian Probabilistic Forecasting for Residential Smart Meter Prediction, Segmentation, and Behavioral Measurement and Verification

As new grid edge technologies emerge—such as rooftop solar panels, battery storage, and controllable water heaters—quantifying the uncertainties of building load forecasts is becoming more critical. The recent adoption of smart meter infrastructures provided new granular data streams, largely unavailable just ten years ago, that can be utilized to better forecast building-level demand. This paper uses Bayesian Structural Time Series for probabilistic load forecasting at the residential building level to capture uncertainties in forecasting. We use sub-hourly electrical submeter data from 120 residential apartments in Singapore that were part of a behavioral intervention study. The proposed model addresses several fundamental limitations through its flexibility to handle univariate and multivariate scenarios, perform feature selection, and include either static or dynamic effects, as well as its inherent applicability for measurement and verification. We highlight the benefits of this process in three main application areas: (1) Probabilistic Load Forecasting for Apartment-Level Hourly Loads; (2) Submeter Load Forecasting and Segmentation; (3) Measurement and Verification for Behavioral Demand Response. Results show the model achieves a similar performance to ARIMA, another popular time series model, when predicting individual apartment loads, and superior performance when predicting aggregate loads. Furthermore, we show that the model robustly captures uncertainties in the forecasts while providing interpretable results, indicating the importance of, for example, temperature data in its predictions. Finally, our estimates for a behavioral demand response program indicate that it achieved energy savings; however, the confidence interval provided by the probabilistic model is wide. Overall, this probabilistic forecasting model accurately measures uncertainties in forecasts and provides interpretable results that can support building managers and policymakers with the goal of reducing energy use.

Emerging investigator series: disaggregating residential sector high-resolution smart water meter data into appliance end-uses with unsupervised machine learning

Disaggregation of residential smart water meter data can support peak demand quantification and customized conservation and efficiency recommendations.

Separating Feeder Demand Into Components Using Substation, Feeder, and Smart Meter Measurements

Real-time, feeder-level energy disaggregation seeks to separate the measured total demand of a distribution feeder into components (e.g., into the aggregate demand-responsive load and remaining demand on the feeder, or into the aggregate generation and demand on the feeder). It can benefit distribution system operators and demand response providers by providing real-time information about balancing reserve requirements or the aggregate demand-responsive load, among other applications. In this work, we develop a feeder-level energy disaggregation algorithm that uses measurements from multiple sources that are available on different timescales. The algorithm is based on online learning and uses sensor fusion to incorporate output equations associated with disparate active and reactive power flow measurements, complex bus voltage measurements, and residential smart meter measurements. The algorithm also uses distribution substation measurements. Case studies simulate a three-phase distribution feeder model with unbalanced loads, where the active and reactive demand of the loads are modeled as a function of voltage and temperature. These models are parameterized using real-world active power demand data at each time-step. Results indicate that disaggregating three demand components is possible using feeder head active power measurements, that including feeder head reactive power measurements improves disaggregation accuracy by 32% on average, and that including intermittent real-time smart meter measurements further improves disaggregation accuracy.

Transactive Demand Response Operation at the Grid Edge using the IEEE 2030.5 Standard

Abstract This paper presents a transactive demand response (TDR) scheme for a network of residential customers with generation assets that emphasizes interoperability within a transactive energy architecture. A complete laboratory-based implementation provides the first (to our knowledge) realization of a comprehensive TDR use case that is fully compliant with the Institute of Electrical and Electronics Engineers (IEEE) 2030.5 standard, which addresses interoperability within a cybersecure smart energy profile (SEP) context. Verification is provided by a full system integration with commercial hardware using Internet Protocol (IP)-based (local area network (LAN) and Wi-Fi) communication protocols and transport layer security (TLS) 1.2 cryptographic protocol, and validation is provided by emulation using extensive residential smart meter data. The demand response (DR) scheme is designed to accommodate privacy concerns, allows customers to select their DR compliance level, and provides incentives to maximize their participation. The proposed TDR scheme addresses privacy through the implementation of the SEP 2.0 messaging protocol between a transactive agent (TA) and home energy management system (HEMS) agents. Customer response is handled by a multi-input multi-output (MIMO) fuzzy controller that manages negotiation between the customer agent and the TA. We take a multi-agent system approach to neighborhood coordination, with the TA servicing multiple residences on a common transformer, and use a reward mechanism to maximize customer engagement during the event-based optimization. Based on a set of smart meter data acquired over an extended time period, we engage in multiple TDR scenarios, and demonstrate with a fully-functional IEEE 2030.5-compliant implementation that our scheme can reduce network peak power consumption by 22% under realistic conditions.

A Novel Clustering Index to Find Optimal Clusters Size With Application to Segmentation of Energy Consumers

Increased deployment of residential smart meters has made it possible to record energy consumption data on short intervals. These data, if used efficiently, carry valuable information for managing power demand and increasing energy consumption efficiency. An efficient way to analyze these data is to first identify clusters of energy consumers, and then focus on analyzing these clusters. However deciding on the optimal number of clusters is a challenging task. In this article, we propose a clustering index that effectively finds the optimal number of clusters. The proposed index is an entropy-based measure that is obtained from eigenvalue analysis of the correlation matrix of time series of consumption data. A genetic algorithm based feature selection is used to reduce the number of features, which are then fed into clustering algorithms. We apply the proposed clustering index on two ground truth synthetic data sets and two real world energy consumption data set. The numerical simulations reveal the effectiveness of the proposed method and its superiority to a number of existing clustering indices.

Hierarchical Probabilistic Forecasting of Electricity Demand With Smart Meter Data

Decisions regarding the supply of electricity across a power grid must take into consideration the inherent uncertainty in demand. Optimal decision-making requires probabilistic forecasts for demand in a hierarchy with various levels of aggregation, such as substations, cities, and regions. The forecasts should be coherent in the sense that the forecast of the aggregated series should equal the sum of the forecasts of the corresponding disaggregated series. Coherency is essential, since the allocation of electricity at one level of the hierarchy relies on the appropriate amount being provided from the previous level. We introduce a new probabilistic forecasting method for a large hierarchy based on UK residential smart meter data. We find our method provides coherent and accurate probabilistic forecasts, as a result of an effective forecast combination. Furthermore, by avoiding distributional assumptions, we find that our method captures the variety of distributions in the smart meter hierarchy. Finally, the results confirm that, to ensure coherency in our large-scale hierarchy, it is sufficient to model a set of lower-dimension dependencies, rather than modeling the entire joint distribution of all series in the hierarchy. In achieving coherent and accurate hierarchical probabilistic forecasts, this work contributes to improved decision-making for smart grids. Supplementary materials for this article, including a standardized description of the materials available for reproducing the work, are available as an online supplement.

User-Based Load Visualization of Categorical Forecasted Smart Meter Data Using LSTM Network

Electrical load forecasting is an essential feature in power systems planning, operation and control. The non-linearity and non-stationary nature of the data, however, poses a challenge in terms of accuracy. This article explores a deep learning technique, a long short-term memory recurrent neural network-based framework to tackle this tricky issue. The proposed machine learning model framework is tested on real time residential smart meter data showing promising results. A web application has also been developed to allow consumers to have access to greater levels of information and facilitate decision-making at their end. The performance of the proposed model is also comprehensively compared to other methods in the field of load forecasting showing more accurate results for the function of forecasting of load on short term basis.

An Empirical Study of the Acceptance of IoT-Based Smart Meter in Malaysia: The Effect of Electricity-Saving Knowledge and Environmental Awareness

Today's increasing demand for electricity requires solutions that better align with energy demand and supply. Innovative technological solutions such as smart metering applications are gaining popularity among electricity providers. Despite numerous benefits, smart meters, a part of the technology on the Internet of Things (IoT), continue to struggle for widespread consumer acceptance due to limited knowledge on electricity savings and environmental awareness. These factors were examined in isolation and have not been theoretically incorporated or examined. Hence, this study investigates the factors that influence residential consumers' acceptance of smart meters by integrating electricity-saving knowledge and environmental awareness with the second generation of “unified theory of acceptance and use of technology” (UTAUT2). The literature revealed an important link between users' behavioural intention and users' use behaviour. Well-established theories of acceptance like “technology acceptance model” (TAM) and UTAUT, incorporate the behavioural intention variable in the nomological network of technology adoption determinants. This study highlighted the impact of users' behavioural intention on users' use behaviour, which was not examined previously by any of the smart meter acceptance models. The data were collected from 318 consumers of residential smart meters in Putrajaya and Malacca, the cities in Malaysia, and were statistically tested using SME-PLS The study confirms that adding electricity-saving knowledge and environmental awareness to the UTAUT2 leads to a significant increase in the explained variance in consumer acceptance of smart meter.

A Physical Probabilistic Network Model for Distribution Network Topology Recognition Using Smart Meter Data

Given the considerable scale of distribution networks in urban and rural areas, as well as the lack of management records, adjustments of switches during the distribution system operation are poorly documented. Such deficiency results in the inaccuracy of models stored in the distribution network automation system, and thus misleads the state estimation. With the emergence of information and communication technology, a large number of the feeder and residential smart meter data are accumulated. Such data can help recognize the operation modes of distribution networks by analyzing the relationships between the on/off states of switches and the voltage correlations among buses. However, the limited quantity and quality of the sampling data restrict the implementation of data-driven recognition. In this paper, a physical-probabilistic-network (PPN) model applied for inferring overall operation mode of distribution networks is proposed. Based on which, a belief propagation-based algorithm is proposed for the inference even under situations when there are only partial bus voltages data available. Meanwhile, the required variable for inference can be reduced from the active trail analysis. Experiment results are used to compare its performance with classic methods and to prove its effectiveness and advantages.

Consumo residencial: uma proposta de modelo de negócio para medidores inteligentes

São verificadas diversas iniciativas, tanto em empreendimentos como na literatura, para o desenvolvimento de produtos inteligentes voltados a aplicações residenciais, principalmente relacionados ao monitoramento e controle de consumo de energia por meio de medidores inteligentes. O uso de medidores inteligentes tem ganhado espaço em diversos países, porém verifica-se que este tipo de aplicação ainda é muito incipiente no Brasil, resultando em altos custos para as concessionárias na realização de coleta de dados de consumo no país. As pesquisas na literatura sobre medidores inteligentes apresentam principalmente foco em aspectos tecnológicos e engenharia de produto, o que resulta em uma lacuna em relação aos aspectos gerenciais e valor ao cliente. Dessa forma, este trabalho tem como objetivo desenvolver uma proposta de modelo de negócio com arranjo flexível voltada ao desenvolvimento e oferta de serviços associado aos medidores inteligentes (água e/ou energia e/ou gás) para uso doméstico. Para tanto, foram utilizadas diversas fontes de evidência a fim de se levantar dados para o desenvolvimento do modelo. Foi possível observar que, de fato, as fontes de evidência da literatura e os casos identificados apresentam foco nos aspectos tecnológicos e estrutura do produto. Os resultados apresentam uma síntese de diversas possibilidades de configuração de modelos de negócio para medidores inteligentes residenciais. A partir disso, este resultado permite estimular o desenvolvimento de negócios para o controle e fomento do consumo consciente de recursos naturais, tais como água, energia e gás, tão estratégicos e críticos para a manutenção do meio ambiente.

Non-Intrusive Load Monitoring and Classification of Activities of Daily Living Using Residential Smart Meter Data

This paper develops an approach for household appliance identification and classification of household activities of daily living (ADLs) using residential smart meter data. The process of household appliance identification, i.e., decomposing a mains electricity measurement into each of its constituent individual appliances, is a very challenging classification problem. Recent advances have made deep learning a dominant approach for classification in fields, such as image processing and speech recognition. This paper presents a deep learning approach based on multilayer, feedforward neural networks that can identify common household electrical appliances from a typical household smart meter measurement. The performance of this approach is tested and validated using publicly available smart meter data sets. The identified appliances are then mapped to household activities, or ADLs. The resulting ADL classifier can provide insights into the behavior of the household occupants, which has a number of applications in the energy domain and in other fields.