Household electricity load forecasting using historical smart meter data with clustering and classification techniques

Yildiz, Baran; Bilbao, José Ignacio; Dore, Jonathon; Sproul, Alistair B.

doi:10.1109/isgt-asia.2018.8467837

Cited by 36 publications

(25 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead of using all available data to train a machine learning-based model, it adjusts models to capture data patterns assigned to a given cluster. In [135] a K-means clustering algorithm was used to segment electricity readings on a hourly basis, along with a CNN model used to approximate each cluster, while in [136] household electricity loads are predicted by using classification and regression trees, together with self-organising map data clustering. (f) Gaussian process regression (GPR) is a non-parametric method based on gaussian processes (GPs) [137].…”

Section: ) Building Energy Consumptionmentioning

confidence: 99%

A Survey on Home Energy Management

et al. 2020

View full text Add to dashboard Cite

Energy is a vital resource for human activities and lifestyle, powering important everyday infrastructures and services. Currently, pollutant and non-renewable sources, such as fossil fuels, remain the main source of worldwide consumed energy. The environmental impact of their exploitation has boosted research and investments in alternative, clean and renewable sources, including photovoltaic and windbased systems. As a whole, buildings are one of the major energy consumption sectors. Hence, improving energy efficiency in buildings will result in economical and environmental gains. In the case of households, home energy management systems are mainly used for monitoring real-time consumption and to schedule appliance operations so that the energy bill could be minimised, or according to another specific criterion. This work aims to survey the most recent literature on home energy management systems, providing an aggregated and unified perspective in the context of residential buildings. In addition, an updated literature list regarding commonly managed household appliances and scheduling objectives are included. Physical and operational constraints, and how they are addressed by home energy management systems along with security issues are also discussed. INDEX TERMS Energy efficiency, home energy management systems, household appliance models, load management, optimal scheduling, smart homes, security.

show abstract

Section: ) Building Energy Consumptionmentioning

confidence: 99%

A Survey on Home Energy Management

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Moreover, [20] shows how an approach incorporating temporal and spatial information, can be used to identify relevant features among different households and improve the forecasting accuracy. Reference [21] also employs a clustering analysis to extract typical daily consumption patterns that can be used to improve the accuracy of the forecasting tool.…”

Section: Literature Reviewmentioning

confidence: 99%

A Single Scalable LSTM Model for Short-Term Forecasting of Massive Electricity Time Series

2020

View full text Add to dashboard Cite

Most electricity systems worldwide are deploying advanced metering infrastructures to collect relevant operational data. In particular, smart meters allow tracking electricity load consumption at a very disaggregated level and at high frequency rates. This data opens the possibility of developing new forecasting models with a potential positive impact on electricity systems. We present a general methodology that can process and forecast many smart-meter time series. Instead of using traditional and univariate approaches, we develop a single but complex recurrent neural-network model with long short-term memory that can capture individual consumption patterns and consumptions from different households. The resulting model can accurately predict future loads (short-term) of individual consumers, even if these were not included in the original training set. This entails a great potential for large-scale applications as once the single network is trained, accurate individual forecast for new consumers can be obtained at almost no computational cost. The proposed model is tested under a large set of numerical experiments by using a real-world dataset with thousands of disaggregated electricity consumption time series. Furthermore, we explore how geo-demographic segmentation of consumers may impact the forecasting accuracy of the model.

show abstract

“…As discussed before, both the steps in the clustering framework require an intermediate step to determine the optimal number of reduced dimensions (for dimensionality reduction step) and the optimal number of clusters (for clustering step). These steps are popularly referred as the hyper-parameter settings in the literature [6], [16].…”

Section: A Hyper-parameter Settingsmentioning

confidence: 99%

A Clustering Framework for Residential Electric Demand Profiles

Jain

AlSkaif

Dev

2020

2020 International Conference on Smart Energy Systems and Technologies (SEST)

View full text Add to dashboard Cite

The availability of residential electric demand profiles data, enabled by the large-scale deployment of smart metering infrastructure, has made it possible to perform more accurate analysis of electricity consumption patterns. This paper analyses the electric demand profiles of individual households located in the city Amsterdam, the Netherlands. A comprehensive clustering framework is defined to classify households based on their electricity consumption pattern. This framework consists of two main steps, namely a dimensionality reduction step of input electricity consumption data, followed by an unsupervised clustering algorithm of the reduced subspace. While any algorithm, which has been used in the literature for the aforementioned clustering task, can be used for the corresponding step, the more important question is to deduce which particular combination of algorithms is the best for a given dataset and a clustering task. This question is addressed in this paper by proposing a novel objective validation strategy, whose recommendations are then cross-verified by performing subjective validation.

show abstract

Household electricity load forecasting using historical smart meter data with clustering and classification techniques

Cited by 36 publications

References 15 publications

A Survey on Home Energy Management

A Survey on Home Energy Management

A Single Scalable LSTM Model for Short-Term Forecasting of Massive Electricity Time Series

A Clustering Framework for Residential Electric Demand Profiles

Contact Info

Product

Resources

About