Household Load Data Research Articles

Short-term household load forecasting based on Stacking-SCN

The household load has high volatility and uncertainty, which plays an increasingly important role in short-term power load forecasting for individual residential customers in the operation and planning of future power grids. There is still room for improvement in the prediction accuracy of traditional machine learning and deep learning models. This paper firstly analyzes the complexity of household load data and formulates forecasting problems. Then, the principle of stochastic configuration network (SCN), stacking algorithm and a stacking integrated stochastic configuration network (Stacking-SCN) is introduced, which constructs a short-term household load forecasting model. By adaptively constraining the output weight distribution of the network model, the prediction accuracy can be improved. To show the prediction performance of the algorithm, the experiment comparing the Stacking-SCN model with incremental random vector functional link network (IRVFLN) models, the traditional SCN, ensemble-SCN (E-SCN), deep neural networks (DNN) and long short-term memory (LSTM), and verify it with RMSE, MAE and APE performance metrics. Experimental results show that Stacking-SCN can more effectively express the time series relationship of household short-term load, has higher prediction accuracy, effectively prevents network over-fitting problems, and improves the accuracy of prediction capabilities.

Efficient deep generative model for short-term household load forecasting using non-intrusive load monitoring

Home energy management systems (HEMS) enable key strategies and methods to improve residential efficiency and energy utilization. To make informed decisions, HEMS depend on energy monitoring and forecasting systems. However, short-term load forecasting (STLF) at the household level is challenging due to the high uncertainty in load demand caused by the difficulty of expecting customer behavior. Traditional solutions only use household load data at low temporal resolution and cannot adequately consider the users’ power consumption profile, limiting prediction accuracy. This paper addresses these issues and proposes a two-stage approach for the household STLF that leverages past and future appliance load data. In the first stage, the past and future appliance load data is estimated from the aggregate load measurement through a non-intrusive load monitoring (NILM) approach based on a deep generative model using variational autoencoders (VAE). This granular information is then augmented through a novel household load predictive model. We implement six configurations combining three stage-1 NILM models with two stage-2 predictive models based on VAE and temporal convolutional networks. Given the volatile nature of residential loads, identifying the running appliances and their consumption benefits the household STLF. The proposed approach is designed to leverage short-term appliance load forecasts rather than being limited to historical features, which is a key advantage, especially when multi-state appliances are used. The proposed approach is compared to state-of-the-art approaches on the UK-DALE and REFIT datasets and yields competitive results with improvements of 16% and 19% on average for the MAE and MAPE metrics, respectively.

Hybrid Multitask Multi-Information Fusion Deep Learning for Household Short-Term Load Forecasting

With the detailed load data provided by smart meter, the learning of electricity usage behavior for individual household short-term load forecasting has become a hot research topic. The existing methods generally ignore the relation among multiple households and fail to obtain enough information from the limited household load data, resulting in accuracy reduction. To solve this problem, a novel hybrid <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Multitask Multi-information Fusion Deep Learning Framework</i> (MFDL) considering both recent and long-term regular behaviors is proposed by integrating long short-term memory (LSTM), and convolutional neural network (CNN) to predict the multiple individual household short-term electricity consumptions. A multitask framework with two levels of information extraction is developed considering multiple household short-term load forecasting. In the low-level information extraction stage, the data structure including long and short datasets with different time scales is presented to learn the recent electricity usage behavior by LSTM and the long-term regular electricity behavior features by CNN. Then the multiple time-scale information is fused and further extracted in the high-level stage to complete the multiple individual short-term household load forecasting. The developed framework is implemented on one real-world load datasets and the results show the proposed method outperforms the state-of-the-art methods in common metrics.

Short-term apartment-level load forecasting using a modified neural network with selected auto-regressive features

Residential electricity load profiles and their diversity have become increasingly important to realize the benefits of Smart or Transactive Energy Networks (TENs). An important element of TENs will be practical, accurate, and implementable residential load forecasting techniques. While there have been many approaches to short-term load forecasting, few have included forecasting for individual households, partly because the high volatility and idiosyncrasies present in individual household load data can pose significant challenges. In this study, we develop a Convolutional Long Short-Term Memory-based neural network with Selected Autoregressive Features (termed a CLSAF model) to improve short-term household electricity load forecasting accuracy by employing three strategies: autoregressive features selection, exogenous features selection, and a “default” state to avoid overfitting at times of high load volatility. We include aggregations of apartments to floor and building level, because utilities may favor transactive approaches that rely on aggregator models, e.g., a cluster of consumers as opposed to an individual. We demonstrate that the CLSAF model, by virtue of its enhanced feature representation and modest computational resources, can accomplish load forecasting in a multi-family residential building across three spatial granularities (individual apartment/household, floor, and building levels), with an accuracy improvement of up to 25% compared to a persistence model. We propose a data screening technique to characterize time-series electricity-load data. This technique is suitable for integration into a TEN ecosystem and allows one to estimate confidence levels of the load forecasts to optimize computational resources and the risks associated with uncertain forecasts.

Energy efficiency deployment: A pathway to sustainable electrification in Ghana

Abstract Electricity is a key input for socio-economic development and its demand is expected to grow rapidly throughout the world, especially in developing countries. In Ghana, where supply challenges persist with increasing residential demand, promoting energy efficiency in households is a needful pathway. This study aims at estimating energy consumption, energy savings and peak demand reduction by introducing different standards and labelling regimes for some household appliances under different gross domestic product (GDP) scenarios for the period 2020–2050. The analysis concentrates on 5 intensive energy consuming appliances: refrigerator, freezer, air conditioner, washing machine and television; and 4 lighting technologies. A bottom-up and a top-down approach were adopted to estimate electricity consumption with the dynamic characteristics of data such as appliance ownership, population and GDP. The summary results indicate that with the implementation of the various energy efficiency scenarios, the energy consumption of the 5 appliances and 4 lighting technologies can be reduced by 24–51% in 2050 while the estimated potential cumulative energy savings range 31–104 TWh for the 30-year period. The peak demand reduction in 2050 ranges 721–2952 MW for all scenarios considered and represents 40–165% of Ghana's peak demand in 2013. Refrigerators were found to account for the highest share of the cumulative energy savings, 37–57%, while televisions contribute the highest, 28–42%, to peak demand reduction by 2050. The methods deployed in this study can easily be adapted for related studies in other countries with even little household load data, providing an important instrument for helping decision makers set adequate energy efficiency policies.

Resident Load Influence Analysis Method for Price Based on Non-intrusive Load Monitoring and Decomposition Data

In the non-invasive load monitoring mode, the load decomposition can reflect the running state of each load, which will help the user reduce unnecessary energy costs. With the demand side management measures of time of using price, a resident load influence analysis method for time of using price (TOU) based on non-intrusive load monitoring data are proposed in the paper. Relying on the current signal of the resident load classification, the user equipment type, and different time series of self-elasticity and cross-elasticity of the situation could be obtained. Through the actual household load data test with the impact of TOU, part of the equipment will be transferred to the working hours, and users in the peak price of electricity has been reduced, and in the electricity at the time of the increase Electrical equipment, with a certain regularity.

Simulation of a simple PV system for local energy usage considering the time resolution of input data

This article focuses on simulation accuracy of photovoltaic (PV) systems designed for local energy usage in households. The results of simulations can be misleading when the input data is estimated not correctly or the input data has a low time resolution. Load profile is difficult to estimate due to the changes in user behaviour and possible high power fluctuations. In contrast to many other studies, 15s data are used as PV power and household load data input. Using this data for simulations we show the influence of averaging interval on the results and thus we indicate the variance in the simulated system performance (rates of self-consumption, energy independence and grid interaction).

A Graphical Performance-Based Energy Storage Capacity Sizing Method for High Solar Penetration Residential Feeders

This paper presents a graphical, performance-based energy storage capacity sizing method for residential feeders with high solar penetration levels. The rated power and storage capacity of an energy storage device (ESD) are calculated to fulfill a specified operational requirement. Three locations for installing ESDs are investigated: 1) consumer-owned ESDs inside single-family households; 2) utility-owned distribution transformer-level ESDs; and 3) third-party owned ESDs in a community. First, historical solar radiation data, residential household load data, and residential load models are used for creating the net load (load minus solar generation) ensembles at the house level with resolution of 15 min. Then, a novel graphical capacity selection method using equal probability lines on compressed, composite cumulative distribution function curves is developed for sizing the energy storage needs at the house, distribution transformer, and community levels. Demand-side management methods are investigated for further reducing the need of energy storage. Simulation results demonstrate that the proposed method avoids over- or under-sizing ESDs and allows the users to compare the marginal benefit of increasing the capacity of the ESD.

Smart Metering Load Data Compression Based on Load Feature Identification

In recent years, smart meters have been widely installed in households across the world, which has led to problems with big data. The huge amount of household load data requires highly efficient data compression techniques to reduce the great burden on data transmittance, storage, processing, application, etc. This paper proposes the generalized extreme value distribution characteristic for household load data and then utilizes it to identify load features including load states and load events. Finally, a highly efficient lossy data compression format is designed to store key information of load features. The proposed feature-based load data compression method can support highly efficient load data compression with little reconstruction error and simultaneously provide load feature information directly for application. A case study based on the Irish Smart Metering Trial Data validates the high performance of this new approach, including in-depth comparisons with the state-of-art load data compression methods.

The Research on Cloud Platform Considered Privacy Household Load Data Processing

To solve the problems such as randomness user behavior, effective privacy protection and massive data processing shortage, this paper has proposed a household load data processing method with privacy protection, combined data mining and cloud computing. First, it gave the platform architecture and individual protection model. Then, it proposed Mask-k_means database encryption method, algorithm parallelization was implemented by MapReduce. Finally, this paper varied the method was and implemented household load data processing based on statistics and real-measured data respectively. One conclusion is that feasibility of cluster analysis for multi-user was low, while single-user load analysis is high. The other is that this method is simple and practical. This can provide a new way to household load big data under smart electricity consuming.

A novel domestic electric water heater model for a multi-objective demand side management program

This paper presents a novel domestic hot water heater model to be used in a multi-objective demand side management program. The model incorporates both the thermal losses and the water usage to determine the temperature of the water in the tank. Water heater loads are extracted from household load data and then used to determine the household water usage patterns. The benefits of the model are: (1) the on/off state of the water heater and temperature of the water in the tank can be accurately predicted, and (2) it enables the development of water usage profiles so that users can be classified based on usage behaviour. As a result, the amount of ancillary services and peak shaving that can be achieved are accurately predictable and can be maximized without adversely affecting users.