Residential Load Research Articles

A machine learning-based framework for clustering residential electricity load profiles to enhance demand response programs

Load shapes derived from smart meter data are frequently employed to analyze daily energy consumption patterns, particularly in the context of applications like Demand Response (DR). Nevertheless, one of the most important challenges to this endeavor lies in identifying the most suitable consumer clusters with similar consumption behaviors. In this paper, we present a novel machine learning based framework in order to achieve optimal load profiling through a real case study, utilizing data from almost 5000 households in London. Four widely used clustering algorithms are applied specifically K-means, K-medoids, Hierarchical Agglomerative Clustering and Density-based Spatial Clustering. An empirical analysis as well as multiple evaluation metrics are leveraged to assess those algorithms. Following that, we redefine the problem as a probabilistic classification one, with the classifier emulating the behavior of a clustering algorithm, leveraging Explainable AI (xAI) to enhance the interpretability of our solution. According to the clustering algorithm analysis the optimal number of clusters for this case is seven. Despite that, our methodology shows that two of the clusters, almost 10% of the dataset, exhibit significant internal dissimilarity. As a result, these clusters have been excluded from consideration for DR programs. The scalability and versatility of our solution makes it an ideal choice for power utility companies aiming to segment their users for creating more targeted DR programs.

Appliance-wise residential monthly average daily load profile simulation for 2022 and 2030 in Auroville, India

Appliance-wise residential monthly average daily load profile simulation for 2022 and 2030 in Auroville, India

A Secure Federated Learning Framework for Residential Short-Term Load Forecasting

A Secure Federated Learning Framework for Residential Short-Term Load Forecasting

Analysis of multi-objective cooperative optimization strategy for residential load under extreme weather

In recent years, the problem of seasonal power supply shortages in extreme weather has frequently occurred across China, which has brought great inconvenience to people’s production and life. At the same time, with the advancement of the energy revolution, the proportion of electric energy in the terminal energy is increasing, which also poses a severe challenge to China’s energy security. Our research, by constructing a residential electricity consumption model and conducting multiple electricity price response experiments, has collected data that assesses various factors involved in user responses. It has also explored and inferred the impact of these factors on user-side responses. This study introduces new research directions aimed at alleviating electricity supply-demand imbalances during different periods, contributing to the optimization and enhancement of the national power system.

ADMM-Based Hierarchical Single-Loop Framework for EV Charging Scheduling Considering Power Flow Constraints

This paper presents a three-layer hierarchical distributed framework for optimal electric vehicle charging scheduling (EVCS). The proposed hierarchical EVCS structure includes a distribution system operator (DSO) at the top layer, electric vehicle aggregators (EVAs) at the middle layer, and electric vehicles (EVs) charging stations at the bottom layer. A single-loop iterative algorithm is developed to solve the EVCS problem by combining the alternating direction method of multipliers (ADMM) and the distribution line power flow model (DistFlow). Using the single-loop structure, the primal variables of all agents are updated simultaneously at every iteration resulting in a reduced number of iterations and faster convergence. The developed framework is employed to provide charging cost minimization at the EV charging stations level, peak load shaving at the EVAs level, and voltage regulation at the DSO level. In order to further improve the performance of the optimization framework, a neural network-based load forecasting model is implemented to include the uncertainties related to non-EV residential load demand. The efficiency and the optimality of the proposed EVCS framework are evaluated through numerical simulations, conducted for a modified IEEE 13 bus test feeder with different EV penetration levels.

A price-regulated scheduling of electric vehicle with grid supporting photovoltaic and battery storage

With solar photovoltaic (PV) sources integrated into microgrids and electric vehicle (EV) charging stations, users have a variety of charging mode options to meet their needs. However, overloading of the distribution transformer, higher peak demand, and voltage instability are all potential drawbacks associated with EV charging loads. Microgrid optimisation becomes more difficult when dealing with intermittent PV generation and uncertain EV charging and discharging profiles. A Smart Home Energy Management System (SHEMS) is proposed for efficient power flow from PV sources, battery energy storage system (BESS), and an EV to maximise economic advantages for a consumer using the Artificial Hummingbird Algorithm (AHBA)-based optimisation approach. This approach considers the EV's unpredictable energy consumption, variations in residential load demand and seasonal variations in the solar insolation profile throughout the year. Consequentially, potential EV requirements such as initial and predetermined state of charge (SOC) arrangements and arrival and departure hours are also considered. Three energy management scenarios satisfying the allowable loading limit of the distribution transformer are examined, namely power transfers from home to electric vehicle (H2EV), EV to home (EV2H), and EV to grid (EV2G). The reduction in yearly operating costs (YOC) of 49.5 %, 53.2 %, and 70 % was achieved in H2EV, EV2H, and EV2G modes, respectively. This demonstrates that an effective EV and BESS management strategy can reduce the annual operating costs of a residential consumer by up to 70 % while simultaneously achieving peak load shaving.

Optimal strategy to reduce energy waste in an electricity distribution network through direct/indirect bulk load control

In recent decades, the ever-increasing demand for electricity in large cities is leading the electricity distribution network (EDN) to an operational state of fatigue, especially at peak load times. In this context, load congestion, technical losses, and voltage deviations across the EDN can contribute to the high waste of electricity as well as to the worsening of the supply service. To face this problem, this paper proposes a mixed integer linear programming (MILP) model that aims to reshape the demand profile (i.e., load reduction in peak periods without load rebound in the remaining periods of the day) of total consumers while the voltage deviations and technical losses are minimized. In the proposed model, the demand profile reshaping is carried out through direct (DC) and indirect (IC) bulk control of residential, commercial, and industrial loads, while the load factor (LF), i.e., indicator related to the efficient use of electricity, is improved. Due to the scheduling of these loads at certain times of the day, the occurrence of voltage deviations and technical losses can be mitigated by the efficient control of capacitor banks (CBs). To corroborate the applicability of the proposed MILP model, the 33-node IEEE test system was used. The results show the technical and economic gains for the distribution company (DISCO) and the total number of consumers.

Comparative Analysis of Various Topologies for AC and DC Power Distribution Systems on Efficiency Grounds

Efficiency is the factor that once wiped DC out of the power system at the birth of electricity, in nineteenth century. The battle of currents between AC and DC has reignited with the aggravating employment of DC in the sectors of generation, transmission, and utilization. This calls for cutting edge research in the field of comparative efficiency analysis of AC and DC, on distribution scale. The current research effort compares AC and DC distribution systems on efficiency grounds, for various topologies. The two systems are compared employing normal distribution technique for residential load profile. The varying efficiency of power electronic converters (PECs) with stochastic load profiles is modeled using MATLAB/SIMULINK, for each topology. DC takes the lead against AC with efficiency advantage for energy efficient load-based architectures. The architectures with variable speed drive-based loads, the efficiency advantage in favor of DC is between 1% and 2%, however, the efficiency advantage reaches to 12% with inherent DC loads. In contrast, AC takes the lead with conventional distribution, with an efficiency advantage between 4% and 5%. The findings of current research work can prove to be beneficial in making policy decisions for the adoption of DC at distribution scale.

Pattern-driven behaviour for demand-side management: An analysis of appliance use

Energy communities play a key role in the transition to sustainable energy, helping to inform and engage end-users so that they can become active energy consumers. In practice, trials and pilots often risk failure due to misplaced expectations and unforeseen behaviours when it comes to achieving flexible energy demand resources. In order to tackle these challenges, residential electricity load profile datasets and consumer survey results emerge as powerful tools for identifying controllable loads, energy consumption models, and tailored understanding of communities' energy contexts. This paper first outlines and analyses these datasets' capabilities to leverage data-driven decision-making for more efficient deployments of demand-side management (DSM) systems. A number of appliance behaviour patterns are extracted, based on high and flexible loads for shifting, being validated over three different use cases to support turn-key DSM in the presence and absence of renewable supply and bill saving. A genetic algorithm optimization is applied to underpin flexible demand reallocation and optimal community load profiles by combining time-variable tariff of use. Experiments demonstrate that controllable and shiftable appliances can reduce average peak load by up to 29% by increasing renewable self-consumption, leading to a valuable energy bill saving of 9%. Our findings also point to the current limitations of existing load/consumption datasets, which are hindering more efficient DSM design of flexibility and demand response programmes in energy communities.

A novel approach for residential load identification based on dynamic time warping

This paper proposes a new approach for load identification using a combination of fast Fourier transform (FFT) and dynamic time warping (DTW) techniques. In this approach, first, the current signals of each load device are measured individually and in various combinations to generate reference signals. FFT is then used to obtain meaningful features from the measured signals, while DTW is employed to classify the loads. This is the first known study to use the DTW algorithm for residential load identification in this structure. The proposed approach is constructed in two ways: supervised (DTW-SUP) and semi-supervised (DTW-SEM). In DTW-SEM, the reference signals for the load combinations are generated by summing the sample measurements of the individual loads, while in DTW-SUP, the actual sample measurements are used. The approach has been tested on six loads, including different load combinations. It has achieved high accuracy levels of over 90% in identifying single loads and load combinations using the error metrics Precision, Recall, and F-score. The proposed approach can potentially facilitate the implementation of energy-saving strategies in residential buildings and improve energy efficiency.

Zero energy building optimization for a residential complex with a new optimized cogeneration system for electricity, cooling, heating and freshwater production

In this research, a multi-generation system was proposed to meet the energy consumption of a 120-unit zero-energy residential complex with 100 m2 and 2 bedrooms in Sydney, Australia. To meet the heating and cooling demand of the residential complex, a compression chiller was used, and electrical energy was provided using the system's production power. The main goal of this research is to define a zero-energy building by providing the energy consumption of a residential complex with a solar system. Building Energy Optimization (BEopt) Software has been used for the process of simulating the residential complex and extracting the required load of the residential complex. The design of the load supply system of the residential complex was done using EES thermodynamic software. In this research, a large number of inputs have been used in different scenarios to optimize the residential complex, and the tool used in this research by all the modes introduced as inputs turns the problem into a multi-objective optimization problem. The results of the investigation of the residential complex showed that the amount of electricity consumption, heating, and cooling of the residential complex during the year is 36728, 133123, and 17,902 kWh, respectively. Also, by optimizing the energy consumption of the residential complex, it is possible to reduce CO2 emissions and prevent the increase in environmental pollution. The optimal results showed that the system in its most optimal state can reach an exergy efficiency of 25.18 % and a system cost rate of 46.6 $/h. The performance of the system in the city of Sydney showed that the system can produce 166,164 kWh of electricity, 467,945 kWh of heating, and 183,370 kWh of cooling in one year. Comparing the results of residential complex consumption load and system production showed that the proposed system can easily provide the residential complex consumption load throughout the year.

Capturing multiscale temporal dynamics in synthetic residential load profiles through Generative Adversarial Networks (GANs)

High-resolution residential smart meter data play a pivotal role in numerous applications, ranging from assessing hosting capacity in low-voltage grids to evaluating the economic viability of household-level solutions such as the integration of photovoltaic (PV) installations with battery energy storage systems (BESS). However, privacy concerns often restrict access to large-scale residential smart meter datasets, leaving only synthetic load profiles available to the public, which generally only represent averages of individual data. In response to this challenge, this paper proposes an innovative approach leveraging Generative Adversarial Networks (GANs) to generate synthetic residential load profiles. Our method operates in two distinct phases: First, we employ the DoppelGANger (DGAN) architecture to construct annual time series of daily consumption values. DGAN has been selected due to its demonstrated capacity for capturing interday and seasonal dynamics. Second, a wavelet-based decomposition-recombination technique is employed to create stochastic daily profiles with realistic intraday variations and peak demand behavior. Our DGAN-based approach demonstrates remarkable effectiveness, as evidenced by evaluation against a real-world dataset through a series of qualitative microbenchmarks. Moreover, the practical utility of the two-step methodology is explored in three downstream applications: (i) the integration of PV systems, (ii) PV-BESS systems increasing PV self-consumption, and (iii) PV-BESS systems boosting PV self-consumption while performing peak shaving. Our methodology is shown to accurately capture and model the intricate complexities inherent in real smart meter data.

Quantifying the Climate Co-Benefits of Hybrid Renewable Power Generation in Indonesia: A Multi-Regional and Technological Assessment

Quantifying the co-benefits of renewable energy investments can aid policymakers in identifying technologies capable of generating significant social, economic, and environmental benefits to effectively offset mitigation costs. Although there has been a growing body of work evaluating co-benefits, few studies have compared the potential co-benefits of several technologies across different regions in key countries. This study fills this gap by formulating a new modeling structure to assess the environmental–health–economic co-benefits of hybrid renewable energy systems (HRESs) in different parts of Indonesia. The proposed model is unique in that it incorporates various techno-economic activities to assess air quality, health, and economic benefits and then presents results as part of a cost–benefit analysis. From the intervention scenario, the modeling results show that installing 0.5 GW grid-connected solar PV, 100 MW of wind turbines, and a 100 MW biomass generator to cover a total of 1.64 million residential load units in the Bali province can avoid GHGs, PM2.5, disability-adjusted life years (DALYs), and provide health savings of 1.73 Mt/y, 289.02 t/y, 1648, and 6.16 million USD/y, respectively. In addition, it shows that the payback period is enhanced by one year, while the net present value is increased by 28%. In Jakarta, a 3 GW solar PV plant and a 100 MW biomass generator that supply 5.8 million residential load units can deliver 32,490 averted DALYs and 652.81 million USD/y of health care savings. Nationally, the contribution of renewable energy to the electricity supply mix could grow from the 2020 baseline of 18.85% to 26.93%, reducing dependence on oil and coal contribution by 5.32%.

PV output estimation method of power distribution station area based on federated learning framework and improved transformer neural network

In contemporary distribution networks (DNs), characterized by extensive integration of distributed energy, the photovoltaic (PV) power output data from the distribution station areas become crucial for system planning and operational optimization. Since many PVs are installed behind-the-meter (BTM), it is difficult to directly obtain PV power data through measurement devices. Therefore, it is important to estimate the BTM PV power from the aggregating data that can be directly obtained. However, the existing estimation methods usually require centralized large-scale data training, which brings certain privacy leakage risks. In order to solve these problems, we propose a federated learning-based improved Transformer Neural Network strategy to estimate BTM PV generation at the community level with data privacy protection. Initially, enhanced Transformer neural networks, employing a fused-attention mechanism, are deployed to precisely delineate the solar power generation pattern. Subsequently, federated learning principles facilitate the sharing of specific parameters among multiple edge endpoints and a central server. This model bifurcates into two layers: an individual layer, where parameters are retained locally, and an exchange layer, where parameters are collectively shared and conveyed through momentum aggregation. This dual-layer structure effectively synchronizes the capture of both unique and common characteristics. The test on the Australian residential load dataset verifies the effectiveness of the proposed method.

Systematic Evaluation of Possible Maximum Loads Caused by Electric Vehicle Charging and Heat Pumps and Their Effects on Common Structures of German Low-Voltage Grids

In this paper, we present a comprehensive assessment of the effects of residential loads, electric vehicles (EVs), and electric heat pumps (HPs) on low-voltage (LV) grids in urban, suburban, and rural areas of Germany. Firstly, real data are used to determine the typical structures for each LV grid region. Secondly, nine scenarios are defined with different levels of EV and HP penetration. Thirdly, the Low Voltage Load Flow Calculation in the DIgSILENT PowerFactory is performed for all scenarios while taking the simultaneity factor (SF) for each load type into consideration to calculate the minimum voltage and maximum loadings of transformer and lines in each grid; this allows for the grid’s potential bottlenecks to be identified. The network simulations are carried out with the consideration of charging powers of 11 kW and 22 kW in order to evaluate how an increasing EV load in the future may affect the grid’s parameters. To the best of our knowledge, no study in the literature has simultaneously addressed all of the aforementioned topics. The results of this study provide a useful framework that distribution system operators (DSOs) may apply to anticipate the forthcoming challenges and figure out when grid reinforcement will be required.

Optimal dispatching strategy for residential demand response considering load participation

To facilitate the coordinated and large-scale participation of residential flexible loads in demand response (DR), a load aggregator (LA) can integrate these loads for scheduling. In this study, a residential DR optimization scheduling strategy was formulated considering the participation of flexible loads in DR. First, based on the operational characteristics of flexible loads such as electric vehicles, air conditioners, and dishwashers, their DR participation, the base to calculate the compensation price to users, was determined by considering these loads as virtual energy storage. It was quantified based on the state of virtual energy storage during each time slot. Second, flexible loads were clustered using the K-means algorithm, considering the typical operational and behavioral characteristics as the cluster centroid. Finally, the LA scheduling strategy was implemented by introducing a DR mechanism based on the directrix load. The simulation results demonstrate that the proposed DR approach can effectively reduce peak loads and fill valleys, thereby improving the load management performance.

Privacy-Preserving Federated Learning for Value-Added Service Model in Advanced Metering Infrastructure

Advanced metering infrastructure (AMI) is the backbone of the next generation smart city and smart grid; it not only provides near real-time two-way communication between the consumers and the energy systems, but also enables the third parties (TPs) to provide relevant value-added services to the consumers to improve user satisfaction. However, the existing services are implemented in a centralized manner, which has potential and associated security and privacy risks also increased with Internet-of-Things (IoT) devices. To better balance the quality of the services and ensure users’ privacy, a TP AMI service model based on differentially private federated learning (FL) is proposed in this article. Instead of sending the private energy data to the cloud server, the proposed service model trains the neural network models locally, and only model parameters are shared with the central server. Moreover, the identity of individuals is eliminated by adding random Gaussian noise during the secure aggregation. Furthermore, an attention-based bidirectional long short-term memory (ATT-BLSTM) neural network model is adopted to solve the long-range dependency problem of conventional neural networks. In the case study, a residential short-term load forecasting (STLF) task is implemented to evaluate the performance of the proposed model. Compared with other state-of-the-art energy service models, the proposed one can achieve similar accuracy as the typical centralized model and balances the trade-off between privacy loss and prediction accuracy flexibly.

Optimizing energy consumption in smart homes: Load scheduling approaches

AbstractRising fuel prices, global warming, and environmental damage are leading to increased demand for rooftop solar energy systems connected to the power grid. The development of smart grids, modern metering systems, and energy management could promote energy conservation in households. In this study, two different meta‐heuristic optimization techniques were employed to schedule the shiftable load in suitable hours for decreasing electricity costs and minimizing peak to average ratio in a smart home while maintaining optimum user comfort. For power generation and storing energy, a grid‐connected residential load with rooftop solar panels, a battery, and an inverter is considered. First, the problem is theoretically described using a load model and an objective function, with the primary goal of reducing power costs by moving the time of usage for certain home appliances. Simulations validate the proposed strategies, which effectively reduce power costs by 4.5% by shifting the time of use, with both optimization algorithms showing similar output. The residential electricity cost before optimization was 507.12 BDT/day, which decreased to 484.33 BDT/day after optimization without compromising load turn‐off.

Efficient power management based on adaptive whale optimization technique for residential load

Efficient power management based on adaptive whale optimization technique for residential load

Managed residential electric vehicle charging minimizes electricity bills while meeting driver and community preferences

Transitions to electric vehicles (EV) are expected to increase electricity use in residences, where most tend to recharge. We develop a mathematical programming framework for shifting residential EV charging during low electricity pricing hours to minimize the additional electricity costs that a household incurs from charging their vehicle. The model also aims to meet additional household and their community preferences by adopting four secondary objectives: charging as soon as possible when the user arrives at home, charging as late as possible before the user leaves home, charging for valley filling and peak shaving of the residential load, and charging in a shared community hub by using a fast charging station. We analyze granular residential energy data from a sample of Austin households in 2018. We conduct an empirical analysis to compare households’ electricity bills under four electricity pricing schemes, including flat rates and time-of-use rates, both with and without a separate meter for EV charging. Our findings indicate that if charging behaviors remain unchanged, installing a separate meter for EVs is more expensive than treating EVs as another residential plug load. However, under time-of-use EV charging tariffs, households should adjust their charging behaviors, as validated by our optimization model. Implementing all four secondary charging objectives successfully avoids the on-peak periods of the EV charging rate and reduces households’ overall daily electricity costs by 38.87 % during the summer and by 44.3 % during the winter. Charging as soon as possible, charging as late as possible, and charging at a shared station provide drivers with increased flexibility. Charging as soon as possible and charging as late as possible lead to the lowest charger utilization, with the former having the longest charging stop time before home departure and the latter allowing for the longest charging start time upon home arrival. Charging at a shared station gives rise to the lowest share of charging time over dwell time. Charging for valley filling and peak shaving of the residential load offers less flexibility but has the advantage of flattening the load curve and mitigating high peak loads. This proves crucial in safeguarding the community’s energy distribution infrastructure.