Generation of synthetic benchmark electrical load profiles using publicly available load and weather data

Pillai, G. N.; Putrus, Ghanim; Pearsall, Nicola

doi:10.1016/j.ijepes.2014.03.005

Cited by 32 publications

(9 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, synthetic building energy consumption data is especially hard to generate because, in addition to the building properties, use type, and operating schedule, energy consumption also depends on human behavior. Pillai et al [29] proposed an approach for generating synthetic load profiles using available load and weather data. While Pillai et al aim to create benchmark profiles, our work generates data for ML.…”

Section: Data Sets In Energy Forecastingmentioning

confidence: 99%

Generating Energy Data for Machine Learning with Recurrent Generative Adversarial Networks

2019

View full text Add to dashboard Cite

The smart grid employs computing and communication technologies to embed intelligence into the power grid and, consequently, make the grid more efficient. Machine learning (ML) has been applied for tasks that are important for smart grid operation including energy consumption and generation forecasting, anomaly detection, and state estimation. These ML solutions commonly require sufficient historical data; however, this data is often not readily available because of reasons such as data collection costs and concerns regarding security and privacy. This paper introduces a recurrent generative adversarial network (R-GAN) for generating realistic energy consumption data by learning from real data. Generativea adversarial networks (GANs) have been mostly used for image tasks (e.g., image generation, super-resolution), but here they are used with time series data. Convolutional neural networks (CNNs) from image GANs are replaced with recurrent neural networks (RNNs) because of RNN’s ability to capture temporal dependencies. To improve training stability and increase quality of generated data, Wasserstein GANs (WGANs) and Metropolis-Hastings GAN (MH-GAN) approaches were applied. The accuracy is further improved by adding features created with ARIMA and Fourier transform. Experiments demonstrate that data generated by R-GAN can be used for training energy forecasting models.

show abstract

Section: Data Sets In Energy Forecastingmentioning

confidence: 99%

Generating Energy Data for Machine Learning with Recurrent Generative Adversarial Networks

2019

View full text Add to dashboard Cite

show abstract

“…The yearly consumption data doesn't provide any information on daily or seasonal variations, to account for these variations a synthetic load profile was generated based on the raw data from [22]. A synthetic load profile [23] shows how the consumption (of domestic hot water in this case) at a certain point in time, relates to the mean value. The data in [22] comprised 113 dwellings, which show differences in the load profile.…”

Section: Domestic Hot Water Consumptionmentioning

confidence: 99%

Decentralized Drain Water Heat Recovery: Interaction between Wastewater and Heating Flows on a Single Residence Scale

Spriet

McNabola

2018

EWaS3 2018

View full text Add to dashboard Cite

This paper studies the potential for drain water heat recovery on a single residence scale, showing its interaction with the space and domestic hot water heating system. The article performs a Monte Carlo simulation based on measured wastewater characteristics, and heat consumption data provided by the Building Energy Ratings database. It shows the necessity of a back-up heating system, on average between 8 and 42% of the demand can be met by recovered heat. This would signify a reduction in GHG emission varying between 7.6 and 22%, but would increase costs ranging from 120 to 130%. Using a hot water reservoir increases the share of recovered heat in the mix, reducing GHG emissions and costs. However at current traditional heating prices, the drain water heat recovery system for a single residence is not financially competitive with traditional systems, showing the need for different strategies.

show abstract

“…In order to reduce the computation time, and to able to consider a range of PV penetration scenarios in parallel a program was written in MATLAB. The Distflow distribution load flow algorithm for radial networks [22,23] was chosen. As the mitigation of unbalance is a key step dealt with in the power distribution planning process, a balanced system is assumed, so a "Per Phase" analysis was used.…”

Section: Matlab Distribution Load Flowmentioning

confidence: 99%

“…For the load profiles, the temporal resolution used for this study was restricted to what was available from the CLNR dataset (84 daily profiles for a year). However, by using synthetic load profile generation methodologies [30] it is possible to extend the resolution to 365 days if essential input data required for the methodology are available. The conclusions drawn from this analysis will not be much different with an improvement in the accuracy of CLNR data, considering the magnitudes of load profiles, voltage limits and other network parameters.…”

Section: Scenarios Of Pv Generation Curtailmentmentioning

confidence: 99%

The effect of distribution network on the annual energy yield and economic performance of residential PV systems under high penetration

et al. 2017

View full text Add to dashboard Cite

Technological advances, environmental awareness and, in several countries (including the UK), financial incentives lead to the adoption of PV (photovoltaic) systems. Economic viability, an important consideration for investment in residential PV, is dependent on the annual energy yield which is affected by distribution network based factors such as point of connection to network, network hosting capacity, load profiles etc. in addition to the climate of the location. A computational algorithm easy on resources is developed in this work to evaluate the effects of distribution network on the annual energy yield of residential PV systems under scenarios of increasing PV penetration. A case study was conducted for residential PV systems in Newcastle upon Tyne with a generic UK distribution network model. Results identified penetration levels at which PV generation curtailment would occur as a consequence of network voltage rise beyond grid limits and the variation in the percentage of annual energy yield curtailed among the systems connected to the network. The volatility of economic performance of the systems depending on its location within the network is also analysed. The study also looked at the impact of the resolution of PV generation profiles on energy yield estimates and consequently economic performance.

show abstract

Generation of synthetic benchmark electrical load profiles using publicly available load and weather data

Cited by 32 publications

References 29 publications

Generating Energy Data for Machine Learning with Recurrent Generative Adversarial Networks

Generating Energy Data for Machine Learning with Recurrent Generative Adversarial Networks

Decentralized Drain Water Heat Recovery: Interaction between Wastewater and Heating Flows on a Single Residence Scale

The effect of distribution network on the annual energy yield and economic performance of residential PV systems under high penetration

Contact Info

Product

Resources

About