Load demand forecasting of residential buildings using a deep learning model

Wen, Lulu; Zhou, Kaile; Yang, Shanlin

doi:10.1016/j.epsr.2019.106073

Cited by 151 publications

(68 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a few cases, MI-ANN, WNN, GRU, DBN, RBM, ANFIS, and ART network approaches were applied to obtain better performances. Cascade NN, KNN-ANN [44], [48], [63], [65], [66], [73], [74], [75], [96], [109], [124], [129], [131], [133], [134], [137], [ [40], [42], [45] - [65], [67] - [73], [76] - [95], [97], [98], [101] - [103], [105] - [107], [110] - [119], [121], [123] - [126], [130] - [41], [42], [45] - [48], [52], [57], [61] - [63], [65], [69], [70], [72], [76],…”

Section: B Different Ann Techniques In Deep Learning Based Load Forementioning

confidence: 99%

See 1 more Smart Citation

HSIC Bottleneck Based Distributed Deep Learning Model for Load Forecasting in Smart Grid With a Comprehensive Survey

et al. 2020

View full text Add to dashboard Cite

Load forecasting is a vital part of smart grids for predicting the required electrical power using artificial intelligence (AI). Deep learning is broadly used for load forecasting in the smart grid using the artificial neural network (ANN). Generally, computing the deep learning in the smart grid requires massive data aggregation or centralization and significant computational time. This paper presents a survey of deep learning-based load forecasting techniques from 2015 to 2020. This survey discusses the studies based on their deep learning techniques, Distributed Deep Learning (DDL) techniques, Back Propagation (BP) based works, and non-BP based works in the load forecasting process. Consequent to the survey, it was determined that data aggregation dependency would be beneficial for reducing computational time in load forecasting. Therefore, a conceptual model of DDL for smart grids has been presented, where the HSIC (Hilbert-Schmidt Independence Criterion) Bottleneck technique has been incorporated to provide higher accuracy.

show abstract

Section: B Different Ann Techniques In Deep Learning Based Load Forementioning

confidence: 99%

“…BPNN is the updated class of FFNN which contains an additional BP algorithm. Because of its accurate prediction, BP has also been used in RFNN [39], WNN [40], RNN [48], ENN [50], ART network [51], LSTM [61], and other techniques. Additionally, CNN is trained with BP and BP is a necessary part of CNN [146].…”

Section: Bp and Non-bp Based Load Forecastingmentioning

confidence: 99%

HSIC Bottleneck Based Distributed Deep Learning Model for Load Forecasting in Smart Grid With a Comprehensive Survey

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Although the vast majority of research, including the above literature, focuses on the substation level, load forecasting at the low-aggregate level and, particularity, at the single household level have been of interest to researchers in recent years [13][14][15][16][17], mainly due to power system modernization. In this regard, Stephen et al, in [13], apply a clustering technique to provides a forecast for aggregated residential load based on the practice theory of human behavior.…”

Section: Introductionmentioning

confidence: 99%

“…Similar clustering techniques are often a successful approach for the similar day method [7]. Wen et al, in [14], focuses on the load forecasting of a residential building with a one-hour resolution, while Kong et al consider both individual and aggregte levels in a case with half-hourly data [15]. Thanks to monitoring of the household appliances by separate meters, single customer forecasting improves through more meaningful temporal relationships [16].…”

Section: Introductionmentioning

confidence: 99%

“…One of the most efficient deep learning models for residential load forecasting has been presented in [15] based on the long short-term memory (LSTM) neural network. Moreover, a variant of LSTM, called gated recurrent unit (GRU) has been tested in [14]. Application of iterative residual blocks (ResBlocks) in deep neural networks has also been illustrated in extracting spatial and temporal correlations [17].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

From Load to Net Energy Forecasting: Short-Term Residential Forecasting for the Blend of Load and PV Behind the Meter

et al. 2020

View full text Add to dashboard Cite

Forecast‐based modeling and robust frequency control of standalone microgrids considering high penetration of renewable sources

Kumar

Mukherjee

Mathur

et al. 2020

Int Trans Electr Energ Syst

View full text Add to dashboard Cite

Summary This article proposes forecast‐based modeling and robust frequency control strategy in isolated microgrids (MGs) to improve its stability. The intermittency and variability in renewable generation is problem for its smooth integration to MGs considering frequency stability. Continuous rise in penetration levels of renewable energy sources (RESs) is the main motivation behind forecast‐based modeling and controller design for MGs. The disturbances that affect the frequency in the MG may come from the load side and/or the generation sides. In MGs, at the generation side, the forecast of power from RESs is usually obtained to get a rough estimate of available renewable power. The forecasted power always differs from the actual one, so the secondary frequency controller may get overburdened due to forecast error resulting in abnormal frequency deviation that may lead to unstable power system. The proposed H∞ based robust control design considers the forecast error which improves the system stability and performance against disturbances coming from load/generation side. First, a long short term memory based recurrent neural network has been used to forecast the renewable power availability. Thereafter a H∞ controller is designed, and it is shown through extensive simulation studies that the H∞ controller outperforms optimized PID controllers so far as rejecting the effects of uncertainties in RESs, forecasting errors, and model parameter variations are concerned. The proposed robust controller design is also validated with real time simulator (OP4510) made by Opal‐RT. The controller hardware in the loop (CHIL) test has been done taking the simulation time step of 50 μS.

show abstract

Load demand forecasting of residential buildings using a deep learning model

Cited by 151 publications

References 34 publications

HSIC Bottleneck Based Distributed Deep Learning Model for Load Forecasting in Smart Grid With a Comprehensive Survey

HSIC Bottleneck Based Distributed Deep Learning Model for Load Forecasting in Smart Grid With a Comprehensive Survey

From Load to Net Energy Forecasting: Short-Term Residential Forecasting for the Blend of Load and PV Behind the Meter

Forecast‐based modeling and robust frequency control of standalone microgrids considering high penetration of renewable sources

Contact Info

Product

Resources

About