Robust short-term electrical load forecasting framework for commercial buildings using deep recurrent neural networks

Chitalia, Gopal; Pipattanasomporn, Manisa; Garg, Vishal; Rahman, Saifur

doi:10.1016/j.apenergy.2020.115410

Cited by 160 publications

(71 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They demonstrated that the multilayered self-normalizing technology could improve the prediction performance of the GRU and LSTM models through several experiments. Chitalia et al [ 19 ] presented a short-term load forecasting framework that is robust regardless of building types or locations. They collected five commercial buildings of five different building types located at five different locations.…”

Section: Related Studiesmentioning

confidence: 99%

An Attention-Based Multilayer GRU Model for Multistep-Ahead Short-Term Load Forecasting

Jung

Moon

Park

et al. 2021

Sensors

View full text Add to dashboard Cite

Recently, multistep-ahead prediction has attracted much attention in electric load forecasting because it can deal with sudden changes in power consumption caused by various events such as fire and heat wave for a day from the present time. On the other hand, recurrent neural networks (RNNs), including long short-term memory and gated recurrent unit (GRU) networks, can reflect the previous point well to predict the current point. Due to this property, they have been widely used for multistep-ahead prediction. The GRU model is simple and easy to implement; however, its prediction performance is limited because it considers all input variables equally. In this paper, we propose a short-term load forecasting model using an attention based GRU to focus more on the crucial variables and demonstrate that this can achieve significant performance improvements, especially when the input sequence of RNN is long. Through extensive experiments, we show that the proposed model outperforms other recent multistep-ahead prediction models in the building-level power consumption forecasting.

show abstract

Section: Related Studiesmentioning

confidence: 99%

An Attention-Based Multilayer GRU Model for Multistep-Ahead Short-Term Load Forecasting

Jung

Moon

Park

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…On the other hand, the traditional ML algorithms suffer when processing long sequences which results on degradation problems in the network (e.g., vanishing or exploding gradients) [8]. Consequently, the research in traditional Machine-Learning algorithms remained relatively limited because of the aforementioned bottlenecks produced when using large datasets [9].…”

Section: Introductionmentioning

confidence: 99%

Short-Term Load Forecasting Using Encoder-Decoder WaveNet: Application to the French Grid

2021

View full text Add to dashboard Cite

The prediction of time series data applied to the energy sector (prediction of renewable energy production, forecasting prosumers’ consumption/generation, forecast of country-level consumption, etc.) has numerous useful applications. Nevertheless, the complexity and non-linear behaviour associated with such kind of energy systems hinder the development of accurate algorithms. In such a context, this paper investigates the use of a state-of-art deep learning architecture in order to perform precise load demand forecasting 24-h-ahead in the whole country of France using RTE data. To this end, the authors propose an encoder-decoder architecture inspired by WaveNet, a deep generative model initially designed by Google DeepMind for raw audio waveforms. WaveNet uses dilated causal convolutions and skip-connection to utilise long-term information. This kind of novel ML architecture presents different advantages regarding other statistical algorithms. On the one hand, the proposed deep learning model’s training process can be parallelized in GPUs, which is an advantage in terms of training times compared to recurrent networks. On the other hand, the model prevents degradations problems (explosions and vanishing gradients) due to the residual connections. In addition, this model can learn from an input sequence to produce a forecast sequence in a one-shot manner. For comparison purposes, a comparative analysis between the most performing state-of-art deep learning models and traditional statistical approaches is presented: Autoregressive-Integrated Moving Average (ARIMA), Long-Short-Term-Memory, Gated-Recurrent-Unit (GRU), Multi-Layer Perceptron (MLP), causal 1D-Convolutional Neural Networks (1D-CNN) and ConvLSTM (Encoder-Decoder). The values of the evaluation indicators reveal that WaveNet exhibits superior performance in both forecasting accuracy and robustness.

show abstract

“…Thus, the prediction accuracy is improved. At present, deep learning technology has been used in load forecasting [35]- [37], wind power output forecasting [38]- [40], and other fields. It has achieved good prediction results.…”

Section: Introductionmentioning

confidence: 99%

Short-Term Photovoltaic Power Prediction Based on Similar Days and Improved SOA-DBN Model

Zhang

Zhu

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Existing methods in predicting short-term photovoltaic (PV) power have low accuracy and cannot satisfy actual demand. Thus, a prediction model based on similar days and seagull optimization algorithm (SOA) is proposed to optimize a deep belief network (DBN). Fast correlation-based filter (FCBF) method is used to select a meteorological feature set with the best correlation with PV output and avoid redundancy among meteorological factors affecting PV output. In addition, a comprehensive similarity index combining European distance and gray correlation degree is proposed to select the similar day. Then, SOA is used to optimize the number of neurons and the learning rate parameters in DBN. Based on the nonuniform mutation and opposition-based learning method, an improved seagull optimization algorithm (ISOA) with higher optimization accuracy is proposed. Finally, the ISOA-DBN prediction model is established, and the experimental analysis is conducted using the actual data of PV power stations in Australia. Results show that compared with DBN, support vector machine (SVM), extreme learning machine (ELM), radial basis function (RBF), Elman, and back propagation (BP), the mean absolute percentage error indicator of ISOA-DBN is only 1.512% on a sunny day, 5.975 on a rainy day, 3.359 on a cloudy to sunny day, and 1.911% on a sunny to cloudy day. Therefore, the good accuracy of the proposed model is verified. INDEX TERMSPV power generation, prediction model, similar day, seagull optimization algorithm, deep belief network NOMENCLATURE PV photovoltaic SOA seagull optimization algorithm DBN deep belief network FCBF fast correlation-based filter ISOA improved seagull optimization algorithm SVM support vector machine ELM extreme learning machine RBF radial basis function BP back propagation CEEMD complementary ensemble empirical mode decomposition SU symmetric uncertainty GR global radiation DR diffused radiation RBM restricted Boltzmann machines GBRBM Gauss-Bernoulli-restricted Boltzmann machine PSO particle swarm optimization GA genetic algorithm GOA grasshopper optimization algorithm RMSE root mean square error MAPE mean absolute percent error

show abstract

Robust short-term electrical load forecasting framework for commercial buildings using deep recurrent neural networks

Cited by 160 publications

References 51 publications

An Attention-Based Multilayer GRU Model for Multistep-Ahead Short-Term Load Forecasting

An Attention-Based Multilayer GRU Model for Multistep-Ahead Short-Term Load Forecasting

Short-Term Load Forecasting Using Encoder-Decoder WaveNet: Application to the French Grid

Short-Term Photovoltaic Power Prediction Based on Similar Days and Improved SOA-DBN Model

Contact Info

Product

Resources

About