Electricity consumption prediction based on LSTM with attention mechanism

Abstract

Power data analysis in power system, such as electricity consumption prediction, has always been the basis for the power department to adjust electricity price, substation regulation, total load prediction and peak avoidance management. In this paper, a short‐term time‐phased electricity consumption prediction model based on Long Short‐Term Memory (LSTM) with an attention mechanism is proposed. First, the attention mechanism is used to assign weight coefficients to the input sequence data. Then, the output value of every cell of LSTM is calculated according to the forward propagation method, and the error between the real value and the predicted value is calculated using the back‐propagation method. The gradient of each weight is calculated according to the corresponding error term, and the weight of the model is updated by the gradient descent direction to make the error smaller. Using modeling and predicting experiments on different types of electricity consumption, the results show that the prediction accuracy of the model proposed increased by 6.5% compared to the state‐of‐the‐art model. The model has a good effect on electricity consumption prediction. Not only can it be close to actual results numerically, but it can also better predict the development trend of data. © 2020 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

Electricity consumption forecasting using DFT decomposition based hybrid ARIMA-DLSTM model

“…Compared with the other benchmark models, ANN is the best performing single model for one-day-ahead and one-week-ahead predictions, however, for 10-day ahead prediction Prophet out performed ANN. LSTM with attention seeking proposed in [21], performed better than the other single and hybrid referenced models in one-day-ahead prediction, however, its performance degraded in oneweek-ahead and ten-day ahead prediction as the other models outperformed it. The one-day-ahead, oneweek-ahead and 10 days-ahead predictions achieved by the benchmark models and the proposed ARIMA-DLSTM model are depicted in Figure 3, Figure 4, and Figure 5 respectively.…”

“…To improve the prediction results, they incorporated the concept of moving-window-based active learning. An LSTM-based short-term time-phased electricity consumption prediction model with an attention mechanism was also proposed by [21]. The technique assigns a weight coefficient to the input sequence data.…”

Research on Electric Load Forecasting and User Benefit Maximization Under Demand-Side Response

“…Some scholars try to apply attention mechanism to power load prediction to improve the accuracy of load prediction. LIN et al(2020) proposed an LSTM model based on attention mechanism, and tested the effectiveness of the model by using four different types of real load data: housing, large industry, commerce and agriculture. Shao et al(2021) established VMD-IdbigRU load prediction model based on attention mechanism.…”

Review of Low Voltage Load Forecasting: Methods, Applications, and Recommendations

“…Petros , anu et al [176] find that the combination of an LSTM with an NARX Feedback Neural Network can produce improvements over either alone. In [141] an LSTM with an attention mechanism is proposed for short-term load forecasting in the China Southern Power Grid and four types of aggregated load; 'Residential', 'Large Industrial electricity', 'Business' and 'Agricultural'. Particularly, the attention mechanism is used to assign weight coefficients to the input sequence data so that the specific features can be accurately extracted.…”