Deep Learning for Household Load Forecasting—A Novel Pooling Deep RNN

Abstract

The key challenge for household load forecasting lies in the high volatility and uncertainty of load profiles. Traditional methods tend to avoid such uncertainty by load aggregation (to offset uncertainties), customer classification (to cluster uncertainties) and spectral analysis (to filter out uncertainties). This paper, for the first time, aims to directly learn the uncertainty by applying a new breed of machine learning algorithms -deep learning. However simply adding layers in neural networks will cap the forecasting performance due to the occurrence of overfitting. A novel pooling-based deep recurrent neural network (PDRNN) is proposed in this paper which batches a group of customers' load profiles into a pool of inputs. Essentially the model could address the over-fitting issue by increasing data diversity and volume. This work reports the first attempts to develop a bespoke deep learning application for household load forecasting and achieved preliminary success. The developed method is implemented on Tensorflow deep learning platform and tested on 920 smart metered customers from Ireland. Compared with the state-of-art techniques in household load forecasting, the proposed method outperforms ARIMA by 19.5%, SVR by 13.1% and classical deep RNN by 6.5% in terms of RMSE.

Review of Smart Meter Data Analytics: Applications, Methodologies, and Challenges

“…Different resolutions ranging from one minute to one week have been tested. A pooling-based deep recurrent neural network (RNN) was proposed in [28] to learn spatial information shared between interconnected customers and to address the over-fitting challenges. It outperformed ARIMA, SVR, and classical deep RNN on the Irish CER residential dataset.…”

Short-term Load Forecasting at Different Aggregation Levels with Predictability Analysis

“…Compared with conventional feedforward neural networks, RNN has the particular advantage to cope with historical data through a feedback connection. As an extension of RNN, longshort-term memory (LSTM) has been introduced in load forecasting area in the past few years [11]- [13].…”

Short-Term Load Forecasting Using Optimized LSTM Networks Based on EMD

“…Shi et al proposed a novel poolingbased deep recurrent neural network (PDRNN) for household load forecasting. The experiments on 920 smart metered customers from Ireland confirm the outperformance of PDRNN [17]. However, the main disadvantage of applying RNN to load forecasting is that the activation function of RNN uses chain rules to operate the gradient descent algorithm.…”