Abstract
In most deep learning-based load forecasting, an intact dataset is required. Since many real-world datasets contain missing values for various reasons, missing imputation using deep learning is actively studied. However, missing imputation and load forecasting have been considered independently so far. In this article, we provide a deep learning framework that jointly considers missing imputation and load forecasting. We consider a family of autoencoder/long short-term memory (LSTM) combined models for missing-insensitive load forecasting. Specifically, autoencoder (AE), denoising autoencoder (DAE), convolutional autoencoder (CAE), and denoising convolutional autoencoder (DCAE) are considered for extracting features, of which the encoded outputs are fed into the input of LSTM. Our experiments show that the proposed DCAE/LSTM combined model significantly improves forecasting accuracy no matter what missing rate or type (random missing, consecutive block missing) occurs compared to the baseline LSTM.
Highlights
Load forecasting is essential to balance power supply and demand
To demonstrate the performance of the proposed denoising convolutional autoencoder (DCAE)/long short-term memory (LSTM), the experimental results are presented in three aspects
We compare two domains used to input the forecasting model: 1) feature domain LSTM where the output of the encoder enters the input of the forecasting model, 2) time domain LSTM where the output of the decoder enters the input of the forecasting model
Summary
Load forecasting is essential to balance power supply and demand. Long-term load forecasting supports power system infrastructure planning while medium-term and shortterm load forecastings are used for power system operation. Deletion is to discard missing values from the dataset This method reduces the size of valid data and cannot be used for typical forecasting models that require a complete intact data set. The historical average method replaces missing values with hour-ahead, day-ahead, or week-ahead metering data [18]. We confirm that the proposed DCAE/LSTM combined model outperforms the conventional models that separately process missing imputation and forecasting. We perform feature extraction using various autoencoders and use the feature data as input to the forecasting model. We train intact data only once for each customer and do not train multiple times, no matter what missing patterns (random/block) or various missing rates appear in the test set. We evaluate the performance with test set to demonstrate the feasibility of the proposed model
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
