|Load prediction for peak demand limiting
|Wang, Shengwei (BEEE)
Xiao, Fu (BEEE)
|Electric power systems
Hong Kong Polytechnic University -- Dissertations
|Department of Building Environment and Energy Engineering
|viii, 77 pages : color illustrations
|In commercial and non-residential building, the electricity bill always contains the power consumption and the peak demand in the bill period, and the latter one may contribute to more than half of the bill with a relatively short occurrence time. To reduce the peak demand, many demand limiting strategies have been developed, and load predication is an important part which can guide the use of the limiting method. This project aims to analyze the performance of different machine learning models and effects of different prediction horizons on building load prediction.
In this project, Artificial Neural Network (ANN) and Recurrent Neural Network (RNN) are learned, and Multi-Layer Perceptron (MLP), Long Short-Term Memory Network (LSTM), Gated Recurrent Unit (GRU) are used to realize the load prediction of a commercial building. The training and prediction of the model only concern the on-peak time, the 7-days ahead prediction result are used to compare the accuracy of the prediction model with three different algorithms. The result shows that the RNN is better than the ANN in the prediction accuracy, and GRU is better than LSTM. As for the different input and output horizons, when the input horizon is the same, the one-day output horizon has the highest prediction accuracy. When the output horizon is the same, the seven-days input horizon has the highest prediction accuracy. Based on the different time of the input, two different structures of the prediction model are proposed. In the common one, the history data are used as the input combined with the weather forecast data, and in the parallel structure, historical data and weather forecast data are used as two separate inputs to the model. The result shows that the parallel structure has higher prediction accuracy than the common one.
|All rights reserved
Files in This Item:
|For All Users (off-campus access for PolyU Staff & Students only)
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: