Sensitivity analysis to reduce duplicated features in ANN training for district heat demand prediction

Chen, S., Ren, Y., Friedrich, D., Yu, Z. and Yu, J. (2020) Sensitivity analysis to reduce duplicated features in ANN training for district heat demand prediction. Energy and AI, 2, 100028. (doi: 10.1016/j.egyai.2020.100028)

Abstract

Artificial neural network (ANN) has become an important method to model the nonlinear relationships between weather conditions, building characteristics and its heat demand. Due to the large amount of training data required for ANN training, data reduction and feature selection are important to simplify the training. However, in building heat demand prediction, many weather-related input variables contain duplicated features. This paper develops a sensitivity analysis approach to analyse the correlation between input variables and to detect the variables that have high importance but contain duplicated features. The proposed approach is validated in a case study that predicts the heat demand of a district heating network containing tens of buildings at a university campus. The results show that the proposed approach detected and removed several unnecessary input variables and helped the ANN model to reduce approximately 20% training time compared with the traditional methods while maintaining the prediction accuracy. It indicates that the approach can be applied for analysing large number of input variables to help improving the training efficiency of ANN in district heat demand prediction and other applications.

• Item Type: Articles
• Additional Information: The research presented in this article was undertaken as part of a project joint founded by Energy Technology Partnership (ETP), SP Distribution PLC (Scottish Power), grant number 146.
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Yu, Dr James and Ren, Dr Yaxing and Chen, Dr Si and Yu, Professor Zhibin
• Authors: Chen, S., Ren, Y., Friedrich, D., Yu, Z., and Yu, J.
• College/School: College of Science and Engineering > School of Engineering; College of Science and Engineering > School of Engineering > Infrastructure and Environment; College of Science and Engineering > School of Engineering > Systems Power and Energy
• Journal Name: Energy and AI
• Publisher: Elsevier
• ISSN: 2666-5468
• ISSN (Online): 2666-5468
• Published Online: 25 September 2020
• Copyright Holders: Copyright © 2020 The Author(s).
• First Published: First published in Energy and AI 2:100028
• Publisher Policy: Reproduced under a Creative Commons license