Omar, Nadhirah (2022) Predicting synthetic load profile using anfis based on electricity usage behaviour. Masters thesis, Universiti Tun Hussein Onn Malaysia.

Preview	Text 24p NADHIRAH OMAR.pdf Download (695kB) | Preview
	Text (Copyright Declaration) NADHIRAH OMAR COPYRIGHT DECLARATION.pdf Restricted to Repository staff only Download (477kB) | Request a copy
	Text (Full Text) NADHIRAH OMAR WATERMARK.pdf Restricted to Registered users only Download (23MB) | Request a copy

Abstract

Predicting the energy consumption in a building is an effective technique for reducing energy demand and improving energy efficiency. Hence, a predictive tool is required to predict the energy consumption. In this study, proposed methodology adapted the adaptive neuro-fuzzy inference system (ANFIS) method to learn the patterns of electricity usage behaviour with the corresponding load profile. The proposed ANFIS was applied to the energy consumption pattern of university students at a residential college. The result showed that the developed method predicted the synthetic load profile accurately even when tested with unpredicted input datasets of electricity usage behaviour. This study showed that the MAPE and RMSE of the proposed ANFIS were 3.96% and 0.9204, respectively. In addition, the proposed ANFIS outperformed the state-of-the-art synthetic load profile prediction techniques of support-vector machine (SVM), multiple linear regression (MLR), Gaussian process regression (GPR), and fuzzy inference system (FIS). Finally, the effectiveness of various demand-side management (DSM) strategies was implemented at the residential college to validate the performance of the proposed ANFIS. In this study, the load-clipping strategy was better than other DSM techniques, as it saved more energy of up to 57%

Actions (login required)

View Item