While prepaid electricity services provide better flexibility, it comes with an additional step for the customer. Instead of paying a monthly bill based on electric usage, a prepaid system requires customers to actively predict their electricity usage before they pay for the correct electricity value. This presents a challenge because Underestimating electricity usage may lead to a power outage. Therefore, a system that monitors electricity can be developed to address this issue. There are two approaches to developing an electric monitoring system: designing the electric meter equipped with monitoring features or designing an external capturing device to work with the current electric meter. The first approach is costly and requires a meter disassembly. Thus, in this paper, the second approach is used. By utilizing image processing and a Random Forest machine learning algorithm, a monitoring device can be developed to read the digital meter's display. Although it may affect performance due to the low-power device, Raspberry Pi 3 and Raspberry Camera are used to provide automation. This method yields an accuracy of 97% using 375 images.
The technology development from year to year is increasing rapidly, especially in the electronics devices such as notebooks and smartphones. With the rapid development of technology, lifestyle habits have also changed. This can lead to an increase in the use of electrical energy. In addition, the negligence of electricity users in monitoring electricity usage at the place of the electricity meter also causes an increase in electrical energy. Monitoring the electricity meter in real time can limit the user from manage their electricity efficiently. This study aims to create a web-based electrical energy usage prediction system. This system can make it easier for users to manage and reduce waste of electrical energy. In the development of this system, it begins by collecting image data of remaining electricity which are processed manually into electrical energy consumption data. Then the data is pre-processed so that the data is clean and ready to use. The clean data is carried out by the process of making a Long-Short Term Memory (LSTM) model which was chosen because it can overcome Time Series and Non-Linear data types. LSTM model is designed to be able to predict the use of electrical energy. Then do the web application design as an interface on the predictive data. Based on the results of the test, the LSTM model can predict the use of electrical energy with a Loss Mean Square Error (MSE) value of 0.0071. While the results of website testing carried out with the alpha test get an accuracy of 100% and a beta test of 82.64%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.