Prediksi Konsumsi Energi Listrik Menggunakan Pendekatan Non-Intrusive Load Monitoring Dengan Adaptive Weighted Recurrence Graph dan Convolutional Neural Network

Hakiki, Makhi Hakim (2024) Prediksi Konsumsi Energi Listrik Menggunakan Pendekatan Non-Intrusive Load Monitoring Dengan Adaptive Weighted Recurrence Graph dan Convolutional Neural Network. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6026221014-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (1MB) | Request a copy

Abstract

Konsumsi energi terutama tenaga listrik memiliki cakupan yang luas sehingga sangat dibutuhkan. Dengan meningkatnya permintaan, prediksi dan klasifikasi yang akurat atas konsumsi energi rumah tangga menjadi sangat penting. Non-Intrusive Load Monitoring (NILM) adalah sebuah pendekatan yang memungkinkan pemilahan konsumsi listrik secara keseluruhan menjadi informasi tingkat peralatan individu tanpa perlu peralatan sensor yang mengganggu. Dalam penelitian ini, kami mengusulkan metode prediksi energi listrik dan klasifikasi peralatan listrik dengan pendekatan NILM menggunakan metode Adaptive Weighted Recurrence Graph dan Convolutional Neural Network untuk menghasilkan akurasi prediksi dan klasifikasi. Keefektifan model yang diusulkan akan divalidasi dengan melakukan eksperimen pada ECO (Energy Consumption and Occupancy) dataset. Analisis performa dilakukan terhadap metode NILM dan deep learning yang sudah ada untuk menilai keakuratan dan efisiensi model. Hasil yang didapat dengan menggunakan metode Adaptive Weighted Recurrence Graph dan Convolutional Neural Network dalam melakukan prediksi konsumsi energi dan klasifikasi peralatan listrik menghasilkan MAPE terbaik dengan nilai 28,85%, nilai RMSE terbaik sebesar 1,6296, dan nilai akurasi terbaik sebesar 68%.
===================================================================
The consumption of energy especially electrical energy has a wide scope hence is in high demand. With the increasing demand, accurate prediction and classification of household energy consumption is crucial. Non-Intrusive Load Monitoring (NILM) is an approach that allows the disaggregation of overall electricity consumption into individual appliance-level information without the need for intrusive sensor equipment. In this study, we propose an electrical energy prediction and appliance classification method that combines the NILM approach using an Adaptive Weighted Recurrence Graph and Convolutional Neural Network to improve prediction and classification accuracy. The effectiveness of the proposed model is validated by conducting experiments on ECO (Energy Consumption and Occupancy) dataset. Performance analysis is conducted against existing NILM and deep learning methods to assess the accuracy, and efficiency of the model. The results obtained by using the Adaptive Weighted Recurrence Graph and Convolutional Neural Network methods in predicting energy consumption and classification of electrical appliances produce the best MAPE with a value of 28.85%, RMSE value of 1.6296, and accuracy value of 68%.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Adaptive Weighted Recurrence Graph, Convolutional Neural Network, Konsumsi Energi Listrik, Klasifikasi Peralatan, Non-Intrusive Load Monitoring, electrical energy consumption, appliance classification, non-intrusive load monitoring, adaptive weighted recurrence graph, convolutional neural network
Subjects:	T Technology > T Technology (General) > T174 Technological forecasting T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 59101-(S2) Master Thesis
Depositing User:	Makhi Hakim Hakiki
Date Deposited:	31 Jul 2024 06:40
Last Modified:	31 Jul 2024 06:40
URI:	http://repository.its.ac.id/id/eprint/111071

Actions (login required)

View Item