Sistem Prediksi Penggunaan Listrik Dan Jumlah Produksi Gula Menggunakan Metode Long Short Term Memory

Kresna, Kresna Arya Nugraha (2026) Sistem Prediksi Penggunaan Listrik Dan Jumlah Produksi Gula Menggunakan Metode Long Short Term Memory. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040201003-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Prediksi konsumsi energi listrik dan jumlah produksi gula merupakan aspek penting dalam meningkatkan efisiensi operasional pabrik gula. Penelitian ini bertujuan membangun model prediksi berbasis Long Short-Term Memory (LSTM) untuk memperkirakan konsumsi energi listrik total dan produksi gula harian di PG Krebet Baru II menggunakan data historis berbentuk deret waktu (time series). Tahapan penelitian meliputi pra-pemrosesan data yang terdiri atas pembersihan data, interpolasi nilai yang hilang, normalisasi menggunakan metode Min-Max, pembentukan data time series dengan teknik sliding window berukuran 10 hari, pelatihan model LSTM, serta evaluasi kinerja model menggunakan metrik Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), dan koefisien determinasi (R²). Variabel masukan yang digunakan adalah total konsumsi energi listrik (KWH Total), yaitu penjumlahan konsumsi energi dari PLN dan turbin. Kedua sumber energi tersebut beroperasi secara komplementer dan saling melengkapi, sehingga penggunaan total konsumsi energi lebih representatif dalam menggambarkan kondisi operasional pabrik. Selain itu, dari perspektif proses produksi, jumlah gula yang dihasilkan dipengaruhi oleh ketersediaan energi total yang digunakan untuk mengoperasikan seluruh mesin produksi. Hasil penelitian menunjukkan bahwa model LSTM mampu mempelajari hubungan temporal pada data dengan baik. Model yang dihasilkan memperoleh nilai RMSE sebesar 2,82%, MAE sebesar 2,55%, dan R² sebesar 0,999, yang menunjukkan tingkat akurasi prediksi. Hasil prediksi juga mampu mengikuti pola data aktual secara konsisten, sehingga model yang dikembangkan berpotensi digunakan sebagai pendukung pengambilan keputusan dalam perencanaan kebutuhan energi, pengendalian produksi, serta peningkatan efisiensi operasional di pabrik.
========================================================================================================================================
Predicting electrical energy consumption and the amount of sugar production are important aspects in improving the operational efficiency of sugar mills. This study aims to build a prediction model based on Long Short-Term Memory (LSTM) to estimate total electrical energy consumption and daily sugar production in PG Krebet Baru II using historical data in the form of time series. The research stages include pre-processing of data consisting of data cleaning, interpolation of missing values, normalization using the Min-Max method, formation of time series data with a 10-day sliding window technique, LSTM model training, and evaluation of model performance using Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (R²) metrics. The input variable used is total electrical energy consumption (Total KWH), which is the sum of energy consumption from PLN and turbines. The two energy sources operate in a complementary and complementary manner, so that the use of total energy consumption is more representative in describing the operational conditions of the factory. In addition, from the perspective of the production process, the amount of sugar produced is affected by the availability of the total energy used to operate the entire production machine. The results of the study show that the LSTM model is able to study the temporal relationships in the data well. The resulting model obtained an RMSE value of 2.82%, MAE of 2.55%, and an R² of 0.999, indicating a level of prediction accuracy. The prediction results are also able to follow the actual data pattern consistently, so that the developed model has the potential to be used as a decision-making support in energy demand planning, production control, and improving operational efficiency in factories.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Long Short-Term Memory (LSTM), prediksi, konsumsi energi listrik, produksi gula, time series, Long Short-Term Memory (LSTM), electricity consumption forecasting, sugar production, time series, predictive modeling.
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1001 Production of electric energy or power
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Kresna Arya Nugraha
Date Deposited:	10 Aug 2026 01:00
Last Modified:	10 Aug 2026 01:00
URI:	http://repository.its.ac.id/id/eprint/144238

Actions (login required)

View Item