Optimasi Net Plant Heat Rate PLTU Menggunakan Model Prediksi dan Surrogate Berbasis Jaringan Saraf Tiruan

Nulhakim, Lukman (2025) Optimasi Net Plant Heat Rate PLTU Menggunakan Model Prediksi dan Surrogate Berbasis Jaringan Saraf Tiruan. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032232142-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Net Plant Heat Rate (NPHR) pembangkit listrik tenaga uap (PLTU) adalah ukuran kinerja penting yang sangat mempengaruhi biaya bahan bakar (mencakup 70-80% dari biaya produksi). Peluang perbaikan Net Plant Heat Rate salah satunya adalah dengan cara memperkecil variabilitas parameter operasi. Penelitian ini bertujuan untuk membangun machine learning yang dapat digunakan sebagai alat yang dapat memberikan rekomendasi pengaturan parameter-parameter operasi sedemikian sehingga net plant heat rate minimal. Dengan alat ini, pengoperasian antar setiap regu operasi relatif akan lebih seragam (variability lebih baik). Machine Learning dibangun secara offline dengan menggunakan algoritma jaringan saraf tiruan multilayer perceptron (MLP) untuk model prediksi NPHR, selanjutnya dilakukan penentuan parameter operasi yang memberikan nilai NPHR minimal dengan menggunakan jaringan saraf tiruan sebagai surrogate model. Algoritma yang digunakan untuk optimasi ini adalah Particle Swarm Optimization (PSO). Objek studi adalah PLTU subkritis berbahan bakar pulverized coal dengan kapasitas terpasang 400 MW. Penelitian memberikan hasil model prediksi dengan kinerja yang sangat baik sebagai berikut: MAE 25.1370 (training set) dan 27,1685 (test set); RMSE 32,7582 (training set) dan 35,2527 (test set); MAPE 0.8% (training set) dan 0.9% (test set); serta R2 0,9722 (training set) dan 0,9685 (test set). Optimasi dengan PSO memberikan perbaikan NPHR sebesar 0.089% - 4.90% dengan rata-rata sebesar 2.36%. Hasil penelitian ini menunjukkan bahwa machine learning dapat diterapkan untuk alat pengambilan keputusan untuk optimasi NPHR dari suatu pembangkit listrik batubara.

======================================================================================================================================
The Net Plant Heat Rate (NPHR) of coal-fired power plants is a key performance indicator that significantly influences fuel cost, which accounts for approximately 70-80% of the total production cost. One opportunity to improve NPHR is to reduce the variability of operating parameters. This study aims to develop a machine learning model that can be utilized as a tool to provide recommendations for adjusting operating parameters to minimize NPHR. With this tool, plant operation among different operator shifts becomes more uniform (showing improved variability). The machine learning model was developed offline using the Artificial Neural Network Multilayer Perceptron (MLP) algorithm to predict NPHR. Subsequently, the optimal operating parameters that yield minimum NPHR were determined using ANN as a surrogate model. The optimization process in this study was carried out using the Particle Swarm Optimization (PSO) algorithm. The object of study is a pulverized coal-fired steam-generating power plant with a 400 MW installed capacity. Prediction model of this study achieve good performance with metric evaluated are MAE of 25.1370 (training set) and 27.1685 (test set); RMSE of 32.7572 (training set) and 35.2527 (test set); MAPE of 0.8% (training set) and 0.9% (test set); and R2 of 0.9722 (training set) and 0.9685 (test set). Optimization using PSO resulted in an improvement of NPHR ranging from 0.86% to 4.9%, with an average improvement of 2.36%. The findings of this study demonstrate that machine learning can be applied as a decision-support tool for optimizing the NPHR of coal-fired power plants.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	jaringan saraf tiruan, multilayer perceptron (MLP), Net Plant Heat Rate, Particle Swarm Optimization (PSO), PLTU batu bara, aritificial neural network, multilayer perceptron (MLP), Net Plant Heat Rate, Particle Swarm Optimization (PSO), coal-fired power plant.
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1322.6 Electric power-plants
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Lukman Nulhakim
Date Deposited:	27 Jan 2026 04:21
Last Modified:	27 Jan 2026 04:21
URI:	http://repository.its.ac.id/id/eprint/130414

Actions (login required)

View Item