Search for collections on Undip Repository

ANALISIS PENGARUH PARAMETER OPERASIONAL TERHADAP EFISIENSI DAN PEMULIHAN PANAS PADA BOILER CIRCULATING FLUIDIZED BED DI PEMBANGKIT LISTRIK TENAGA UAP

SYACHJAYA, Muhamad Iqbal and Christwardana, Marcelinus and Sutanto, Heri (2026) ANALISIS PENGARUH PARAMETER OPERASIONAL TERHADAP EFISIENSI DAN PEMULIHAN PANAS PADA BOILER CIRCULATING FLUIDIZED BED DI PEMBANGKIT LISTRIK TENAGA UAP. Masters thesis, UNIVERSITAS DIPONEGORO.

[thumbnail of HAL COVER-1.pdf] Text
HAL COVER-1.pdf

Download (206kB)
[thumbnail of COVER LENGKAP.pdf] Text
COVER LENGKAP.pdf
Restricted to Repository staff only

Download (958kB)
[thumbnail of BAB I.pdf] Text
BAB I.pdf

Download (71kB)
[thumbnail of BAB II.pdf] Text
BAB II.pdf

Download (783kB)
[thumbnail of BAB III.pdf] Text
BAB III.pdf
Restricted to Repository staff only

Download (306kB)
[thumbnail of BAB IV.pdf] Text
BAB IV.pdf
Restricted to Repository staff only

Download (2MB)
[thumbnail of BAB V.pdf] Text
BAB V.pdf
Restricted to Repository staff only

Download (48kB)
[thumbnail of DAFTAR PUSTAKA.pdf] Text
DAFTAR PUSTAKA.pdf

Download (203kB)
[thumbnail of LAMPIRAN.pdf] Text
LAMPIRAN.pdf
Restricted to Repository staff only

Download (449kB)

Abstract

Pembangkit Listrik Tenaga Uap (PLTU) berbasis Circulating Fluidized Bed (CFB) menawarkan fleksibilitas bahan bakar yang tinggi dan emisi rendah. Namun, fluktuasi parameter operasional riil sering kali menyebabkan penurunan efisiensi termal dan ketidakoptimalan pemulihan panas (heat recovery). Sebagian besar penelitian terdahulu menganalisis variabel operasional secara parsial melalui simulasi numerik atau skala laboratorium. Penelitian ini bertujuan untuk menganalisis pengaruh simultan dari delapan parameter operasional utama (Bed Pressure, Bed Temperature, Cyclone Temperature, Oxygen Excess Air, Air Fuel Ratio/AFR, Medium Coal Size Percentage, Ratio Primary to Secondary Air, dan Turbine Load) terhadap dua indikator kinerja utama, yaitu Efisiensi Langsung Boiler (Y_1) dan Pemulihan Panas Boiler (Y_2).
Penelitian ini menggunakan pendekatan kuantitatif observasional analitik retrospektif dengan mengevaluasi data sekunder historis 1 tahun (365 hari operasi pada tahun 2024) dari sistem Distributed Control System (DCS) dan uji laboratorium di PLTU CFB 2x100 MW Kalimantan Tengah. Data dianalisis secara komprehensif menggunakan analisis bivariat, regresi multivariat (Linier, Eksponensial, dan Polinomial Orde Dua dengan dan tanpa Principal Component Analysis/PCA), K-Means Clustering, serta divalidasi dengan temporal hold-out data tahun 2025.
Hasil analisis menunjukkan bahwa model Regresi Polinomial Orde Dua tanpa PCA menghasilkan akurasi prediksi tertinggi pada data pengujian 2025. Untuk Efisiensi Langsung Boiler (Y_1), model Polinomial mencapai R^2=0,891 dan "RMSE"=1,670, dengan empat variabel asosiasi utama yaitu Excess Air, Air Fuel Ratio (AFR), Secondary Air Flow, dan Turbine Load. Adanya suku kuadratik pada Excess Air membuktikan hubungan non-linier parabolis dengan titik operasi optimum. Untuk Pemulihan Panas Boiler (Y_2), model Polinomial menghasilkan R^2=0,399 dan "RMSE"=0,773, yang secara dominan dipengaruhi oleh Bed Pressure dan Bed Temperature,
serta didukung interaksi sinergis antara Primary Air dengan Bed Pressure dan AFR dengan Bed Temperature. Penerapan PCA terbukti menurunkan akurasi prediksi akibat hilangnya varians spesifik variabel fisik. Integrasi K-Means Clustering (K=2) berhasil memetakan mode operasi kinerja tinggi (efisiensi rata-rata 78,46%) yang dicirikan oleh manajemen AFR tinggi dan Excess Air minimal, serta membuktikan bahwa Bed Pressure dan Bed Temperature berfungsi sebagai variabel stabilitas batas hidrodinamika. Temuan ini memberikan panduan preskriptif bagi operator untuk mengoptimalkan titik setel pembakaran secara dinamis guna meningkatkan efisiensi energi pembangkit.
Kata Kunci: Boiler CFB, Efisiensi Langsung, Pemulihan Panas, Regresi Polinomial Multivariat, Principal Component Analysis, K-Means Clustering

Circulating Fluidized Bed (CFB) coal-fired power plants offer high fuel flexibility and low environmental emissions. However, real-world operational parameter fluctuations often cause thermal efficiency degradation and sub-optimal heat recovery. Most previous studies evaluated operational variables partially through numerical simulations or laboratory-scale testing. This study aims to analyze the simultaneous effects of eight key operational parameters (Bed Pressure, Bed Temperature, Cyclone Temperature, Oxygen Excess Air, Air Fuel Ratio/AFR, Medium Coal Size Percentage, Ratio Primary to Secondary Air, and Turbine Load) on two main performance indicators: Boiler Direct Efficiency (Y_1) and Boiler Heat Recovery (Y_2).
A quantitative retrospective analytical observational approach was conducted using 1-year historical operational secondary data (365 operating days in 2024) retrieved from the Distributed Control System (DCS) and laboratory testing at a 2x100 MW CFB power plant in Central Kalimantan. Data were comprehensively processed through bivariate analysis, multivariate regression (Linear, Exponential, and Second-Order Polynomial with and without Principal Component Analysis/PCA), and K-Means Clustering, followed by validation using a temporal hold-out dataset from 2025.
The modeling results demonstrated that the Second-Order Polynomial Regression without PCA achieved superior predictive performance on the 2025 test dataset. For Boiler Direct Efficiency (Y_1), the Polynomial model attained R^2=0.891 and "RMSE"=1.670, associated primarily with four operational variables: Excess Air, Air-Fuel Ratio (AFR), Secondary Air Flow, and Turbine Load. The quadratic term of Excess Air confirmed a parabolic non-linear relationship with an optimum operating threshold. For Boiler Heat Recovery (Y_2), the Polynomial model yielded R^2=0.399 and "RMSE"=0.773, driven predominantly by Bed Pressure and Bed Temperature, coupled with synergistic interactions between Primary Air with Bed Pressure and AFR with Bed Temperature.
Applying PCA reduced predictive accuracy due to the loss of physical variance details. The integration of K-Means Clustering (K=2) successfully isolated a high-performance operating mode (mean efficiency of 78.46%) characterized by high AFR management and minimal Excess Air, while proving that Bed Pressure and Bed Temperature act as fundamental hydrodynamic stability constraints. These findings provide prescriptive operational guidelines for plant control operators to dynamically adjust set-points for optimal thermal efficiency.
Keywords: CFB Boiler, Direct Efficiency, Heat Recovery, Multivariate Polynomial Regression, Principal Component Analysis, K-Means Clustering

Item Type: Thesis (Masters)
Uncontrolled Keywords: Boiler CFB, Efisiensi Langsung, Pemulihan Panas, Regresi Polinomial Multivariat, Principal Component Analysis, K-Means Clustering
Subjects: Engineering
Divisions: Postgraduate Program > Master Program in Energy
Depositing User: ekana listianawati
Date Deposited: 19 Nov 2026 02:48
Last Modified: 19 Nov 2026 02:48
URI: https://eprints2.undip.ac.id/id/eprint/64448

Actions (login required)

View Item View Item