Pengendalian Kualitas Main Steam Pada PLTU X Menggunakan Diagram Kontrol Max-MCUSUM Berbasis Residual Model Multioutput Least Square Support Vector Regression (MLS-SVR)

Situmeang, Sri Elsa (2025) Pengendalian Kualitas Main Steam Pada PLTU X Menggunakan Diagram Kontrol Max-MCUSUM Berbasis Residual Model Multioutput Least Square Support Vector Regression (MLS-SVR). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003211134-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Salah satu faktor yang sangat berpengaruh terhadap kinerja sistem PLTU dalam memproduksi listrik adalah kualitas main steam, yaitu uap panas bertekanan tinggi yang dihasilkan dari proses pemanasan air di dalam boiler, yang kemudian dialirkan ke turbin untuk menghasilkan energi listrik. Tekanan, suhu, dan laju aliran yang tinggi pada main steam akan meningkatkan kinerja turbin, sehingga energi listrik yang dihasilkan akan optimal. Namun, apabila ketiga variabel tersebut terlalu tinggi maupun terlalu rendah, maka dapat mengakibatkan kerusakan unit maupun masalah pada konversi energi dalam sistem PLTU. Oleh karena itu, perlu dilakukan pengendalian kualitas secara statistik pada main steam PLTU X untuk menjaga kestabilan dan efisiensi kinerjanya dalam memproduksi listrik. Pada penelitian ini, pengendalian kualitas dilakukan menggunakan diagram kontrol Maximum Multivariate Cumulative Sum (Max-MCUSUM) dengan berbasis residual model Multi-output Least Square Support Vector Regression (MLS-SVR) untuk mengatasi autokorelasi. Hasil pemodelan MLS-SVR dengan hyper-parameter optimal γ^'=2^9, γ^''=2^(-6), dan σ=2^(-15) menunjukkan bahwa residual telah memenuhi asumsi white noise, sehingga dapat digunakan pada diagram kontrol Max-MCUSUM. Pengendalian kualitas fase I dengan k = 0,75 menunjukkan bahwa diagram telah terkendali secara statistik setelah diatasi. Namun pada fase II dengan menggunakan k yang sama, diagram belum terkendali sacara statistik. Out of control tersebut diduga disebabkan oleh interaksi variabel suhu dengan laju aliran. Analisis kapabilitas proses menunjukkan bahwa secara multivariat proses telah kapabel dan memiliki tingkat presisi dan akurasi proses yang baik.
==========================================================================================================================================
One of the key factors significantly affecting the performance of coal-fired power plants (CFPP) in electricity generation is the quality of the main steam, high-pressure superheated steam produced through the water heating process in the boiler, which is then directed to the turbine to generate electrical energy. High pressure, temperature, and flow rate of the main steam enhance turbine performance, thus optimizing the generated electricity. However, when these three variables are excessively high or low, they may cause equipment damage or disruptions in the energy conversion process within the CFPP system. Therefore, it is essential to implement statistical quality control of the main steam at CFPP X to maintain stable and efficient power generation. In this study, quality control is carried out using the Maximum Multivariate Cumulative Sum (Max-MCUSUM) control chart based on residuals from a Multi-output Least Square Support Vector Regression (MLS-SVR) model to address autocorrelation issues. The MLS-SVR modeling results, with optimal hyper-parameters γ^'=2^9, γ^''=2^(-6), and σ=2^(-15) , indicate that the residuals meet the white noise assumption and are therefore suitable for use in the Max-MCUSUM control chart. Phase I quality control with k = 0.75 shows that the diagram is statistically in control after adjustments. However, in Phase II, using the same k value, the diagram remains statistically out of control. This out of control condition is suspected to be caused by the combination of temperature and flow rate variables. Process capability analysis indicates that, from a multivariate perspective, the process is capable and demonstrates good precision and accuracy.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Main Steam, Max-MCUSUM, MLS-SVR, PLTU, Coal-Fired Power Plant (CFPP)
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > Q Science (General) > Q337.5 Pattern recognition systems Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics) T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models. T Technology > TJ Mechanical engineering and machinery > TJ164 Power plants--Design and construction T Technology > TJ Mechanical engineering and machinery > TJ223.P76 Programmable controllers
Divisions:	Faculty of Civil Engineering and Planning > Civil Engineering > 22401-(D3) Diploma 3
Depositing User:	Sri Elsa Situmeang
Date Deposited:	01 Aug 2025 08:04
Last Modified:	01 Aug 2025 08:04
URI:	http://repository.its.ac.id/id/eprint/125526

Actions (login required)

View Item