Pengendalian Kualitas pada Proses Produksi Pupuk ZA I di PT Petrokimia Gresik Menggunakan Diagram Kontrol Max-MCUSUM Berbasis Residual MLS-SVR

Himmah, Faiqotul (2026) Pengendalian Kualitas pada Proses Produksi Pupuk ZA I di PT Petrokimia Gresik Menggunakan Diagram Kontrol Max-MCUSUM Berbasis Residual MLS-SVR. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221091_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (10MB) | Request a copy

Abstract

Pupuk memiliki peranan penting dalam meningkatkan produktivitas pertanian. PT Petrokimia Gresik merupakan pabrik pupuk terlengkap di Indonesia yang memproduksi Pupuk ZA I. Proses produksinya merupakan proses kimia kontinu dengan karakteristik kualitas yaitu Nitrogen (N), Air (H₂O), dan Sulfur (S) yang saling berhubungan dan menunjukkan adanya autokorelasi. Autokorelasi dapat menyebabkan false alarm dan pengambilan keputusan yang tidak tepat. Penelitian ini menggunakan algoritma Multi-output Least Square Support Vector Regression (MLS-SVR) untuk mengatasi autokorelasi, kemudian dilakukan pengendalian menggunakan diagram kontrol Max-MCUSUM. Input model ditentukan berdasarkan lag signifikan dari PACF. Pemodelan MLS-SVR dengan kernel RBF menghasilkan hyper-parameter optimal yaitu γ^'=2^(-5), γ^''=2^(-4), dan σ=2^1 dengan MSE sebesar 0,0046. Residual yang diperoleh digunakan untuk pembuatan diagram kontrol Max-MCUSUM. Pada fase I dengan k=0,5, didapatkan diagram terkendali secara statistik setelah diatasi. Pada fase II dengan k yang sama, terdapat titik keluar batas kontrol yang disebabkan oleh Nitrogen, Air, dan Sulfur. Hasil kapabilitas proses secara univariat dan multivariat telah memenuhi batas spesifikasi. Dengan demikian, kinerja proses produksi Pupuk ZA I di PT Petrokimia Gresik telah kapabel.
=====================================================================================================================================
Fertilizer plays an important role in increasing agricultural productivity. PT Petrokimia Gresik is the most complete fertilizer factory in Indonesia that produces ZA I fertilizer. The production process is a continuous chemical process with quality characteristics namely Nitrogen (N), Water (H₂O), and Sulfur (S) which are interrelated and indicate autocorrelation. Autocorrelation can cause false alarms and incorrect decision-making. This study uses the Multi-output Least Square Support Vector Regression (MLS-SVR) algorithm to overcome autocorrelation, then performs process control using a Max-MCUSUM control chart. The model input is determined based on significant lags from PACF. MLS-SVR modeling with RBF kernel yields optimal hyper-parameters γ^'=2^(-5), γ^''=2^(-4), and σ=2^1 with MSE of 0,0046. The residuals obtained are used to construct the Max-MCUSUM control chart. In Phase I with k=0,5, the chart is statistically in-control after the out-of-control points are addressed. In Phase II with the same k value, there are out-of-control points caused by Nitrogen, Water, and Sulfur. The process capability both univariately and multivariately has met the specification limits. Thus, the production process performance of ZA I fertilizer at PT Petrokimia Gresik is capable.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Autokorelasi, Max-MCUSUM, MLS-SVR, Pengendalian Kualitas, Pupuk ZA I, Autocorrelation, Max-MCUSUM, MLS-SVR, Quality Control, ZA I Fertilizer
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD9980.5 Service industries--Quality control. T Technology > TS Manufactures > TS156 Quality Control. QFD. Taguchi methods (Quality control)
Divisions:	Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Faiqotul Himmah
Date Deposited:	28 Jul 2026 02:41
Last Modified:	28 Jul 2026 02:41
URI:	http://repository.its.ac.id/id/eprint/138344

Actions (login required)

View Item