Pengendalian Kualitas Statistika Pupuk ZA III di PT Petrokimia Gresik Menggunakan Maximum Half Multivariate Control Chart Berbasis Kernel (Max-KH-Mchart) pada Residual Model Multioutput Least Square Support Vector Regression (MLS-SVR)

Metana, Lintang Genis (2026) Pengendalian Kualitas Statistika Pupuk ZA III di PT Petrokimia Gresik Menggunakan Maximum Half Multivariate Control Chart Berbasis Kernel (Max-KH-Mchart) pada Residual Model Multioutput Least Square Support Vector Regression (MLS-SVR). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221141-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Pengendalian kualitas pada proses produksi Pupuk ZA III di PT Petrokimia Gresik memerlukan pendekatan statistik multivariat karena karakteristik mutu utama yaitu kadar Nitrogen (N), Air (H₂O), dan Sulfur (S), saling berkorelasi serta sering kali tidak memenuhi asumsi normalitas multivariat. Penelitian ini bertujuan untuk melakukan pengendalian kualitas proses produksi Pupuk ZA III menggunakan diagram kendali Maximum Half Multivariate Control Chart berbasis Kernel (Max-KH-Mchart) pada residual model Multioutput Least Square Support Vector Regression (MLS-SVR), serta menganalisis kapabilitas proses produksinya. Data Fase I mencakup periode 1 Januari – 31 Oktober 2025 (292 pengamatan), sedangkan data Fase II mencakup periode 1 November – 31 Desember 2025 (61 pengamatan). Pemodelan MLS-SVR dilakukan dengan input berdasarkan lag signifikan dari model VAR (1), dengan hyper-parameter optimal γ' = 2^(-3),γ'' =2^(-10), dan σ = 2^3 menghasilkan MSE sebesar 0,0039. Karena residual tidak memenuhi asumsi distribusi normal multivariat, diagram kendali Max-KH-Mchart yang bersifat non-parametrik berbasis Kernel Density Estimation (KDE) dipilih sebagai alternatif yang robust. Hasil monitoring Fase II menunjukkan terdapat tujuh pengamatan out of control. Analisis kapabilitas proses secara univariat menunjukkan nilai C_Npk untuk N = 1,88, H₂O = 6,03, dan S = 1,81, sedangkan secara multivariat diperoleh MC_Npk = 2,69. Secara univariat maupun multivariat nilai indeks kapabilitas melebihi 1.00, yang berarti proses produksi Pupuk ZA III di PT Petrokimia Gresik telah kapabel.
======================================================================================================================================
Quality control in the production process of ZA III fertilizer at PT Petrokimia Gresik requires a multivariate statistical approach, as the primary quality characteristics, namely Nitrogen (N), Water (H₂O), and Sulfur (S) content are mutually correlated and frequently violate the multivariate normality assumption. This study aims to conduct quality control of the ZA III fertilizer production process using the Kernel-based Maximum Half Multivariate Control Chart (Max-KH-Mchart) applied to the residuals of a Multioutput Least Square Support Vector Regression (MLS-SVR) model, as well as to analyze the process capability. Phase I data covers the period from January 1 to October 31, 2025 (292 observations), while Phase II data covers the period from November 1 to December 31, 2025 (61 observations). The MLS-SVR model was constructed using inputs based on significant lags identified from the VAR (1) model, with optimal hyperparameters γ' = 2^(-3),γ'' =2^(-10), and σ = 2^3 yielding an MSE of 0.0039. Since the residuals did not satisfy the multivariate normal distribution assumption, the non-parametric Max-KH-Mchart based on Kernel Density Estimation (KDE) was selected as a robust alternative. Phase II monitoring results revealed seven out-of-control observations. Univariate process capability analysis yielded C_Npk values of 1.88 for Nitrogen (N), 6.03 for Water (H₂O), and 1.81 for Sulfur (S), while the multivariate process capability index MC_Npk was obtained as 2.69. Both univariate and multivariate capability index values exceed 1.00, indicating that the ZA III fertilizer production process at PT Petrokimia Gresik is considered capable.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Pupuk, Kernel, Max-KH-Mchart, Nonparametrik, Multivariat, Fertilizer, Kernel, Nonparametric, Multivariate
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD9980.5 Service industries--Quality control. Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Lintang Genis Metana
Date Deposited:	04 Aug 2026 04:08
Last Modified:	04 Aug 2026 04:08
URI:	http://repository.its.ac.id/id/eprint/140073

Actions (login required)

View Item