Pengendalian Kualitas Proses Produksi Pupuk NPK Phonska PT Petrokimia Gresik Menggunakan Peta Kendali MEWMV dan SSRM Berbasis Multioutput Least Square Support Vector Regression (MLS-SVR)

Rachman, Annisa (2025) Pengendalian Kualitas Proses Produksi Pupuk NPK Phonska PT Petrokimia Gresik Menggunakan Peta Kendali MEWMV dan SSRM Berbasis Multioutput Least Square Support Vector Regression (MLS-SVR). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003211124-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

PT Petrokimia Gresik merupakan Badan Usaha Milik Negara (BUMN) yang memproduksi berbagai macam pupuk salah satunya adalah pupuk NPK Phonska. Pupuk tersebut memiliki empat kandungan utama, yaitu Air, Nitrogen, Fosfat, dan Kalium. Semua kandungan yang terdapat di pupuk NPK Phonska harus sesuai dengan spesifikasi yang telah ditetapkan perusahaan agar manfaat dari pupuk tersebut didapatkan secara maksimal. Oleh karena itu, untuk menjaga kualitas dari proses produksi pupuk tersebut dilakukan pengendalian kualitas menggunakan diagram kendali. Keempat karakteristik kualitas pupuk NPK Phonska memiliki hubungan satu sama lain dan tidak berdistribusi normal multivariat, selain itu juga terdapat autokorelasi pada data produksi. Sehingga, digunakan diagram kendali Multivariate Exponentially Weighted Moving Variance (MEWMV) dan Spatial Signed Rank MEWMA (SSRM) berbasis Multioutput Least Square Support Vector Regression (MLS-SVR). Pemodelan dilakukan untuk mengatasi autokorelasi pada data. Pada proses pemodelan diperoleh hyperparameter yang terpilih, yaitu γ^'=2^9 ,γ^''=2^4, dan σ=2^(-15) dengan nilai MSE sebesar 0,741. Pada diagram kendali MEWMV dengan kombinasi pembobot ω=0,4, λ=0,2, dan L=3,4461 diperoleh bahwa variabilitas proses fase I telah terkendali secara statistik setelah dilakukan 3 kali penanganan pada pengamatan out of control. Pengendalian variabilitas proses fase II juga telah terkendali secara statistik setelah dilakukan 1 kali penanganan pada pengamatan out of control. Selanjutnya, untuk pengendalian rata-rata proses menggunakan diagram kendali SSRM dengan pembobot λ=0,05 didapatkan bahwa proses telah terkendali secara statistik setelah dilakukan 45 kali penanganan terhadap pengamatan out of control. Pada pengendalian rata-rata proses fase II tidak terkendali secara statistik karena masih terdapat 31 pengamatan out of control. Pada analisis kapabilitas proses diperoleh bahwa kinerja proses produksi pupuk NPK Phonska PT Petrokimia Gresik belum kapabel secara multivariat.
======================================================================================================================================
PT Petrokimia Gresik is a State-Owned Enterprise (SOE) or State-Owned Company (commonly known in Indonesia as BUMN) that produces various types of fertilizers, one of which is NPK Phonska fertilizer. This fertilizer contains four main components: Water, Nitrogen, Phosphate, and Potassium. All the components in NPK Phonska fertilizer must comply with the specifications set by the company to ensure its maximum effectiveness. Therefore, to maintain the quality of the production process, quality control is carried out using control charts. The four quality characteristics of NPK Phonska fertilizer are interrelated and do not follow a multivariate normal distribution. Additionally, there is autocorrelation present in the production data. Thus, the Multivariate Exponentially Weighted Moving Variance (MEWMV) control chart and the Spatial Signed Rank MEWMA (SSRM) based on Multioutput Least Squares Support Vector Regression (MLS-SVR) are used. Modeling is conducted to address the autocorrelation in the data. During the modeling process, the selected hyperparameters were γ^'=2^9 ,γ^''=2^4, and σ=2^(-15), with a resulting MSE value of 0,741. For the MEWMV control chart using the weight combination ω=0,4, λ=0,2, and L=3,4461, it was found that the Phase I process variability was statistically in control after three adjustments to the out-of-control observations. The process variability in Phase II was also found to be statistically in control after one adjustment to an out-of-control observation. Furthermore, for monitoring the process mean using the SSRM control chart with a weight of λ=0,05, the process was statistically in control after 45 adjustments to out-of-control observations. However, the monitoring of the process mean in Phase II was not statistically in control, as there were still 31 out-of-control observations. The process capability analysis showed that the production process performance of NPK Phonska fertilizer at PT Petrokimia Gresik is not yet capable in a multivariate sense.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	MEWMV, MLS-SVR, Pengendalian Kualitas, Pupuk NPK Phonska, SSRM MEWMV, MLS-SVR, NPK Phonska Fertilizer, Quality Control, SSRM
Subjects:	H Social Sciences > HD Industries. Land use. Labor H Social Sciences > HD Industries. Land use. Labor > HD9980.5 Service industries--Quality control.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Annisa Rachman
Date Deposited:	31 Jul 2025 06:16
Last Modified:	31 Jul 2025 06:16
URI:	http://repository.its.ac.id/id/eprint/124678

Actions (login required)

View Item