Pengendalian Kualitas Crude Palm Oil (CPO) di PT Perkebunan Nusantara IV PKS Sosa Menggunakan Diagram Kontrol Adaptive Multivariate Exponentially Weighted Moving Average (AMEWMA) Berbasis Residual Model Multioutput Least Square Support Vector Regression (MLS-SVR)

Harahap, Nur Farahizam Sari (2023) Pengendalian Kualitas Crude Palm Oil (CPO) di PT Perkebunan Nusantara IV PKS Sosa Menggunakan Diagram Kontrol Adaptive Multivariate Exponentially Weighted Moving Average (AMEWMA) Berbasis Residual Model Multioutput Least Square Support Vector Regression (MLS-SVR). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06211940007003_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (3MB) | Request a copy

Abstract

Crude Palm Oil atau disebut dengan minyak kelapa sawit merupakan bahan dasar pokok mentah yang diolah menjadi bahan baku minyak goreng. Salah satu perusahaan agribisnis yang bergerak fokus pada produksi Crude Palm Oil (CPO) yaitu PT Perkebunan Nusantara IV Sosa dengan kapasitas produksi sekitar 45ton TBS per jam, yang diolah menjadi Crude Palm Oil (CPO). Pada penelitian ini karakteristik kualitas yang digunakan yaitu kadar air, kadar kotoran dan asam lemak bebas yang masing-masing karakteristik kualitas memiliki hubungan. Hubungan pada ketiga variabel menunjukkan adanya kasus autokorelasi. Adanya autokorelasi menyebabkan false alarm dan mendapatkan pengambilan keputusan yang tidak tepat. Analisis yang digunakan yaitu dengan model algoritma Multi-output Least Square Support Vector Regression (MLS-SVR) untuk mengatasi data berautokorelasi dan dilakukan pengendalian menggunakan diagram kontrol AMEWMA. Input model MLS-SVR ditentukan berdasarkan lag yang signifikan dari plot PACF. Hasil input model MLS-SVR dan kombinasi hyper-parameter optimal yaitu γ'=2^(-5),γ''=2^(-10), dan σ=2^(-15) dengan nilai MSE minimum sebesar 0.958531 menghasilkan nilai residual yang dapat mengurangi autokorelasi. Nilai residual digunakan pada pembuatan diagram kontrol AMEWMA. Pada fase I dengan pembobot k=4, λ= 0,07 dan h*=0,3736 didapatkan diagram terkendali secara statistik setelah diatasi, dan pada fase II dengan menggunakan pembobot yang sama menunjukkan masih terdapat titik-titik yang keluar batas kontrol. Hasil identifikasi variabel penyebab out of control didapatkan variabel kadar air dan kadar asam lemak bebas yang memiliki faktor-faktor yang menjadi penyebab terjadinya out of control.
===============================================================================================================================
Crude Palm Oil or called palm oil is a basic raw material that is processed into cooking oil raw materials. One of the agribusiness companies that focuses on the production of Crude Palm Oil (CPO) is PT Perkebunan Nusantara IV Sosa with a production capacity of around 45ton FFB per hour, which is processed into Crude Palm Oil (CPO). In this study, the quality characteristics used are moisture content, dirt content and free fatty acids, each of which has a relationship. The relationship between the three variables shows a case of autocorrelation. The existence of autocorrelation causes false alarms and gets inappropriate decision making. The analysis used is the Multi-output Least Square Support Vector Regression (MLS-SVR) algorithm model to overcome autocorrelated data and control using AMEWMA control charts. The MLS-SVR model input is determined based on the significant lag of the PACF plot. The results of the MLS-SVR model input and the optimal hyper-parameter combination are γ'=2^(-5),γ''=2^(-10), and σ=2^(-15) with a minimum MSE value of 0.958531 produces residual values that can reduce autocorrelation. The residual value is used in the construction of the AMEWMA control diagram. In phase I with weight k = 4, λ = 0.07 and h * = 0.3736, the diagram is statistically controlled after being overcome, and in phase II using the same weight shows that there are still points that are out of control. The results of identifying the variables that cause out of control obtained moisture content and free fatty acid content variables that have factors that cause out of control.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	CPO, Autokorelasi, AMEWMA, MLS-SVR, Pengendalian kualitas CPO, Autocorrelation, Quality Control
Subjects:	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:	Nur Farahizam Sari Harahap
Date Deposited:	24 Aug 2023 03:32
Last Modified:	24 Aug 2023 03:32
URI:	http://repository.its.ac.id/id/eprint/104529

Actions (login required)

View Item