Perancangan Sistem Kendali Dynamic Positioning pada Kapal Platform Supply Vessel Berbasis Integrasi Model Predictive Control dan Gaussian Process

Raditia, Raihan (2026) Perancangan Sistem Kendali Dynamic Positioning pada Kapal Platform Supply Vessel Berbasis Integrasi Model Predictive Control dan Gaussian Process. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009221059-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Kapal Platform Supply Vessel (PSV) yang beroperasi pada kegiatan lepas pantai memerlukan sistem kendali yang mampu mempertahankan posisi dan orientasi kapal secara akurat meskipun dipengaruhi oleh gangguan lingkungan dan ketidakpastian model. Model Predictive Control (MPC) dapat digunakan untuk memenuhi kebutuhan tersebut karena mampu memprediksi respons sistem dan menentukan aksi kendali melalui proses optimasi. Namun, performa MPC bergantung pada ketepatan model prediksi sehingga perbedaan antara model internal pengendali dan dinamika kapal dapat menimbulkan model mismatch. Penelitian ini bertujuan merancang sistem kendali Dynamic Positioning pada kapal PSV melalui integrasi MPC dan Gaussian Process (GP). Model gerak kapal dibatasi pada tiga derajat kebebasan, yaitu surge, sway, dan yaw. MPC dirancang menggunakan model linear hasil linearisasi dan diskritisasi, sedangkan model nonlinear tiga derajat kebebasan berbasis Marine Systems Simulator (MSS) digunakan sebagai plant simulasi. GP dilatih untuk mempelajari residual pada arah surge yang diperoleh dari selisih antara respons aktual plant nonlinear dan prediksi model linear. Dataset GP dibentuk dari tiga variasi kondisi awal dengan total 903 data mentah dan menghasilkan 900 pasangan transisi valid. Hasil evaluasi GP pada data uji memperoleh nilai RMSE sebesar 0,00023596 dan MAE sebesar 0,00013780. Hasil simulasi menunjukkan bahwa MPC mampu menjaga kestabilan posisi dan heading kapal pada seluruh variasi pengujian, tetapi masih menghasilkan steady-state error arah surge sekitar 0,1752 m. Setelah GP diintegrasikan, nilai tersebut menurun menjadi sekitar 0,1017–0,1025 m atau mengalami peningkatan akurasi sekitar 41,5–42,0%. Pada Variasi 2, integrasi GP juga membuat respons surge memenuhi kriteria settling time 5% dalam waktu 24,5 s, sedangkan sebelumnya kriteria tersebut belum terpenuhi. Integrasi GP tidak memberikan perubahan berarti pada respons sway dan heading karena kompensasi difokuskan pada arah surge. Hasil penelitian menunjukkan bahwa integrasi MPC-GP mampu mengurangi residual akibat model mismatch dan meningkatkan akurasi pengendalian posisi longitudinal tanpa mengurangi kestabilan sistem. Pengembangan metode kendali cerdas tersebut juga mendukung Sustainable Development Goals (SDGs) 9, yaitu Industry, Innovation and Infrastructure, melalui penerapan inovasi teknologi kendali untuk meningkatkan keandalan dan kinerja sistem operasi kapal lepas pantai.
==================================================================================================================================
A Platform Supply Vessel (PSV) operating in offshore activities requires a control system capable of accurately maintaining the vessel’s position and heading despite environmental disturbances and model uncertainties. Model Predictive Control (MPC) can be applied to address this requirement due to its ability to predict system responses and determine optimal control actions. However, MPC performance strongly depends on the accuracy of its prediction model; therefore, differences between the internal controller model and the actual vessel dynamics may result in model mismatch. This study aims to design a Dynamic Positioning control system for a PSV through the integration of MPC and Gaussian Process (GP). The vessel motion is limited to three degrees of freedom: surge, sway, and yaw. The MPC is designed using a linearized and discretized model, while a nonlinear three-degree-of-freedom model based on the Marine Systems Simulator (MSS) is used as the simulation plant. The GP is trained to learn the residual in the surge direction, calculated from the difference between the nonlinear plant response and the linear model prediction. The GP dataset is generated from three initial-condition variations, comprising 903 raw samples and 900 valid transition pairs. Evaluation on the test data produces an RMSE of 0.00023596 and an MAE of 0.00013780. Simulation results show that MPC maintains vessel position and heading stability under all test variations, although it produces a steady-state error of approximately 0.1752 m in the surge direction. Following GP integration, this value decreases to approximately 0.1017–0.1025 m, corresponding to an accuracy improvement of approximately 41.5–42.0%. In Variation 2, GP integration also enables the surge response to satisfy the 5% settling-time criterion within 24.5 s, whereas the criterion was not previously achieved. The GP produces no significant changes in the sway and heading responses because the compensation is specifically applied in the surge direction. These results demonstrate that MPC-GP integration can reduce residual errors caused by model mismatch and improve longitudinal position-control accuracy without compromising system stability. The development of this intelligent control method also supports Sustainable Development Goal 9, Industry, Innovation and Infrastructure, through the application of innovative control technology to improve the reliability and performance of offshore vessel operations.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Dynamic Positioning, Gaussian Process, Marine Systems Simulator, Model Predictive Control, Platform Supply Vessel, Sustainable Development Goals (SDGs)
Subjects:	T Technology > T Technology (General) T Technology > T Technology (General) > T57.62 Simulation T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Raihan Raditia
Date Deposited:	30 Jul 2026 02:01
Last Modified:	03 Aug 2026 02:01
URI:	http://repository.its.ac.id/id/eprint/139285

Actions (login required)

View Item