PREDICTIVE EMISSION MONITORING SYSTEM OF GAS TURBINE ON FLOATING PRODUCTION STORAGE AND OFFLOADING (FPSO)

Shodiq, Fatahillah Muhammad Daffa (2024) PREDICTIVE EMISSION MONITORING SYSTEM OF GAS TURBINE ON FLOATING PRODUCTION STORAGE AND OFFLOADING (FPSO). Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5019201159-Undergraduate_Thesis.pdf] Text
5019201159-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2026.

Download (3MB) | Request a copy

Abstract

This reseach introduces a novel Predictive Emission Monitoring System (PEMS) for Carbon Monoxide (CO) emissions in an auto-derivative gas turbine within a Floating
Production Storage and Offloading (FPSO) system. The PEMS utilizes a hybrid approach, integrating Artificial Neural Networks (ANN) with a Multi-Layer Perceptron (MLP) architecture and XGBoost for optimization. A thorough hyperparameter tuning process, using GridSearch, was performed to optimize both the ANN and XGBoost models. The ANN model was assessed with 20 parameter combinations, while the XGBoost model underwent fine tuning with an additional 20 combinations. The optimized model achieved a Root Mean Squared Error (RMSE) of 1.67, Mean Absolute Percentage Error (MAPE) of 1.88%, and an Rsquared (R²) of 0.94, highlighting its predictive accuracy. The research also examined the relationship between various engine parameters and CO emissions using visualization techniques like scatter plots, density plots, and box plots. Correlation heatmaps, SHAP(SHapley Additive exPlanations) values, and Permutation Feature Importance analysis identified N1 RPM (rotational speed) and compressor discharge pressure as the most influential parameters affecting CO emissions.

Item Type: Thesis (Other)
Uncontrolled Keywords: Predictive Emission Monitoring System, Artificial Neural Network, Machine Learning Predictive Emission Monitoring System, Artificial Neural Network, Machine Learning
Subjects: V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > Marine gas-turbines
V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering
V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM731 Marine Engines
Divisions: Faculty of Marine Technology (MARTECH) > Marine Engineering > 36202-(S1) Undergraduate Thesis
Depositing User: Fatahillah Muhammad Daffa Shodiq
Date Deposited: 05 Aug 2024 04:16
Last Modified: 05 Aug 2024 04:16
URI: http://repository.its.ac.id/id/eprint/112747

Actions (login required)

View Item View Item