Welcome To UTPedia

We would like to introduce you, the new knowledge repository product called UTPedia. The UTP Electronic and Digital Intellectual Asset. It stores digitized version of thesis, final year project reports and past year examination questions.

Browse content of UTPedia using Year, Subject, Department and Author and Search for required document using Searching facilities included in UTPedia. UTPedia with full text are accessible for all registered users, whereas only the physical information and metadata can be retrieved by public users. UTPedia collaborating and connecting peoples with university’s intellectual works from anywhere.

Disclaimer - Universiti Teknologi PETRONAS shall not be liable for any loss or damage caused by the usage of any information obtained from this web site.Best viewed using Mozilla Firefox 3 or IE 7 with resolution 1024 x 768.

Gas Turbine Health Prognostics Using Artificial Neural Network

Xin Wei, Yeap (2017) Gas Turbine Health Prognostics Using Artificial Neural Network. IRC, UniversitiTeknologi PETRONAS. (Submitted)

[img] PDF
Restricted to Registered users only

Download (6Mb)


The field of prognostics has gained the attention of companies in effort to reduce costs or losses by predicting the equipment’s future life. This project aims to develop a prognostics model that is capable of accurately predicting the remaining useful life (RUL) of gas turbines using an artificial neural network model. The model developed is to incorporate data fusion and can handle multiple-sensory signal input. Lastly, suitable application guidelines for the practical usage of the prognostics model is proposed. The Turbofan Engine Degradation Simulation Data Set published by Saxena and Goebel is used as the benchmark dataset in this research. In this research, the modified Wu’s method ANN model structure is tested in combination with wavelet-based denoising and exponential function fitting. The ANN model takes in input in the form of past two time-step parameters, and outputs the gas turbine’s life percentage, which can be further used to calculate the gas turbine RUL. This model has the advantage of being capable of handling multiple-sensory input at once The prediction results are evaluated using five parameters: total score, MAPE, MAE, RMSE, and correlation coefficient. The proposed ANN model has shown best results using original data only without denoising or exponential function fitting. Without any data filters, the ANN model scored 1,093 for the FD001 with a MAPE of 27%, MAE of 17, and RMSE of 23. Study from this research have also shown an increase in prediction accuracy when data fusion is incorporated in the form of increasing number of input variables. Analysis of results show that the ANN model has similar prediction effectiveness as other publications in terms of MAPE and MAE. However, the ANN model shows higher score value due to its bias of late predictions. Like other health-based prognostics models, this ANN model suffers from data insufficiency in practical applications due to the unknown current life percentage of the gas turbine.

Item Type: Final Year Project
Academic Subject : Academic Department - Mechanical Engineering - Materials - Engineering materials - Metals alloys - Fabrication
Subject: T Technology > TJ Mechanical engineering and machinery
Divisions: Engineering > Mechanical
Depositing User: Ahmad Suhairi Mohamed Lazim
Date Deposited: 01 Aug 2018 09:53
Last Modified: 20 Jun 2019 11:40
URI: http://utpedia.utp.edu.my/id/eprint/17977

Actions (login required)

View Item View Item

Document Downloads

More statistics for this item...