TY  - JOUR
UR  - https://www.scopus.com/inward/record.uri?eid=2-s2.0-85140723871&doi=10.1007%2f978-981-19-1939-8_38&partnerID=40&md5=e239adcb3753ed846417b4dc4a486940
JF  - Lecture Notes in Mechanical Engineering
A1  - Ahsan, S.
A1  - Alemu Lemma, T.
A1  - Baqir Hashmi, M.
A1  - Asmelash Gebremariam, M.
EP  - 501
Y1  - 2023///
N2  - The Original equipment manufacturers (OEM) advice to follow a fixed cycle for an overhaul which is required to recover the deteriorations.However, there is a need to investigate prognosis-based maintenance to find optimum time for maintenance activities.This paper proposes a simultaneous state and parameter estimation technique to evaluate the performance and prognostics of gas turbine using particle filter approach based on Bayesian inference framework and isentropic efficiency.Particle Filter is reported to better foresee health status of non-linear systems guaranteeing design efficiency, improved safety and reduced maintenance cost.The component level deterioration due to fouling is investigated for compressor and turbine section.The simulation data used for prognosis is generated from industry-standard Gas Turbine Simulation Program software developed by the Netherlands Aerospace Centre.The prognosis of gas turbine has been performed by using multi-sensory data in form of isentropic efficiency.The results of prognosis provided information about the accuracy and precision of the proposed approach with 95 and 90 confidence interval.The results obtained reveal a promising prospect of the proposed method for predicting remaining useful life of deteriorating gas turbines and can support in maintenance decision making for industries using gas turbines.A parametric study is also performed to see the effect of noise on prognosis accuracy.It is concluded that with increased noise in data, the proposed approach predicts pessimistic RUL. © 2023, Institute of Technology PETRONAS Sdn Bhd.
N1  - cited By 0; Conference of 7th International Conference on Production, Energy and Reliability, ICPER 2020 ; Conference Date: 14 July 2020 Through 16 July 2020; Conference Code:284729
KW  - Bayesian networks; Computer software; Decision making; Deterioration; Efficiency; Gas turbines; Gases; Inference engines; Linear systems; Systems engineering
KW  -  Amplitude modulated pseudo random signal; Amplitude-modulated; Linear filters; Non linear; Non-linear filter; Particle filter; Performance; Prognose; Pseudorandom signals; Transient operation
KW  -  Monte Carlo methods
TI  - Performance Prognostics of Gas Turbines Using Nonlinear Filter
SP  - 479
ID  - scholars19511
AV  - none
ER  -