A framework with nonlinear system model and nonparametric noise for gas turbine degradation state estimation

Liu, Jie; Banerjee, Avisekh; Chen, Ying

doi:10.1088/0957-0233/26/6/065604

Cited by 26 publications

(13 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Along with the improvement of the component level model precision and the increase of the hardware calculation speed, the use of an engine model for the performance estimation and fault diagnosis has been a key concept in the engine health management [17][18][19]. Due to utilizing all model information available, the model-based approaches are more accurate and can be used to deal with faults that may be unknown.…”

Section: Introductionmentioning

confidence: 99%

Performance Estimation and Fault Diagnosis Based on Levenberg–Marquardt Algorithm for a Turbofan Engine

Lü

Huang

2018

Energies

View full text Add to dashboard Cite

Establishing the schemes of accurate and computationally efficient performance estimation and fault diagnosis for turbofan engines has become a new research focus and challenges. It is able to increase reliability and stability of turbofan engine and reduce the life cycle costs. Accurate estimation of turbofan engine performance counts on thoroughly understanding the components' performance, which is described by component characteristic maps and the fault of each component can be regarded as the change of characteristic maps. In this paper, a novel method based on a Levenberg-Marquardt (LM) algorithm is proposed to enhance the fidelity of the performance estimation and the credibility of the fault diagnosis for the turbofan engine. The presented method utilizes the LM algorithm to figure out the operating point in the characteristic maps, preparing for performance estimation and fault diagnosis. The accuracy of the proposed method is evaluated for estimating performance parameters in the transient case with Rayleigh process noise and Gaussian measurement noise. The comparison among the extended Kalman filter (EKF) method, the particle filter (PF) method and the proposed method is implemented in the abrupt fault case and the gradual degeneration case and it has been shown that the proposed method has the capability to lead to more accurate result for performance estimation and fault diagnosis of turbofan engine than current popular EKF and PF diagnosis methods.

show abstract

Section: Introductionmentioning

confidence: 99%

Performance Estimation and Fault Diagnosis Based on Levenberg–Marquardt Algorithm for a Turbofan Engine

Lü

Huang

2018

Energies

View full text Add to dashboard Cite

show abstract

“…PF is a numerical method for sequential state estimation that applies on Monte Carlo sequential method to estimate the posterior distribution of the evolving state by a set of sample scenarios, which are represented by a set of particles and their corresponding assigned weights [89]. Some PF variants have been developed to improve the filtering performance [90][91][92], and this scheme is finding increasing number of applications for state estimation in nonlinear/non-Gaussian systems, including degradation monitoring and prognostics in GTEs [93][94][95][96][97][98].…”

Section: Gte Fault Detection and Diagnosticsmentioning

confidence: 99%

“…As per section 5.2.1, three ambient conditions ( , , ) and three operating parameters ( , , ) are used to calculate the performance indicators at each time step. It was shown that the short-term performance deterioration is the dominant part of the performance indicator [97]. As a result, the indicators can be effectively used to quantify the compressor fouling, and the corresponding variation rates of the indicators will indicate the rate of fouling.…”

Section: Effective Factors On the Rate Of Performance Deteriorationmentioning

confidence: 99%

“…GTE operating systems commonly measure and monitor the power ( ), the rotational speed ( ) and the exhaust gas temperature ( ). Once the GTE model is constructed and calibrated with the design parameters ( ), can be predicted by the model, given the ambient condition and two of the three mentioned operating parameters [97]. Therefore, for a demanded power and controlled rotational speed ,…”

Section: Gas-path Model Of the Degrading Turbinementioning

confidence: 99%

See 1 more Smart Citation

Gas Turbine Engine Performance Estimation and Prediction

Hanachi

Self Cite

View full text Add to dashboard Cite

Modern health management approaches for gas turbine engines (GTE) aim to acquire precise information about the health state of the GTE components to optimize the maintenance decisions with respect to both the economy and safety. The task becomes more challenging for the GTE parts inaccessible to direct measurements with the available sensors of the GTE control system. This article-based thesis integrates a set of five coherent research work to address this problem.A detailed nonlinear thermodynamic model for single shaft GTEs is developed to predict the expected cycle parameters for the GTE in the healthy condition. In reality, the measured cycle parameters gradually deviate from the prediction due to performance deterioration. Physics-based performance indicators are defined based on the deviations in the measured performance parameters, compared to the respective model predictions.The indicators can effectively monitor the GTE performance deterioration in both shortterm and long-term regimes. In the next step, effect of the air humidity is taken into account to enhance the GTE model, and it is shown that the enhanced model can improve the performance monitoring by reducing the uncertainties.In order to separate the effects of different fault modes, an inference-based model is developed to predict the short-term recoverable performance deterioration due to the compressor fouling under different ambient and operating conditions. For the long-term non-recoverable performance deterioration due to the degradation mechanisms in the turbine hot section, two steps are undertaken; 1) a state estimation framework is developed for nonlinear/non-Gaussian systems with non-uniform time steps to track a degradation symptom of the turbine, i.e., loss of isentropic efficiency, using the observable performance indicators, and 2) the state estimation framework is extended to multidimensional dynamical systems with stochastic inputs for simultaneous tracking of two degradation symptoms, i.e., loss of isentropic efficiency and increase of the mass flow, using the observable parameters, provided by the GTE operating system.The developed techniques and frameworks are verified and validated, using a set of three-year operating data from an industrial GTE in a power plant. iii Preface This thesis is an integrated article-based thesis, consisted of the peer-reviewed journal articles and a conference paper, either published or currently under review. In all of the work presented henceforth, Houman Hanachi was the principal contributor in all areas of conceptualization, methodology development, implementation and validation, computer code writing, data preprocessing and analysis, as well as manuscript composition, including tables and figures. Professor Jie Liu provided the guidance and supervision of the overall project as well as timely and constructive feedbacks to all the research work. Dr. Avisekh Banerjee provided his ideas and guidance regarding the work as well as editorial suggestions. Dr. Ying Chen provided supportive explanatio...

show abstract

“…[55], which was obtained using a fixed five-day sampling period after each wash without distinguishing between operating conditions. was defined as the ratio of efficiency deviation to the actual efficiency, as in Equation (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16). With an extended sample size and condition-based sampling criteria, assessed by the present interfacing process showed a relatively modest and gradual growth trend compared to that by Hanachi et al The difference was likely due to the following two reasons: (1) compared to sampling periods greater than 10 days, a five-day sampling period was still insufficient at gathering enough data to form a clear distribution of and (2) without distinguishing between operating conditions, i.e., , the distribution of will be influenced by the result entries associated with a significantly lower than , resulting in differences in corrected rotor speed and lossto-roughness sensitivity.…”

Section: Establishment Of the "Actual-turbine" Performance Characterimentioning

confidence: 99%

Physics-Based Online Assessment of Surface Roughness Severity for High-Pressure Turbines

Li¹

View full text Add to dashboard Cite

Surface roughness is a critical indicator of the health of turbine blades, due to its implications on blade surface heat transfer and structural integrity. The present work proposes a physics-based online assessment framework for industrial gas turbine engines (GTE), in order to assess the blade surface roughness in a high-pressure turbine without engine shutdown. The framework consolidates gas path analysis (GPA) based performance monitoring models and meanline turbomachinery analysis using a novel GPA-meanline matching process. This extracts meaningful performance deviation trends from GPA, while resolving the uncertainties associated with the measurements and modeling. To relate efficiency loss to surface roughness severity, a meanline-based system-identification process has been developed to establish the meanline representation of the turbine stage and to incorporate an empirical surface roughness loss correlation system. The roughness loss correlations have been evaluated against recent transonic test data in the literature. A modification to the compressibility correction factor has been made according to the evaluation outcome, which yielded improved loss predictions compared to the experimental measurements.The successful application of the framework on three-year operational data of a cogenerative GTE validated the framework's potential as an effective online surface roughness monitoring tool. The estimated surface roughness index achieved magnitudelevel and trend agreement with the measurements reported in the literature.iii "No, in all these things we are more than conquerors through him who loved us." -Romans 8:37, ESV iv Acknowledgements I wish to thank my supervisor, Professor Jie Liu, for providing me the opportunity to work on this research project. This work could not have been accomplished without his encouragement and sound advice. I would also like to gratefully acknowledge my supervisor from my fourth-year capstone project, Professor Steen A. Sjolander, for introducing me to the field of turbomachinery. I am also truly grateful for his intermittent, but invaluable, guidance and encouragement over the course of this work.

show abstract

A framework with nonlinear system model and nonparametric noise for gas turbine degradation state estimation

Cited by 26 publications

References 28 publications

Performance Estimation and Fault Diagnosis Based on Levenberg–Marquardt Algorithm for a Turbofan Engine

Performance Estimation and Fault Diagnosis Based on Levenberg–Marquardt Algorithm for a Turbofan Engine

Gas Turbine Engine Performance Estimation and Prediction

Physics-Based Online Assessment of Surface Roughness Severity for High-Pressure Turbines

Contact Info

Product

Resources

About