In an effort to better compare particular gas turbine diagnostic solutions and recommend the best solution, the software tool called Propulsion Diagnostic Method Evaluation Strategy (ProDiMES) has been developed. This benchmarking platform includes a simulator of the aircraft engine fleet with healthy and faulty engines. The platform presents a public approach, at which different investigators can verify and compare their algorithms for the diagnostic stages of feature extraction, fault detection, and fault identification. Using ProDiMES, some different diagnostic solutions have been compared so far. This study presents a new attempt to enhance a gas turbine diagnostic process. A data-driven algorithm that embraces the mentioned three diagnostic stages is verified on the basis of ProDiMES. At the feature extraction stage, this algorithm uses a polynomial model of an engine baseline to compute deviations of actual gas path measurements from the corresponding values of a healthy engine. At the fault detection and fault identification stages, a common classification for fault detection and fault identification is firstly constructed using deviation vectors (patterns). One of the three chosen pattern recognition techniques then performs both fault detection and fault identification as a common process. Numerous numerical experiments have been conducted to select the best configurations of the baseline model, a pertinent structure of the fault classification, and the best recognition technique. The experiments were accompanied by a computational precision analysis for each component of the proposed algorithm. The comparison of the final diagnostic ProDiMES metrics obtained under the selected optimal conditions with the metrics of other diagnostic solutions shows that the proposed algorithm is a promising tool for gas turbine monitoring systems.

No abstract

Gas turbine diagnostic algorithms widely use fault simulation schemes, in which measurement errors are usually given by theoretical random number distributions, like the Gaussian probability density function. The scatter of simulated noise is determined on the basis of known information on maximum errors for every sensor type. Such simulation differs from real diagnosis because instead of measurements themselves the diagnostic algorithms work with their deviations from an engine baseline. In addition to simulated measurement inaccuracy, the deviations computed for real data have other error components. In this way, simulated and real deviation errors differ by amplitude and distribution. As a result, simulation-based investigations might result in too optimistic conclusions on gas turbine diagnosis reliability. To understand error features, deviations of real measurements are analyzed in the present paper. To make error presentation more realistic, it is proposed to extract an error component from real deviations and to integrate it in fault description. Finally, the effect of the new noise representation mode on diagnostic reliability is estimated. It is shown that the reliability change due to inexact error simulation can be significant.

Zero-dimensional models based on the description of the thermo-gas-dynamic process are widely used in the design of engines and their control and diagnostic systems. The models are subjected to an identification procedure to bring their outputs as close as possible to experimental data and assess engine health. This paper aims to improve the stability of engine model identification when the number of measured parameters is small, and their measurement error is not negligible. The proposed method for the estimation of engine components’ parameters, based on multi-criteria identification, provides stable estimations and their confidence intervals within known measurement errors. A priori information about the engine, its parameters and performance is used directly in the regularized identification procedure. The mathematical basis for this approach is the fuzzy sets theory. Synthesis of objective functions and subsequent scalar convolutions of these functions are used to estimate gas-path components’ parameters. A comparison with traditional methods showed that the main advantage of the proposed approach is the high stability of estimation in the parametric uncertainty conditions. Regularization reduces scattering, excludes incorrect solutions that do not correspond to a priori assumptions and also helps to implement the gas path analysis with a limited number of measured parameters. The method can be used for matching thermodynamic models to experimental data, gas path analysis and adapting dynamic models to the needs of the engine control system.

In modern gas turbine health monitoring systems, the diagnostic algorithms based on gas path analysis may be considered as principal. They analyze gas path measured variables and are capable of identifying different faults and degradation mechanisms of gas turbine components (e.g. compressor, turbine, and combustor) as well as malfunctions of the measurement system itself. Gas path mathematical models are widely used in building fault classification required for diagnostics because faults rarely occur during field operation. In that case, model errors are transmitted to the model-based classification, which poses the problem of rendering the description of some classes more accurate using real data. This paper looks into the possibility of creating a mixed fault classification that incorporates both model-based and data-driven fault classes. Such a classification will combine a profound common diagnosis with a higher diagnostic accuracy for the data-driven classes. A gas turbine power plant for natural gas pumping has been chosen as a test case. Its real data with cycles of compressor fouling were used to form a data-driven class of the fouling. Preliminary qualitative analysis showed that these data allow creating a representative class of the fouling and that this class will be compatible with simulated fault classes. A diagnostic algorithm was created based on the proposed classification (real class of compressor fouling and simulated fault classes for other components) and artificial neural networks. The algorithm was subjected to statistical testing. As a result, probabilities of a correct diagnosis were determined. Different variations of the classification were considered and compared using these probabilities as criteria. The performed analysis has revealed no limitations for realizing a principle of the mixed classification in real monitoring systems.