Gas Path Performance Research Articles

Transient gas path fault diagnosis of aero-engine based on domain adaptive offline reinforcement learning

Real-time measurement parameters are crucial for diagnosing faults in aero-engine gas path performance, ensuring engine reliability, and mitigating potential economic losses. Traditional aero-engines performance diagnosis was mainly based on the measurements of steady-state condition and lacked the utilization of data under transient conditions. Gas path diagnosis of aero-engines under transient conditions is crucial for early fault detection and safety of flight within the envelope. The challenge lies in the inconsistent distribution of performance deviations caused by variable operating conditions, especially with complex fault types, which can undermine diagnostic credibility. To improve reliability of gas path diagnosis under transient conditions, an offline reinforcement learning fault diagnosis framework based on a transient aero-engine performance model is proposed. To address the issue of variable operating conditions during transient states, a domain adaptive approach is utilized to reconstruct the measurement baseline and facilitate the transfer of different performance deviation distributions. Additionally, by adding spool acceleration as a measurement parameter, the multi-component fault coupling is solved. Finally, validation with actual operating data simulates fault cases, demonstrating the proposed method's efficacy in quantitatively detecting gradual, sudden, and multiple component faults under transient conditions with high accuracy and efficiency. The method proposed in this study achieves a computational speed improvement by 64% compared to the conventional method, achieving a time of 0.13 seconds, with an average error of less than 0.00389%. Additionally, it demonstrates strong robustness in the presence of noise, with an average error of less than 0.03125%. This proposed method improves real-time fault detection under transient conditions for its higher accuracy and efficiency, and therefore significantly enhance gas path health monitoring and diagnosis capability.

A novel approach to aeroengine performance diagnosis based on physical model coupling data-driven using low-rank multimodal fusion method

Aeroengine health assessment plays a pivotal role in ensuring flight safety and reliability. Traditionally, this process involves diagnosing the performance of the aeroengine gas path. However, owing to the intricacies of operating conditions, non-linear performance, and the interplay of gas path performance fault characteristics, determining the aeroengine health condition directly from engine monitoring information poses a significant challenge, particularly in cases of insufficient sensor data. To address these challenges, a novel digital twin method for aeroengine performance diagnosis has been proposed. This method integrates data-driven and performance models, employing a low-rank multimodal fusion approach. By digitizing the physical system or process through mathematical models and simulation technology, this approach presents distinct advantages compared to previous methods relying solely on models or data. At the aeroengine component level, an adaptive model was implemented, and the data-driven model was constructed using flight data. Gas path fault classification employed support vector machines. The engine digital twin was established through low-order multimodal fusion. Results indicate that the proposed method attains excellent diagnostic accuracy under both steady and transient conditions. It can be harnessed to enhance engine performance monitoring and evaluation, thereby improving the reliability, availability, and efficiency of the engine.

Diagnosis of gas path health states on a real flight data by convolutional neural network and dual-tree complex wavelet

As the core component of aircraft, the reliability and availability of aero-engines directly affect the entire cruise process of aircraft. Implementing health monitoring in the gas path of an aero-engine can ensure engine reliability and availability to a greater extent. It can also enable the move from traditional time-scheduled maintenance to condition-based maintenance and reduce life cycle costs. The existing models were mainly concentrated on simulation data and traditional machine learning methods. This paper provides a new pathway to classify the states of engine gas path performance by combining the convolutional neural network (CNN) with the dual-tree complex wavelet (DTCW) method. A real flight data accompanied by maintenance records collected from the airline was adopted to construct a data set. The data pre-processing and stable points extraction were carried out, and the combined CNN and DTCW method was developed to diagnose the healthy states of the gas path. The final accuracy can achieve higher than 94%. The diagnosis results indicate that the deep learning methods hold particular promise for aero-engine health monitoring.

Knowledge and data jointly driven aeroengine gas path performance assessment method

Aeroengines, as the sole power source for aircraft, play a vital role in ensuring flight safety. The gas path, which represents the fundamental pathway for airflow within an aeroengine, directly impacts the aeroengine's performance, fuel efficiency, and safety. Therefore, timely and accurate evaluation of gas path performance is of paramount importance. This paper proposes a knowledge and data jointly driven aeroengine gas path performance assessment method, combining Fingerprint and gas path parameter deviation values. Firstly, Fingerprint is used to correct gas path parameter deviation values, eliminating parameter shifts caused by non-component performance degradation. Secondly, coarse errors are removed using the Romanovsky criterion for short-term data divided by an equal-length overlapping sliding window. Thirdly, an Ensemble Empirical Mode Decomposition and Non-Local Means (EEMD-NLM) filtering method is designed to “clean” data noise, completing the preprocessing for gas path parameter deviation values. Afterward, based on the characteristics of gas path parameter deviation values, a Dynamic Temporary Blended Network (DTBN) model is built to extract its temporal features, cascaded with Multi-Layer Perceptron (MLP), and combined with Fingerprint to construct a Dynamic Temporary Blended AutoEncoder (DTB-AutoEncoder). Eventually, by training this improved autoencoder, the aeroengine gas path multi-component performance assessment model is formed, which can sufficiently decouple the nonlinear mapping relationship between aeroengine gas path multi-component performance degradation and gas path parameter deviation values, thereby achieving the performance assessment of engine gas path components. Through practical application cases, the effectiveness of this model in assessing the aeroengine gas path multi-component performance is verified.

Aeroengine gas trajectory prediction using time-series analysis auto regressive integrated moving average

PurposeThis study aims to propose the use of time series autoregressive integrated moving average (ARIMA) models to predict gas path performance in aero engines. The gas path is a critical component of an aero engine and its performance is essential for safe and efficient operation of the engine.Design/methodology/approachThe study analyzes a data set of gas path performance parameters obtained from a fleet of aero engines. The data is preprocessed and then fitted to ARIMA models to predict the future values of the gas path performance parameters. The performance of the ARIMA models is evaluated using various statistical metrics such as mean absolute error, mean squared error and root mean squared error. The results show that the ARIMA models can accurately predict the gas path performance parameters in aero engines.FindingsThe proposed methodology can be used for real-time monitoring and controlling the gas path performance parameters in aero engines, which can improve the safety and efficiency of the engines. Both the Box-Ljung test and the residual analysis were used to demonstrate that the models for both time series were adequate.Research limitations/implicationsTo determine whether or not the two series were stationary, the Augmented Dickey–Fuller unit root test was used in this study. The first-order ARIMA models were selected based on the observed autocorrelation function and partial autocorrelation function.Originality/valueFurther, the authors find that the trend of predicted values and original values are similar and the error between them is small.

Gas path diagnosis method for gas turbine fusing performance analysis models and extreme learning machine

The gas path analysis, which can quantify the performance degradation of gas turbine components, has been extensively applied to the gas path diagnosis. However, the precondition of this method is that the number of measurable parameters for the gas turbine to be diagnosed should not be less than the number of its health factors. In the existing research, this precondition can be guaranteed through common approaches such as screening the degraded components by a model-based prediagnosis process or recognizing the degraded components by using tools such as an ANN or a support vector machine. However, the diagnosis speed, recognition accuracy, and robustness of these approaches need to be improved. Therefore, a diagnosis method fusing the gas path performance analysis model and the extreme learning machine was proposed in this paper and applied to a GE LM2500+SAC gas turbine. The working mechanism of similarity ranking-gas path diagnosis-rationality check was introduced in the fusion method, endowing it with a higher recognition accuracy rate, stronger robustness, and higher diagnostic accuracy.

Gas path fault diagnosis for gas turbine engines with fully operating regions using mode identification and model matching

Gas path fault diagnosis is key to improving the reliability and safety of gas turbine engines. Flexible operating conditions bring obstacles to performing accurate gas path performance analysis. Most of the existing methods are developed for specific operating conditions, which are difficult to adapt to fully operating regions. The operating mode identification and targeted diagnostic model matching are effective technologies to solve the gas path fault diagnosis under fully operating regions, which improves diagnostic accuracy and efficiency. The fully operating regions are classified into four typical operating modes, and the targeted diagnostic models are matched according to the mode features. For the typical start-stop state and high dynamic state, the small deviation diagnostic model and transient diagnostic model are established and verified by real fault cases. The small deviation diagnostic model based on boundary parameters reduces the influences of operating conditions on diagnostic results, it accurately monitors the health states. The transient diagnostic model driven by the dynamic model and a designed hybrid solution algorithm markedly improves diagnostic accuracy and efficiency. It shows better performance for the mixed gas path fault modes, more stable diagnostic results, and higher diagnostic efficiency. The proposed technical framework provides an effective way for the fault diagnosis of gas turbine engines under fully operating regions.

Gas path deterioration assessment for turbofan engines based on stochastic dynamics responses in the thermodynamic cycle

Aiming at the adverse effects of strong nonlinearity and high uncertainty existing in the gas path performance deterioration assessment of turbofan engines, in contrast to previously reported techniques on dealing with nonlinearity and uncertainty that this work proposes a new method for assessing the gas path performance based on the stochastic dynamics responses, which fully utilizes the synergy of nonlinearity and uncertainty to accurately identify health status. The nonlinear expressions of turbofan engines are derived from thermodynamic equations and component characteristics. The gas path deterioration assessment method is further established based on the stochastic dynamics, and the thrust and specific fuel consumption integrating the nonlinearity and uncertainty are correspondingly designed to quantify the performance deterioration degree. The influences of nonlinearity and uncertainty on the assessment results are then discussed, which proves the necessity of considering nonlinearity and uncertainty in the performance assessment. Finally, the effectiveness and superiority of the proposed method are verified by comparing with the traditional method under two performance deterioration cases. The results indicate that the proposed method enhances early weak deterioration features through the optimal matching among fuel flow rate, nonlinearity and uncertainty to accurately assess the gas path performance. The statistical accuracy of the proposed method is 94%, which describes performance evolution trends and provides a novel and effective way to synthetically deal with the nonlinearity and uncertainty in the performance assessment of turbofan engines.

Aero engine health monitoring, diagnostics and prognostics for condition-based maintenance: an overview

Abstract Aero engine performance deterioration highly influences its reliability, availability and life cycle. Predictive maintenance is therefore a key figure within Industry 4.0, which guarantees high availability and reduced downtime thus reduced operational costs for both military and civil engines. This leads to maintenance on demand and needs an effective engine health monitoring system. This paper overviews the work carried out on aero engine health monitoring, diagnostic and prognostic techniques based on gas path performance parameters. The inception of performance monitoring and its evolution over time, techniques used to establish a high-quality data base using engine model performance adaptation, and effects of computationally intelligent techniques on promoting the implementation of engine fault diagnosis are reviewed. Generating dependable information about the health condition of the engine is therefore a requisite for a successful implementation of condition-based maintenance. Based on this study, further research can be attempted to predict residual life of critical components using degradation pattern from aero engine performance data bank which will be an invaluable asset for engine designers as well as for operators.

Gas path parameter prediction of aero-engine based on an autoregressive discrete convolution sum process neural network

In order to improve the approximation ability of neural network to functional and improve the prediction accuracy of time series data, this paper proposes an autoregressive discrete convolution sum process neural network prediction model and applies it to the prediction of aero-engine gas path performance parameters. Coupling threshold wavelet de-noising method is applied to the preprocessing of engine gas path parameters, which can effectively remove the noise in the time series data. Process neurons are applied to artificial neural networks to expand the computational functions of artificial neurons. The process neuron can not only realize the spatial aggregation operation of discrete input data and the connection weight, but also realize the time aggregation operation of the product integral of the continuous input function and the connection weight function. The output of discrete time series is affected by the current input/output, and also by the historical input/output value. Therefore, an autoregressive feedback link is added to the network topology, and the value of the network's output node is used as the input to adjust the network connection weight. The convolution sum operation of discrete time series data can realize the time accumulation effect, so the discrete convolution sum operator is used to replace the integral operator to realize the time aggregation function of the process neuron. Since the integral operation of the continuous function is avoided, the weight training process of the process neural network is effectively simplified, and the accuracy loss in the continuous-time function fitting process of discrete input samples is effectively avoided. Bayesian regularization network weight learning algorithm is used to solve the problems of slow convergence speed and easy to fall into local optimum of back propagation learning algorithm, and improve the generalization ability of neural network. It can be seen from the prediction simulation results of the two engines that the average relative error and root mean square error of the engine gas path parameter prediction based on the autoregressive discrete convolution sum process neural network model can be reduced to 0.67% and 0.35 respectively. The stable and high-precision prediction results show that the network model and weight learning algorithm proposed in this paper have better robustness in functional approximation, and have higher accuracy in time series data prediction.

Online monitoring experiments of turbo-shaft engine based on electrostatic sensor

PurposeElectrostatic monitoring technology is a useful tool for monitoring and detecting component faults and degradation, which is necessary for engine health management. This paper aims to carry out online monitoring experiments of turbo-shaft engine to contribute to the practical application of electrostatic sensor in aero-engine.Design/methodology/approachCombined with the time and frequency domain methods of signal processing, the authors analyze the electrostatic signal from the short timescale and the long timescale.FindingsThe short timescale analysis verifies that electrostatic sensor is sensitive to the additional increased charged particles caused by abnormal conditions, which makes this technology to monitor typical failures in aero-engine gas path. The long scale analysis verifies the electrostatic sensor has the ability to monitor the degradation of the engine gas path performance, and water washing has a great impact on the electrostatic signal. The spectrum of the electrostatic signal contains not only the motion information of the charged particles but also the rotating speed information of the free turbine.Practical implicationsThe findings in this article prove the effectiveness of electrostatic monitoring and contribute to the application of this technology to aero-engine.Originality/valueThe research in this paper would be the foundation to achieve the application of the technology in aero-engine.

Aircraft engine hot-section virtual sensor creation and gas path performance monitoring

Various Kalman filter approaches have been presented for performance estimation of aircraft engines provided that sensor measurement numbers are sufficient and all of them are available. However, it is difficult to collect all physical parameters along the gas path since the complex structure limits sensor installation, especially around the high-pressure turbine. The contribution of this article is to create a virtual sensor in the hot-section based on the component physical characteristics and aerothermodynamic theory and develop a dual hypothesis test strategy combined with a state estimator to track abrupt degradation of the engine component. A novel couple algorithm of the unscented Kalman filter using a virtual sensor is developed that has three primary advantages: (i) The performance degradation of four rotatory components can be estimated without sensor P43. (ii) The accuracy of state estimation using the virtual sensor is equivalent to that of sensor P43, and it adapts to the abrupt change of engine operating condition in the whole flight envelope. (iii) If the sensor can be installed in the future, it can be engaged to its analytical redundancy. Simulations are carried out on typical abrupt degradation datasets of aircraft engines, which demonstrate the superiority of anomaly detection of gas path component performance using virtual sensor measurements, especially hot-section components.

Life cycle gas path performance monitoring with control loop parameters uncertainty for aeroengine

This paper addresses the extended Kalman filtering approach to aircraft engine gas path analysis with control loop parameter uncertainty, which results from faulty sensors and actuators. A novel approach of gas path performance anomaly detection is proposed during the life course, which relies on combinations of recovered and actual measurements. A couple of adaptive onboard engine models (AOBEMs) are designed independently tuned to the real engine with state buffers, and sensor and actuator fault isolation (SAFI) logic is developed using hybrid matrices of the AOBEM couple. Neighborhood measurements with prior thermodynamic analysis reconstruct the faulty sensor signal once the control loop fault is recognized by the SAFI logic. The contribution of this paper is to develop engine performance monitoring against uncertain parameters over time in a control loop, and the gas path fault false alarm rate is independent of sensor and actuator faults. The tests involve abrupt fault and degradation datasets of the aircraft engine life cycle in the flight envelope, and the results demonstrate the superiority of the proposed methodology in the uncertain engine control loop.

Sample adaptive aero-engine gas-path performance prognostic model modeling method

An accurate gas-path performance prognostic model would provide supports for aero-engine performance assessment, maintenance plan optimization, fleet management and operation schedule determination. Three important aspects deserve further considerations in aero-engine gas-path performance prediction: (1) time series characteristics, (2) operating conditions, (3) individual-differences among the aero-engines. To deal with the aforementioned three aspects, an aero-engine gas-path performance prognostic model based on long short-term memory (LSTM) network is established for each aero-engine, which is called Single-LSTM model. The established model is able to deal with the time series characteristics and operating conditions simultaneously. Furthermore, a unified and novel prognostic model, the so-called sample adaptive LSTM neural tree (SALNT), is investigated by combining LSTM and decision trees. The developed SALNT model achieves time series characteristics extraction and hierarchical structure analysis. In the SALNT, the sample can adaptively search for the best prognostic gas-path performance model according to its own characteristics. The real-life operation data of aero-engines are adopted to compare the developed Single-LSTM model and SALNT model with several typical prognostic models in short-term prediction. The experiments show that the developed prognostic models significantly improve the accuracy and stability in short-term prediction. The SALNT eliminates the influence of individual-differences among the aero-engines in short-term prediction. Moreover, the SALNT is applied to long-term prediction. The experiments results show that the SALNT is accurate in trend prediction of gas-path performance.

Aero-engine Gas Path Performance Degradation Assessment Based on a Multi-objective Optimized Discrete Feedback Network

In order to solve the problem of neural network algorithm for aero-engine’s gas path performance evaluation under high-dimensional evaluation index with non-equal weights, the trend analysis method and fault fingerprints are used to mine engine’s gas path performance characteristic parameters. A comprehensive weighting method based on game theory is proposed to optimize the weight value of each gas path performance characteristic parameter. A discrete feedback neural network with single-layer and binary output is established. The original gas path performance evaluation index is equivalently expanded according to the weight ratio, and the gas path state evaluation indexes with different weights are mapped into higher-dimensional equivalent evaluation indexes with equal weights. The network attractor is designed according to the engine gas path performance evaluation levels, and the design of discrete feedback neural network weights is transformed into multi-objective programming problem, and a particle swarm optimization algorithm with adaptive inertia weight is used to improve the efficiency and global search ability of particle swarm optimization. The experimental results shows that the proposed model and algorithm can provide a scientific and reasonable machine learning method for the evaluation of high-dimensional evaluation index with non-equal weights.

Regression model for civil aero-engine gas path parameter deviation based on deep domain-adaptation with Res-BP neural network

The variations in gas path parameter deviations can fully reflect the healthy state of aero-engine gas path components and units; therefore, airlines usually take them as key parameters for monitoring the aero-engine gas path performance state and conducting fault diagnosis. In the past, the airlines could not obtain deviations autonomously. At present, a data-driven method based on an aero-engine dataset with a large sample size can be utilized to obtain the deviations. However, it is still difficult to utilize aero-engine datasets with small sample sizes to establish regression models for deviations based on deep neural networks. To obtain monitoring autonomy of each aero-engine model, it is crucial to transfer and reuse the relevant knowledge of deviation modelling learned from different aero-engine models. This paper adopts the Residual-Back Propagation Neural Network (Res-BPNN) to deeply extract high-level features and stacks multi-layer Multi-Kernel Maximum Mean Discrepancy (MK-MMD) adaptation layers to map the extracted high-level features to the Reproduce Kernel Hilbert Space (RKHS) for discrepancy measurement. To further reduce the distribution discrepancy of each aero-engine model, the method of maximizing domain-confusion loss based on an adversarial mechanism is introduced to make the features learned from different domains as close as possible, and then the learned features can be confused. Through the above methods, domain-invariant features can be extracted, and the optimal adaptation effect can be achieved. Finally, the effectiveness of the proposed method is verified by using cruise data from different civil aero-engine models and compared with other transfer learning algorithms.

Hybrid State Estimation for Aircraft Engine Anomaly Detection and Fault Accommodation

This paper is concerned with a novel state estimator to track gas path performance in real time with one sensor failure and packet dropouts for aircraft engine in an advanced distributed architecture. It is common to sensor measurement lost in the distributed network, which results in the decrease of state tracking accuracy. A hybrid extended Kalman filter (HEKF) is proposed for engine performance anomaly detection and sensor fault accommodation from previous studies. Five groups of sensor measurement are divided along gas path related to five local filters, and the local estimated results are fused in the main filter. The reception state matrix is introduced to HEKF to deal with packet dropout, and nonlinear calculation is separated from a local filter to reduce computational burden in the field. Besides, fault diagnosis and isolation strategy of sensor subsets is developed and combined to HEKF by state consistency strategy of distributed network. The contribution of this study is to provide the novel HEKF algorithm to achieve real-time state estimation for sensor-fault-tolerant monitoring of aircraft engine with packet dropout in the distributed structure. The simulation and comparison are systematically carried out, and the superiority of the proposed methodology is confirmed.

A hybrid deep computation model for feature learning on aero-engine data: applications to fault detection

Recently, the safety of aircraft has attracted much attention with some crashes occurring. Gas-path faults, as the most common faults of aircraft, pose a vast challenge for the safety of aircraft because of the complexity of the aero-engine structure. In this article, a hybrid deep computation model is proposed to effectively detect gas-path faults on the basis of the performance data. In detail, to capture the local spatial features of the gas-path performance data, an unfully connected convolutional neural network of one-dimensional kernels is used. Furthermore, to model the temporal patterns hidden in the gas-path faults, a recurrent computation architecture is introduced. Finally, extensive experiments are conducted on real aero-engine data. The results show that the proposed model can outperform the models with which it is compared.

A multi-rate sensor fusion approach using information filters for estimating aero-engine performance degradation

Gas-path performance estimation plays an important role in aero-engine health management, and Kalman Filter (KF) is a well-known technique to estimate performance degradation. In previous studies, it is assumed that different kinds of sensors are with the same sampling rate, and they are used for state estimation by the KF simultaneously. However, it is hard to achieve state estimation using various kinds of sensor measurements at the same sampling rate due to a complex network and physical characteristic differences between sensors, especially in an advanced multi-sensor architecture. For this purpose, a multi-rate sensor fusion using the information filtering approach is proposed based on the square-root cubature rule, which is called Multi-rate Square-root Cubature Information Filter (MSCIF) to track engine performance degradation. Soft measurement synchronization of the MSCIF is designed to provide a sensor fusion condition for multiple sampling rates of measurement, and a fault sensor is isolated by maximum likelihood validation before state estimation. The contribution of this paper is to supply a novel multi-rate information filter approach for sensor fault tolerant health estimation of an aero-engine in a multi-sensor system. Tests are conducted for aero-engine performance degradation estimation with multiple sampling rates of sensor measurement on both digital simulation and semi-physical experiment. Experimental results illustrate the superiority of the proposed algorithm in terms of degradation estimation accuracy and robustness to sensor failure in a multi-sensor system.

Application of adaptive square root cubature Kalman filter in turbofan engine gas path performance monitoring

Kalman filters are very popular in turbofan engine community for health monitoring purposes. In this study, in order to get better performance in terms of filtering accuracy and noise adaptability, an Adaptive Square Root Cubature Kalman Filter (ASRCKF) was proposed to estimate health parameters of the turbofan engine gas path components. In the ASRCKF algorithm, the mean value and covariance of the engine nonlinear function were calculated by cubature rule-based numerical integration method and used as a substitute for nonlinear model in nonlinear Kalman filter. The latest information of measurement parameters in the recursion and filtering process was used to estimate and self-adjust the noises cross-covariance by removable window method to get higher filtering accuracy. Compared with the Extended Kalman Filter (EKF) and Square Root Cubature Kalman Filter (SRCKF), the simulation results in the gradual and rapid deterioration process of turbofan engine gas path components indicate that the higher accuracy and faster convergence can be obtained through ASRUKF, which can be used in health parameters estimation and condition monitoring of turbofan engine gas path.1