This study investigates mechanism-based modeling and simulation of a single-shaft heavy-duty industrial gas turbine. Taking the PG9171E gas turbine as the case study, component-level steady-state and dynamic models are developed. The steady-state model is established using the constant mass flow (CMF) method. For dynamic modeling, both the CMF approach and the inter-component volume (ICV) approach are implemented to enable a comparative assessment of the two methods. On the basis of the steady-state model, an improved Dung Beetle Optimization (DBO) algorithm is proposed to perform model correction using measured operational data from the gas turbine. After model correction, the maximum relative error between the simulated results and the measured operating data is reduced to 1.01 & times; 10-5%. Following high-accuracy model correction, sensitivity analysis and a comparative dynamic study are conducted for the two dynamic modeling approaches. The results indicate that the most influential sensitivity parameter is the rotor rotational inertia, followed by the virtual volume of the combustor. Moreover, the primary discrepancy between the ICV and CMF approaches arises from differences in the operating trajectories on component characteristic maps. The ICV-based model exhibits a pronounced response lag; however, it requires less computational time than the CMF-based model, making it more suitable for rapid engineering simulation and practical applications.

This paper proposes a rapid fusion generation of sensor analytical redundancy for aero-engine in flight envelope. The proposed methodology consists of three key components. Firstly, a physical-based component-level model is introduced, from which onboard rapid sensor analytical redundancy models are developed by simplifying gas path calculations using measurable parameters and reducing the dimensionality of nonlinear equation sets. Secondly, to address the accuracy degradation of the shaft speed onboard rapid calculation model caused by the non-unique solution problem in characteristic maps, a shaft speed analytical redundancy deep learning network is designed based on the multi-head self-attention mechanism of the Transformer architecture. Furthermore, an integrated analytical redundancy strategy is proposed by incorporating a multi-objective fuzzy fusion algorithm. The main contributions of this work lie in the simplified model that significantly enhances real-time performance with minimal accuracy sacrifice, the deep neural network that effectively extends nonlinear representation capability and compensates for the rapid calculation model's accuracy, and the fusion strategy that comprehensively addresses the characteristics of the two methods under different operating conditions and sensor types, thereby improving accuracy and stability across the full flight envelope. Simulation results demonstrate the superior performance of the proposed methodology.

Prediction of turboshaft engine performance parameters is essential for engine health management, however, traditional physics-based models and shallow neural networks struggle to effectively model the complex nonlinear characteristics of time-series data. To overcome these limitations, a hybrid model integrating a Temporal Convolutional Network (TCN), an attention mechanism, and a Bidirectional Long Short-Term Memory (BiLSTM) network is proposed for predicting turboshaft engine performance parameters. Temporal convolution is employed to extract local temporal features, while the BiLSTM network is used to capture long-term bidirectional dependencies. An attention mechanism is further incorporated to assign greater weights to critical time steps. Simulation results show that the proposed model achieves higher prediction accuracy compared to standalone LSTM and TCN models. When applied to the performance prediction of turboshaft engines under different inlet air temperature conditions, the proposed framework consistently maintained the root mean square error values for training and testing between 0.05 and 0.09 across three key performance parameters, while effectively suppressing measurement noise. These results demonstrate that the model possesses excellent generalization performance and holds significant potential for practical engineering applications.

Turbofan engine performance gradually degrades during service due to fouling, erosion, and clearance growth, which alters the engine dynamics and input-output mapping. Under such conditions, existing advanced control methods still face practical challenges caused by the trade-off between modeling accuracy and onboard computational constraints, while conventional onboard proportional-integral (PI) controllers often suffer from degraded closed-loop tracking performance due to characteristic mismatch. To address this problem, this paper proposes a model-free adaptive discrete sliding-mode control method, referred to as FFDL-DSMC, for robust command tracking of degraded turbofan engines. The proposed method develops a degradation-oriented robust tracking framework by embedding full-form dynamic linearization (FFDL) and online pseudo-gradient (PG) estimation into a discrete sliding-mode control structure. In this framework, the online data model is used to characterize degradation-induced variations in engine dynamics, while a continuous reaching law is introduced to alleviate sliding-mode chattering. In addition, an adaptive weighting mechanism is incorporated to reduce controller conservatism during large-scale operating-condition transitions. The key controller parameters are optimized offline at the design point, resulting in a fixed parameter set used throughout the simulations. Simulation results demonstrate that the proposed method can maintain tracking performance comparable to that of onboard PI controllers while adaptively compensating for performance deviations caused by different degradation levels. Compared with conventional MFAC and FFDL-SMC, the proposed controller achieves faster response, smaller overshoot, and negligible steady-state error. Moreover, the maximum chattering amplitude of the controlled variable is reduced to 2.83% of that under FFDL-SMC. Across the tested degraded modes and operating conditions within the flight envelope, the proposed method maintains stable and consistent tracking performance for the controlled variables.

The fuel-servo actuator regulates fuel flow and nozzle area, and dual electrohydraulic servo valves usually provide redundancy for the aeroengine control system. However, time delay in the actuator during the main-backup switchover induces pulsating pressure fluctuations. These phenomena result in fuel-flow oscillations and engine speed ripple, which pose a potential threat to flight safety. Traditional controller methods are less capable of adapting to complex conditions, such as friction and zero bias of actuators, and they reduce the system's reliability. This article proposes a fault-tolerant control strategy of the main-backup switchover for fuel-servo control systems. First, a high-fidelity actuator model is built in AMESim to capture nonlinearities, where the typical fault scenarios are injected. Second, an active disturbance rejection control (ADRC) is designed to leverage an extended state observer to estimate the total disturbance from the gradual degradation of the actuator in real time. The channel switchover logic is triggered when the magnitude of degradation reaches a certain level. The time-window long short-term memory (LSTM) predictor is employed to forecast the transient differential-pressure trajectory. A control transformation then maps the predicted value to a feedforward compensation signal, and it produces a fault-tolerant control strategy combining linear ADRC feedback and LSTM feedforward. The proposed control algorithm and logic are conducted and integrated with the fuel-servo actuator system in the cosimulation of AMESim and Simulink to enhance the global robustness. The main contribution of this article is to launch a fault-tolerant control of the main-backup fuel-servo system, which is to weigh off the switchover disturbance suppression and tracking accuracy in the flight mission. Simulations show that the proposed method gains better control quality as the channel switchover in the typical actuator fault modes; the fluctuation magnitude is reduced by up to 64.73% . Even with moderate degradation, control accuracy improves by more than 62% compared with traditional methods.

Online gas path fault diagnosis of aeroengine transient performance is crucial in the field of the gas path fault diagnosis. Nevertheless, traditional methods present challenges for onboard applications due to their time-consuming, high memory usage and modeling errors. Therefore, an online gas path fault detection method especially for the transient process is proposed for aeroengines. Firstly, a combination filter named iB-EKF is presented, which consists of the extended Kalman filter (EKF) and the single-rank inverse Broyden (Rank-1 iB) algorithm. The Jacobi matrices of the component-level model (CLM) and EKF are both calculated by Rank-1 iB, and the iterations of CLM are replaced by EKF. The iB-EKF significantly reduces the call times to CLM, which reduces the fault detection computation time. Secondly, to reduce the impact of modelling error, the derivative of residuals plays the input of the iB-EKF module. The residual derivative tracking strategy (RDT) makes the derivatives of CLM outputs follow those of actual engine outputs. Thirdly, the data processing units are organized by a self-selecting fitting module and a curve fitting module to improve fault diagnostic stability. The function with the highest fitting accuracy is automatically matched, and it is introduced to adapt real engine outputs. Simulations show that the proposed method effectively diagnoses the gas path faults online with the modeling errors in transient processes. Compared to the traditional EKF, the computation time is reduced by more than 50% and the diagnosis accuracy is improved with the existence of modeling errors.

Adaptive cycle engine exhibits strong nonlinearity and multivariable coupling, making it difficult to meet coordinated control requirements in multiple modes. During performance degradation, control strategies often ignore the actual characteristics of the fuel system, leading to a disconnect between the control system and the actuators. This results in insufficient response speed and robustness. This paper proposes a data-driven fuzzy multivariable adaptive control algorithm. A compact form dynamic linearization (CFDL) model for ACE is established, with the pseudo-Jacobian matrix (PJM) updated dynamically. Then, a rule base is constructed for different degradation scenarios, including gas path faults and actuator degradation. Furthermore, a fuzzy PID controller is developed to achieve rapid nonlinear compensation. The contribution of this paper is to balance the dynamic response speed and control accuracy of the system in multiple work modes. The AMESim and MATLAB/Simulink co-simulation platform is used for verification under diverse operating modes and degradation scenarios. Simulation results show that the proposed controller outperforms conventional single control algorithms, it achieves a regulation time of less than 2.5 s and an overshoot below 0.5 %. Moreover, it exhibits faster response dynamics and smaller overshoot during performance degradation, thereby verifying the effectiveness and robustness of the presented method.

Monitoring the inter-turbine temperature T43 of turbofan engines is critical for performance assessment and safety margin management. Traditional physical sensors become unreliable under extreme operating conditions, while virtual sensor methods are prone to failure when faced with model mismatch and component degradation. This paper proposes a virtual sensor for T43 by integrating rotor inertia power balance (RPB) with a physics-informed neural network (PINN). First, based on engine thermodynamics and rotor dynamics, we extract rotor inertia power as a characteristic quantity and derive an RPB-based constraint that links measurable variables to T43. The derived constraint is then embedded into the PINN training objective. Automatic differentiation is used to compute the required derivatives, and an explicit constraint form is adopted to improve numerical stability and facilitate loss balancing between the data term and the physics term. Simulations under multiple turbine degradation scenarios show that the proposed method maintains stable accuracy compared with gas-path-based and purely data-driven baselines. In our setup, an intermediate physics weight provides a favorable trade-off between physical consistency and overall loss reduction. The proposed model also achieves shorter per-step prediction time while delivering robust T43 predictions across the operating envelope.

To address the challenges of limited access to full-life-cycle data and insufficient labeled samples in gas turbine health management, a Bidirectional Long Short-Term Memory-Domain Adversarial Neural Network (BiLSTM-DANN) is adopted to achieve cross-domain health assessment for gas turbines. The model extracts temporal health features with a two-layer BiLSTM network and integrates DANN to achieve cross-domain feature alignment, thereby learning domain-invariant health representations. The simulation results demonstrate that the BiLSTM-DANN model outperforms the traditional BiLSTM and DCNN models on both the FD001 and FD003 datasets of C-MAPSS. Health assessment tests conducted on real gas turbine operation datasets indicate that the BiLSTM-DANN model can effectively depict the long-term operational health evolution trend of the entire unit and accurately reflect the health changes of the gas turbine before and after water washing. Therefore, the method studied in this paper provides a transferable solution for assessing the health of the entire gas turbine under conditions of scarce labels.

The fuel servo actuator regulates fuel flow and nozzle area, and dual electrohydraulic servo valves usually provide redundancy for the aero-engine control system. However, time-delay in the actuator during the main-backup switchover induces pulsating pressure fluctuations. These phenomena result in fuel-flow oscillations and engine speed ripple, which pose a potential threat to flight safety. Traditional controller methods are less capable of adapting to complex conditions, such as friction and zero bias of actuators, and they reduce the system reliability. This paper proposes a fault tolerant control strategy of main-backup switchover for fuel servo control systems. Firstly, a high-fidelity actuator model is built in AMESim to capture nonlinearities, where the typical fault scenarios are injected into. Secondly, an active disturbance rejection control (ADRC) is designed to leverage an extended state observer to estimate the total disturbance from the gradual degradation of the actuator in real time. The channel switchover logic is triggered when the magnitude of degradation reaches a certain level. The time-window long short-term memory (LSTM) predictor is employed to forecast the transient differential pressure trajectory. A control transformation then maps the predicted value to a feedforward compensation signal, and it produces a fault tolerant control strategy combining linear ADRC feedback and LSTM feedforward. The proposed control algorithm and logic are conducted and integrated with the fuel servo actuator system in the co-simulation of AMESim and Simulink to enhance the global robustness. The main contribution of this paper is to launch a fault tolerant control of the main-backup fuel servo system, which is to weigh off the switchover disturbance suppression and tracking accuracy in the flight mission. Simulations show that the proposed method gains better control quality as the channel switchover in the typical actuator fault modes, the fluctuation magnitude is reduced by up to 64.73%. Even with moderate degradation, control accuracy improves by more than 62% compared with traditional methods.

Variable cycle engines (VCEs) represent a crucial development direction for future advanced fighter aircraft propulsion systems. The high-flow dual variable cycle engine (HDVCE), featuring a large flow capacity and triple-bypass configuration, represents a significant technical advancement. However, introducing this configuration presents thrust control challenges across broad operational envelopes. To address the variable thrust estimation requirements under HDVCE’s complex dynamics, this paper proposes a robust thrust estimation method that considers individual differences. First, neural networks compensate for measurement parameter deviations caused by individual differences. Second, a nonlinear filter utilizes the compensated measurements for state estimation. Finally, an intelligent fusion module generates high-confidence thrust estimation by integrating model-based and data-driven strategies. Comprehensive verification under fully digital and hardware-in-the-loop (HIL) conditions demonstrates that the proposed robust estimation method reduces the impact of engine individual differences on filter performance and ensures estimation accuracy at off-design conditions. Thrust estimation accuracy exceeds 98.5

With the increasing demand for flexible operation in modern power systems, gas turbines are frequently required to operate under wide load ranges and varying environmental conditions. In practical operation, compressor surge margin is strongly influenced by operating conditions and environmental disturbances, which makes accurate surge-margin regulation challenging. Consequently, industrial control systems typically adopt conservative anti-surge strategies with large safety margins. These strategies ensure operational safety while causing substantial energy efficiency loss. A hypernetwork-based Twin Delayed Deep Deterministic Policy Gradient (TD3-Hyper) control framework is proposed for active surge-margin regulation under varying operating conditions. The proposed method generates condition-specific policy parameters in a dynamic manner, which enables adaptive control across heterogeneous operating regimes. A dual-boundary safety–efficiency reward formulation is introduced for simultaneous enforcement of surge safety and energy efficiency requirements. A five-stage curriculum learning strategy is incorporated for improved training stability and convergence behavior. The overall framework improves controller adaptability under wide operating conditions while enhancing the generalization capability and training stability of hypernetwork-based reinforcement learning methods. Experimental results on a high-fidelity 300 MW gas turbine simulation platform show that TD3-Hyper reduces the root mean square error (RMSE) of speed tracking by 44.8% and decreases convergence iterations by 58% compared with conventional TD3. The proposed method achieves a 3.3% reduction in fuel consumption under variable-load operation. Robust performance is also observed under unseen operating conditions, together with consistent energy-saving behavior enabled by active surge-margin regulation, indicating its potential for safe and energy-efficient gas turbine operation.

A combined structure for autoupdating rotating component characteristic maps is proposed for an individual aeroengine model. The designed combined structure, known as the Newton-Raphson (NR)-particle swarm optimization (PSO) method, mainly composed of three parts, namely the equilibrium equations design, stability improved strategy, and reduced optimization logics. The equilibrium equations are designed to ensure the accuracy of the modified model. However, in some scenarios, the optimization process has poor stability. Therefore, a stability improved strategy is designed through limiting the range of values of the optimal parameters. In addition, reduced optimization logic is designed to reduce the computing time of the optimization algorithm. The NR-PSO method uses the Newton-Raphson method to increase the accuracy of model outputs. Meanwhile, the time consumed for optimization and the number of equilibrium equations are decreased by particle swarm optimization. The suggested method for automatic model correction has higher model output accuracy, quicker optimization speed, and stronger algorithm stability than particle swarm optimization. The simulation results showed the proposed method can transform the average performance model into the individual model matching the actual rig test data of an individual engine, and the maximum error of outputs of individual model are less than 1.5%.

The varying operational parameters and random noise make it difficult to determine the fault diagnosis thresholds for engine sensors under different working conditions. Therefore, an adaptive threshold-based fault diagnosis method for aeroengine sensors is proposed. A multivariable control system based on the MFAC method is established for the aeroengine. The OS-ELM algorithm employs historical sensor data to train and update the engine baseline model. MFAC dynamically establishes a linear model based on the pseudo-gradient change of control variables from the current sensor data and designs a baseline model tracker to calculate reasonable diagnostic thresholds based on historical sensor data characteristics, thereby improving the efficiency of threshold calculation and diagnostic accuracy. The experimental results validate that this method improves the fault detection rate by at least 30% while ensuring a low false alarm rate, reduces the minimum detectable fault magnitude by 39%, and keeps the fault detection time within 0.2 s.

A novel degradation estimation method is proposed to draw out the engine performance deterioration rules of multiple rotational-component parameters simultaneous variations from available sensors. The proposed methodology consists of the extended Kalman filtering (EKF) and an enhanced optimization strategy. Gas-path parameter analysis on the impact of components performance is implemented to determine the key degradation feature. The unmeasured feature is randomly generated to estimate performance drifts of rotational components by the EKF group and steady model. The achieved performance parameters stream to the engine dynamic model to yield measurements’ estimated series, which are utilized to construct the fitness function. Thus, the optimization gains the unmeasured feature and the deteriorations of engine performance. This paper reaches simultaneous performance drifts of all components under the limited sensors. Simulation results show the suggested approach provides the precise unmeasured feature, and quantify variations in the simultaneous drift of engine performance during its usage.

Aero-engine performance monitoring is a core component of the engine health management system and an important approach to enhancing flight safety and reliability. Meanwhile, to improve engine operation efficiency, control systems are evolving from traditional centralized architectures to distributed control architectures. To alleviate the negative impact of network uncertainties, this paper proposes a Distributed Adaptive Kalman Filter (DAKF), which resolves the estimation performance degradation of the classical Kalman Filter under network uncertainty by designing measurement reconstruction and buffer-based signal fusion strategies, expanding the engineering applicability of the Kalman Filter in distributed control architectures. Furthermore, a distributed hardware architecture was established based on the time-triggered protocol/class (TTP/C) bus protocol, communication programs between simulation nodes were developed, and the proposed DAKF algorithm was deployed in the hardware architecture for experimental validation. This study focuses on the steady-state operations of the turboshaft engine to investigate the performance of the proposed distributed Kalman Filter algorithm under network uncertainties. The results demonstrated the effectiveness of the proposed method, providing a basis for the engineering application of distributed performance monitoring methods.

Modeling the dynamic performance of modern aero-engines is crucial for ensuring the reliability and efficiency of control systems. However, traditional steady-state performance matching methods struggle to accommodate complex dynamic conditions, leading to significant modeling errors. This paper proposes a fast dynamic performance matching method based on an optimized combination strategy. First, the CN-TOPSIS method is employed to analyze the sensitivity of adaptive factors, enhancing numerical stability. Second, a self-adjusting damping coefficient is introduced to improve the stability and convergence of steady-state performance matching. Finally, by leveraging steady-state reference data, dynamic characteristic curves are constructed to achieve individualized dynamic performance adaptation for aero-engines. Experimental results demonstrate that the proposed method significantly improves computational efficiency while maintaining modeling accuracy, providing robust support for high-fidelity aero-engine modeling and performance correction.

Due to the unique propeller of turboprop engines and the conservative control strategies implemented for safety, traditional turboprop engines are unable to meet current control requirements. Hence, this article proposes a method for achieving multivariable performance recovery control of turboprop engines using a data-driven predictive model. The method designs a main controller for the turboprop engine based on the model-free adaptive predictive control algorithm, incorporating constraints into performance indices to realize overload protection for constrained parameters. The main contributions are as follows: 1) The integrated control of the propeller and engine is achieved by combining inner and outer loop control, reducing the pilot's operational burden. 2) The method uses a dynamically established linear model to transform the complex nonlinear constrained optimization problem into a quadratic programming problem, which ensures good real-time performance. 3) Under performance degradation, the predictive model ensures that constrained parameters remain within protection boundaries, simplifying the control system structure while effectively completing control tasks. Through turboprop engine simulation comparison experiments, the real-time performance, response to performance parameters, and overload protection capability of the controller were verified. The method demonstrates advantages in both control performance and time efficiency, providing an excellent alternative control solution in the case of engine performance degradation.

In order to investigate the effect of different oxygenation schemes on oxygen enrichment under different working conditions during the construction of the tunnel face of a large section tunnel on a plateau highway and to solve the quantitative problem of the oxygen enrichment model, a full-size numerical model of the tunnel face and a 1:30 experimental model were established for a plateau highway tunnel. These were used for numerical simulation and experimental research. The results indicate that the oxygen enriched area is piled up in the shape of a round bench when a construction bench is present, whereas in the absence of a construction bench, the oxygen enriched area assumes the shape of a slanting droplet. Furthermore, the oxygen enriched programs with oxygen delivery distances of 9 m, 8 m, and 7 m are unable to deliver the oxygen to the tunnel face. In the absence of a construction bench, the oxygenation effect is enhanced when the distance between the oxygen outlet and the tunnel face is 0.5 m. The oxygen released from different locations in multiple ports is distributed along the tunnel’s oxygen distribution range, which extends from three ports to two ports to one port. The number of oxygen ports affects the initial velocity of oxygen, which is susceptible to wind flow field influence. It is challenging to direct the flow of oxygen along the oxygen ports. The oxygen enriched volume of the different oxygen enriched programs in the tunnel under different working conditions can be quantified according to the multivariate linear regression oxygen enriched model, which can serve as a reference for optimizing oxygen enriched programs.

The development of the adaptive cycle engine is a crucial direction of advanced fighter power sources in the near future.However,this new technology brings more uncertainty to the design of the control system.To address the versatile thrust demand under complex dynamic characteristics of the adaptive cycle engine,this paper proposes a direct thrust estimation and control method based on the Model-Free Adaptive Control (MFAC) algorithm.First,an improved Sliding Mode Control-MFAC (SMC-MFAC) algorithm has been developed by introducing a sliding mode variable structure into the standard Full Format Dynamic Linearization-MFAC (FFDL-MFAC)and designing self-adaptive weight coefficients.Then a trivariate double-loop direct thrust control structure with a controller-based thrust estimator and an outer command compensation loop has been established.Through thrust feedback and command correction,accurate control under multi-mode and operation conditions is achieved.The main contribution of this paper is the improved algorithm that combines the tracking capability of the MFAC and the robustness of the SMC,thus enhancing the dynamic performance.Considering the requirements of the online thrust feedback,the designed MFAC-based thrust estimator significantly speeds up the calculation.Additionally,the proposed command correction module can achieve the adaptive thrust control without affecting the operation of the inner loop.Simulations and Hardware-in-Loop (HIL) experiments have been performed on an adaptive cycle engine component-level model to investigate the estimation and control effect under different modes and health conditions.The results demonstrate that both the thrust estimation precision and operation speed are significantly improved compared with Extended Kalman Filter (EKF).Furthermore,the system can accelerate the response of the controlled plant,reduce the overshoot,and realize the thrust recovery within the safety range when the engine encounters the degradation.