An adaptive compressor characteristic map method based on the Bézier curve

There is inevitably a performance deviation between an engine model and an actual engine that is influenced by unpredictable factors such as the unsuspected environmental conditions and the natural performance degradation in the process of use. Because the engine model precision largely depends on the accuracies of the component maps, it is possible to revise the engine model to determine a better trend for the engine performance from recorded measurements by adjusting the maps. This paper presents a new method for updating the variable geometry component maps of a variable cycle engine (VCE) by using a set of scaling factors estimated with the cubature Kalman filter (CKF). A mapping function is created between the scaling factors and the component characteristic scaling coefficients for the adjustments of the maps. The proposed method is applied to a VCE model according to the VCE benchmark steady-state performance data. The results show that the maximum simulation error of the engine steady-state model decreases from 5.33 to 0.93%, and the CKF-based adaptation method provides a much faster computing rate than the particle swarm optimization (PSO) based adaptation method, which verifies the effectiveness and engineering applicability of the variable geometry characteristic adaptive correction method.

There is inevitably a performance deviation between an engine model and an actual engine that is influenced by unpredictable factors such as the unsuspected environmental conditions and the natural performance degradation in the process of use. Because the engine model precision largely depends on the accuracies of the component maps, it is possible to revise the engine model to determine a better trend for the engine performance from recorded measurements by adjusting the maps. This paper presents a new method for updating the variable geometry component maps of a variable cycle engine (VCE) by using a set of scaling factors estimated with the cubature Kalman filter (CKF). A mapping function is created between the scaling factors and the component characteristic scaling coefficients for the adjustments of the maps. The proposed method is applied to a VCE model according to the VCE benchmark steady-state performance data. The results show that the maximum simulation error of the engine steady-state model decreases from 5.33 to 0.93%, and the CKF-based adaptation method provides a much faster computing rate than the particle swarm optimization (PSO) based adaptation method, which verifies the effectiveness and engineering applicability of the variable geometry characteristic adaptive correction method.

Adaptive modification of turbofan engine nonlinear model based on LSTM neural networks and hybrid optimization method

This paper describes a Genetic Algorithm (GA) convergence study for a highly multi-modal fitness function with non-ordered parameters. The measures of GA performance used are best single solution performance, effectiveness in finding the optimum and percentage of total search space (PTSS) covered. We developed several ways of adapting the crossover and mutation probabilities, and we compare the results of these methods with a canonical GA, a mutation-only GA, and the Srinivas' adaptive method. The results indicate that a large constant probability of crossover, regardless of the mutation method used does not provide high efficiency, for medium and large populations if covering a small PTSS. The most effective method while covering the smallest PTSS, is an adaptive mutation-only method. Our results suggest that when convergence speed is of utmost interest, for functions with non-ordered parameters mutation is more important than crossover despite massive multi-modality of the function optimized. Methods with adaptive crossover can, however, also give good results as long as mutation with a constant high probability is also performed.

In this paper, a new sensorless deadbeat control method is proposed. In the deadbeat method, the desired voltage is calculated via the model of the induction motor and inverter (prediction model). This voltage impels the motor to track the references of the torque and flux in the next control interval. Robustness is an important issue about the deadbeat method. Two new techniques are used to reach a robust speed-independent sensorless deadbeat method. A speed-independent model is sued for prediction. Therefore, the estimated speed will not be used in the prediction model. It will reduce the drift error problem. Also, a new adaptive predictive method is proposed for simultaneous estimation of the stator resistance and speed. Only direct-axis equation is used in the adaptive method. This will reduce the calculation burden. The new adaptive function is achieved via the Lyapunov technique. The stability of the multiple-input multiple-output system for simultaneous adaptation is analyzed for the gain design problem. Simulation and experimental results in wide range of speed are depicted in order to verify the proposed method.

A new method of fault diagnosis for aeroengines with dispersedly clumped gas path parameters

The dynamic characteristics of the turbofan engine vary greatly in the full flight envelope, which makes the problem of dynamic uncertainty and input uncertainty very prominent. This brings different degrees of performance impact to the engine control system and even makes it lose stability. This paper proposes an adaptive variable parameter control method for dealing with multivariable dynamic uncertainty and input uncertainty. In this paper, the dynamic uncertainty and input uncertainty are mathematically converted into standard matched uncertainty, which can be handled more conveniently. Firstly, in the state space model, for the case where the number of state variables is less than or equal to the number of input variables and the input matrix satisfies the full-rank condition of the row, the existence of the right pseudo-inverse matrix of the input matrix can be guaranteed. So the dynamic uncertainty can be separated from the system matrix, and the input uncertainty can be separated from the input matrix. Thus these uncertainties are equivalently transformed into parametric matched uncertainty. Then the matched uncertainty model with two vectors of bounded basis functions is established. Secondly, the Lyapunov quadratic function is constructed by the closed-loop tracking error vector and the adaptively adjustable control parameter estimation errors, and the Lyapunov stability constraint is considered. Then, under the premise of considering the dynamic characteristics of the actuator, an adaptive control algorithm for multivariable matched uncertainty model of turbofan engine is derived. Finally, ground and high altitude simulations are carried out on the dual-loop control system based on the nonlinear dynamic model of the turbofan engine. The results show that the control system has robust stability and anti-interference performance for dynamic uncertainty and input uncertainty of turbofan engine in the full flight envelope. The fan speed control loop basically achieves no static error tracking. The dynamic error of the core speed control loop is less than 0.6% and the steady state error is less than 0.05%. By introducing stronger parameter change rate information to the controller, its performance can be further improved, and the transient state control is more stable.

Accurate gas turbine modeling is crucial for performance evaluation, control optimization, and fault diagnosis, especially under off-design conditions or component degradation. Traditional thermodynamic models rely on complete compressor and turbine maps, but these are often proprietary, outdated, or incomplete, limiting their applicability. Hybrid approaches that integrate physical models with data-driven techniques offer a promising alternative but face challenges due to limited real-world data, difficulty in capturing nonlinear relationships in degraded component maps, and inadequacies in modeling spatial relationships in compressor maps. To address these challenges, this study proposes a novel framework that generates degraded compressor maps, samples operating curves from these maps, and trains a neural network to predict performance under varying conditions. A stage-stacking method with embedded decay factors creates diverse degraded maps for training, simulatingawide range of operating scenarios. A Spatial Transformer Network is then employed to model the affine transformation between generalized performance maps and degraded conditions, enabling adaptive adjustments to reflect real-world scenarios. The results of the numerical experiments demonstrate that the proposed framework effectively captures the relationship between operating lines and affine transformations, enhancing the adaptability of compressor maps. This approach shows significant potential for improving gas turbine monitoring and control. Future work will focus on improving training data with more accurate degradation models, exploring advanced neural network architectures, and validating the framework with real-world operational data.

Optimization configuration of gas path sensors using a hybrid method based on tabu search artificial bee colony and improved genetic algorithm in turbofan engine

The purposes of modeling and simulation of a turbojet engine are the steady state analysis and transient analysis. From the steady state analysis, which consists in the investigation of the operating, equilibrium regimes and it is based on appropriate modeling describing the operation of a turbojet engine at design and off-design regimes, results the performance analysis, concluded by the engine's operational maps (i.e. the altitude map, velocity map and speed map) and the engine's universal map. The mathematical model that allows the calculation of the design and off-design performances, in case of a single spool turbojet is detailed. An in house code was developed, its calibration was done for the J85 turbojet engine as the test case. The dynamic modeling of the turbojet engine is obtained from the energy balance equations for compressor, combustor and turbine, as the engine's main parts. The transient analysis, which is based on appropriate modeling of engine and its main parts, expresses the dynamic behavior of the turbojet engine, and further, provides details regarding the engine's control. The aim of the dynamic analysis is to determine a control program for the turbojet, based on the results provided by performance analysis. In case of the single-spool turbojet engine, with fixed nozzle geometry, the thrust is controlled by one parameter, which is the fuel flow rate. The design and management of the aircraft engine controls are based on the results of the transient analysis. The construction of the design model is complex, since it is based on both steady-state and transient analysis, further allowing the flight path cycle analysis and optimizations. This paper presents numerical simulations for a single-spool turbojet engine (J85 as test case), with appropriate modeling for steady-state and dynamic analysis.

Turbofan engine technology has evolved over several decades resulting in highly efficient and reliable propulsion systems for commercial airliners. Maintenance costs have decreased, but still represent a major part of the overall aircraft operating costs. To further minimize these, engine maintenance needs to be planned timely and strategically. Advanced diagnostics and health monitoring methods are being developed at KLM Engine Services (ES) to improve engine maintenance. For health monitoring, gas path analysis is used which requires accurate engine performance models. The modelling task is complicated further by the reduced number of measured gas path parameters with modern turbofan engines. This paper presents a systematic approach to overcome these complications. An engine model usually comprises of a cycle reference point, representing a design point such as maximum take-off thrust. As a first step, a genetic algorithm is used to determine the set of unknown cycle reference point parameters and component efficiencies best matching a set of known engine measurement data. Additionally, physical relations were used as constraints to compensate for the missing data. Off-design performance is calculated by solving a set of non-linear algebraic equations which depend on the unknown component performance maps. The customary method of deriving the performance maps by scaling similar maps at the cycle reference point only, often suffers from large deviations at off-design conditions. Consequently, these ‘baseline’ maps require corrections across the entire operating envelope. In the second step of the method, genetic algorithms determine the best off-design performance estimations at multiple measured operating points by finding the optimal coefficients of polynomial scaling functions for map parameters such as efficiency, corrected pressure ratio and mass flow. The modelling method has been verified by developing CF6-80C2 and GEnx-1B turbofan engine models using test cell data. The GEnx-1B engine model has subsequently been validated using on-wing operating data. The largest validation error was attained at cruise flight conditions and was found to be equal to 3.9%. The resulting method provides a systematic way to deal with missing data and can be used for developing accurate engine models for better gas path analysis reliability, resulting in more effective engine maintenance.

Automated repair processes and adaptive machining strategies constitute an important task in today’s aero-engine and industrial gas turbine maintenance, repair and overhaul (MRO) industry (figure 1). Currently, the repair of blisks is a central issue whenever consideration is given to replacing bladed stages with blisks; the feasibility of such a step hinges on the available capabilities for automated repair. The standard repairs are also influenced by these innovative approaches. Today, most of the processes for the MRO of engine components are carried out manually. In many cases, however, manual operations are not satisfactory from the point of view of costs and reliability. The MRO steps which are especially time consuming and require a high degree of accuracy are inspection, welding, milling and polishing. Adaptive machining methods can compensate for part-to-part variation as well as inaccurate clamping positions and keep the tolerances for the actual parts within a minimal range. The geometrical adaptation of the NC paths to the actual part geometry is performed automatically using in-process measuring techniques, mathematical best-fit strategies and adaptation methods. With the present state-of-the-art, it is possible to automate MRO work steps currently performed manually and to reduce costs and throughput times while boosting quality and precision. A further important aspect for the automation of component repair is the data management which should constitute the core of automated overhaul systems. As part of an innovative data management solution, the single repair process modules are integrated to build an automated repair cell for aero engine components. Furthermore, it is possible to establish “virtual” MRO workshops. The data management system generates a data set for each individual component and handles the logistics of the components and the accompanying data sets. As result, different MRO processes can be carried out at different facilities without loss of information, efficiency or quality. In addition, the approach described supports efficient life cycle monitoring.

This paper presents an adaptive causal model method (adaptive CMM) for fault diagnosis and recovery in complex multi-robot teams. We claim that a causal model approach is effective for anticipating and recovering from many types of robot team errors, presenting extensive experimental results to support this claim. To our knowledge, these results show the first, full implementation of a CMM on a large multi-robot team. However, because of the significant number of possible failure modes in a complex multi-robot application, and the difficulty in anticipating all possible failures in advance, our empirical results show that one cannot guarantee the generation of a complete a priori causal model that identifies and specifies all faults that may occur in the system. Instead, an adaptive method is needed to enable the robot team to use its experience to update and extend its causal model to enable the team, over time, to better recover from faults when they occur. We present our case-based learning approach, called LeaF (for Learning-based Fault diagnosis), that enables robot team members to adapt their causal models, thereby improving their ability to diagnose and recover from these faults over time.

Updating subsystem-level fault-symptom relationships for Temperature and Humidity Control Systems with redundant functions

At present, the main purpose of gas turbine fault prediction is to predict the performance decline trend of the whole system, but the quantitative and thorough performance health index (PHI) research of every major component is lacking. Regarding the issue above, a long-short term memory and gas path analysis (GPA) based gas turbine fault diagnosis and prognosis method is proposed, which realizes the coupling of fault diagnosis and prognosis process. The measurable gas path parameters (GPPs) and the health parameters (HP) of every main component of the goal engine are obtained through the adaptive modeling strategy and the gas path diagnosis method based on the thermodynamic model. The predictive model of the Long-Short Term Memory (LSTM) network combines the measurable GPPs and the diagnostic HPs to predict the HPs of each major component in the future. Simulation experiments show that the proposed method can effectively diagnose and predict detailed, quantified, and accurate PHIs of the main components. Among them, the maximum root mean square error (RMSE) of the diagnosed component HPs do not exceed 0.193%. The RMSE of the best prediction model is 0.232%, 0.029%, 0.069%, and 0.043% in the HP prediction results of each component, respectively.