Aero-Engine On-Board Dynamic Adaptive MGD Neural Network Model Within a Large Flight Envelope

Zheng, Qiangang; Zhang, Haibo; Li, Yongjin; Hu, Zhongzhi

doi:10.1109/access.2018.2789935

Cited by 26 publications

(12 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As shown in Figure 3, the DNN is a nonlinear mapping from multi-input to multi-output 21,23,24 and can be described as…”

Section: Linear Predictive Model Based On Ol-sw-dnnmentioning

confidence: 99%

“…The engine nonlinear model based CLM estimates the unmeasurable parameter and the engine sensors measure the measurable parameters. The OL-SW-DNN, 20 which has stronger fitting capacity than other shallow network structure such as traditional NN 21 and suppose vector regression, 22 is adopted to fitting the transient and steady process of engine. The SVM can be calculated through linearizing the OL-SW-DNN model and chosen as predictive model.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Aero-engine direct thrust control with nonlinear model predictive control based on linearized deep neural network predictor

Zheng

Yong

Sun

et al. 2019

Proceedings of the Institution of Mechanical Engineers, Part I:

Self Cite

View full text Add to dashboard Cite

A novel nonlinear model predictive control method for aero-engine direct thrust control is proposed to improve engine response ability and reduce computational complexity of nonlinear model predictive control. The control objective of the proposed method is the thrust directly instead of the measurable parameters. The linearized model based on online sliding window deep neural network is proposed as predictive model. The online sliding window deep neural network has strong fitting capacity for nonlinear object and adopted to fitting the transient process of engine. The back propagation is adopted to obtain linearized model of online sliding window deep neural network, which greatly reduce the calculated amount. The comparison simulations of the popular nonlinear model predictive control based on extended Kalman filter and the proposed one are carried out. The simulation results show that compared with the popular nonlinear model predictive control, the proposed nonlinear model predictive control not only has the better response ability but also has reduced computational complexity greatly, nearly reduce computation time more than 35 ms.

show abstract

“…As shown in Figure 3, the DNN is a nonlinear mapping from multi-input to multi-output 21,23,24 and can be described as…”

Section: Linear Predictive Model Based On Ol-sw-dnnmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Aero-engine direct thrust control with nonlinear model predictive control based on linearized deep neural network predictor

Zheng

Yong

Sun

et al. 2019

Proceedings of the Institution of Mechanical Engineers, Part I:

Self Cite

View full text Add to dashboard Cite

show abstract

“…The min-batch gradient descent (MBGD) method [18], [19] divides the training set data into M groups randomly, and each group has N b training sets.…”

Section: Acceleration Optimization Control For Turboshaft Enginementioning

confidence: 99%

A Study on the Acceleration Optimization Control Method for the Integrated Helicopter/Engine System Based on Torsional Vibration Suppression

et al. 2019

Self Cite

View full text Add to dashboard Cite

In order to solve the problem of insufficient power supply for helicopter and to increase the response speed of engine output power during the climbing process, the research on acceleration optimization control for the integrated helicopter/engine system is conducted. Meanwhile, the optimization control method with remarkable robustness for turboshaft engine based on torsional vibration suppression and min-batch gradient descent-neural network is proposed. The modified notch filter is available to suppress the torsional vibration from the frequency domain. Selecting the maximum output power as the optimization objective, the response speed of the engine output power can increase significantly through relaxing the restriction boundary of power turbine speed and taking into account the constraints of rotational speed, static strength, and temperature. The results show that the optimization control method can effectively reduce the low-order torsional amplitude by more than 70% and can decrease the response time of output power by more than 3s compared with the conventional cascade PID control, which allows the turboshaft engine to reach the potential on the premise of hardly exceeding the limit boundary.

show abstract

“…On the contrary, the CLM has high accuracy, but the model complexity of CLM is high. Therefore, some scholars proposed support vector regression (SVR) [25][26][27][28] and neural network (NN) 29,30 for modeling engine dynamics. These two methods have better accuracy than PLM and better real time than CLM.…”

Section: Introductionmentioning

confidence: 99%

Research on hybrid optimization and deep learning modeling method in the performance seeking control

Zheng

Chen

Wang

et al. 2020

Proceedings of the Institution of Mechanical Engineers, Part G:

Self Cite

View full text Add to dashboard Cite

A novel performance seeking control method based on hybrid optimization algorithm and deep learning modeling method is proposed to get a better engine performance. The deep learning modeling method, deep neural network, which has strong representation capability and can deal with big training data, is adopted to establish an on-board engine model. A hybrid optimization algorithm—genetic algorithm particle swarm optimization–feasible sequential quadratic programming—is proposed and applied to performance seeking control. The genetic algorithm particle swarm optimization–feasible sequential quadratic programming not only has the global search ability of genetic algorithm particle swarm optimization, but also has the high local search accuracy of feasible sequential quadratic programming. The final simulation experiments show that, compared with feasible sequential quadratic programming, genetic algorithm particle swarm optimization, and genetic algorithm, the proposed optimization algorithm can get more installed thrust, decrease fuel consumption between 2% to 3%, and decrease turbine blade temperature larger than 15k, while meeting all of the constraints. Moreover, it also shows that the proposed modeling method has high accuracy and real-time performance.

show abstract

Aero-Engine On-Board Dynamic Adaptive MGD Neural Network Model Within a Large Flight Envelope

Cited by 26 publications

References 29 publications

Aero-engine direct thrust control with nonlinear model predictive control based on linearized deep neural network predictor

Aero-engine direct thrust control with nonlinear model predictive control based on linearized deep neural network predictor

A Study on the Acceleration Optimization Control Method for the Integrated Helicopter/Engine System Based on Torsional Vibration Suppression

Research on hybrid optimization and deep learning modeling method in the performance seeking control

Contact Info

Product

Resources

About