Chen Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Chen Zhang has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
Fuxin Wang
4 shared publicationsHong Liu
4 shared publications- Hong Liu
Shanghai Jiao Tong University
3 shared publications - Jiaxin Yuan
College of Air Transportation, Shanghai University of Engineering Science, Shanghai 201620, China
2 shared publications Shufan Hu
Shanghai Jiao Tong University
2 shared publications- Tao Chen
China Agricultural University
2 shared publications
Latest Publications
- Open Access
- 6 October 2026
- Published by Elsevier in SSRN Electronic Journal
Comparative Analysis of Two-Dimensional Airfoil Ice Shape Prediction with Different Neural Network Architectures
Airfoil icing involves complicated multi-physics coupling. This study presents a comparative analysis of three neural network architectures—Multilayer Perceptron (MLP), Residual MLP (ResMLP), and Kolmogorov–Arnold Network (KAN)—for two-dimensional airfoil ice shape prediction. The three models are trained under identical conditions and comparable parameter budgets on sets of 2000, 5000, and 8000 samples, with dimensionality reduction applied to the high-dimensional ice displacement field to facilitate efficient regression. Their performance is systematically evaluated from three perspectives: prediction accuracy, training efficiency, and data sensitivity. The results show that at 2000 training samples, KAN achieves the highest prediction accuracy, followed by ResMLP and then MLP. As the training set increases to 5000 and 8000 samples, all three architectures improve and the performance gap narrows substantially; by 8000 samples, the three converge to comparable accuracy levels. KAN also requires more training iterations to converge under L-BFGS optimization, reflecting a trade-off between prediction accuracy and training speed. These findings provide preliminary guidance for architecture selection in data-driven icing prediction under different data availability conditions.
- 6 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessComparative Analysis of Two-Dimensional Airfoil Ice Shape Prediction with Different Neural Network Architectures
Airfoil icing involves complicated multi-physics coupling. This study presents a comparative analysis of three neural network architectures—Multilayer Perceptron (MLP), Residual MLP (ResMLP), and Kolmogorov–Arnold Network (KAN)—for two-dimensional airfoil ice shape prediction. The three models are trained under identical conditions and comparable parameter budgets on sets of 2000, 5000, and 8000 samples, with dimensionality reduction applied to the high-dimensional ice displacement field to facilitate efficient regression. Their performance is systematically evaluated from three perspectives: prediction accuracy, training efficiency, and data sensitivity. The results show that at 2000 training samples, KAN achieves the highest prediction accuracy, followed by ResMLP and then MLP. As the training set increases to 5000 and 8000 samples, all three architectures improve and the performance gap narrows substantially; by 8000 samples, the three converge to comparable accuracy levels. KAN also requires more training iterations to converge under L-BFGS optimization, reflecting a trade-off between prediction accuracy and training speed. These findings provide preliminary guidance for architecture selection in data-driven icing prediction under different data availability conditions.
- Open Access
Binary Icing Shapes Prediction via Principal Component Analysis and Deep Learning Method
Aircraft icing prediction is crucial for aerodynamic design and airworthiness assessment. Traditional physics-based models struggle with complex multi-physical processes, while existing AI methods (function-based characterization or direct image learning) face issues like multi-valued mapping, high data dependency, or lack of physical interpretability. This study proposes a deep learning framework based on point set displacement description, transforming the icing process into airfoil boundary point movements. PCA dimensionality reduction mitigates the curse of dimensionality while retaining physical meaning. A neural network is used to map environmental parameters to low-dimensional principal components. Comparative analysis shows the 64 × 64 network achieves optimal fitting; 2000 samples reproduce complex ice shapes, and 800 low samples characterize simple ones. Balancing efficiency, accuracy, and interpretability with reduced data dependency, this method provides a new approach for rapid engineering icing prediction.
Open AccessBinary Icing Shapes Prediction via Principal Component Analysis and Deep Learning Method
Aircraft icing prediction is crucial for aerodynamic design and airworthiness assessment. Traditional physics-based models struggle with complex multi-physical processes, while existing AI methods (function-based characterization or direct image learning) face issues like multi-valued mapping, high data dependency, or lack of physical interpretability. This study proposes a deep learning framework based on point set displacement description, transforming the icing process into airfoil boundary point movements. PCA dimensionality reduction mitigates the curse of dimensionality while retaining physical meaning. A neural network is used to map environmental parameters to low-dimensional principal components. Comparative analysis shows the 64 × 64 network achieves optimal fitting; 2000 samples reproduce complex ice shapes, and 800 low samples characterize simple ones. Balancing efficiency, accuracy, and interpretability with reduced data dependency, this method provides a new approach for rapid engineering icing prediction.
- Open Access
- 1 January 2026
- Published by National Library of Serbia in Thermal Science
- No. 00,p. 60
Effect of honeycomb structure layout on heat transfer performance of a high-temperature wind tunnel storage heater
To meet the demand for stable high-temperature air sources in hypersonic ground testing, this study employs three-dimensional unsteady conjugate heat transfer simulations to investigate the thermo-fluid performance of honeycomb regenerators with different pore densities under constant porosity and pitch-to-diameter ratio constraints. The results show that the high-density configuration exhibits the best thermal output and sustained heat-supply capability. At the end of 120 s of discharge, its outlet temperature is 74 K higher than that of the low-density configuration, while its outlet-temperature drop over 60-120 s is reduced by 39%. At 120 s, its solid-temperature non-uniformity decreases from 18% in the upstream section to 1.7% in the downstream section. Under high-flow conditions, the high-density configuration also shows lower sensitivity to flow-rate disturbances and achieves a 13% improvement in energy extraction efficiency. However, this thermal advantage is accompanied by a greater flow penalty, with an average resistance coefficient 2 times that of the low-density configuration.
- 1 January 2026
- Published by National Library of Serbia in Thermal Science
- No. 00,p. 60
Open AccessEffect of honeycomb structure layout on heat transfer performance of a high-temperature wind tunnel storage heater
To meet the demand for stable high-temperature air sources in hypersonic ground testing, this study employs three-dimensional unsteady conjugate heat transfer simulations to investigate the thermo-fluid performance of honeycomb regenerators with different pore densities under constant porosity and pitch-to-diameter ratio constraints. The results show that the high-density configuration exhibits the best thermal output and sustained heat-supply capability. At the end of 120 s of discharge, its outlet temperature is 74 K higher than that of the low-density configuration, while its outlet-temperature drop over 60-120 s is reduced by 39%. At 120 s, its solid-temperature non-uniformity decreases from 18% in the upstream section to 1.7% in the downstream section. Under high-flow conditions, the high-density configuration also shows lower sensitivity to flow-rate disturbances and achieves a 13% improvement in energy extraction efficiency. However, this thermal advantage is accompanied by a greater flow penalty, with an average resistance coefficient 2 times that of the low-density configuration.
- Open Access
- 7 May 2024
- Published by MDPI AG in Fractal and Fractional
- Vol. 8 (5) , 278
Adaptive Neural Control for a Class of Random Fractional-Order Multi-Agent Systems with Markov Jump Parameters and Full State Constraints
Based on an adaptive neural control scheme, this paper investigates the consensus problem of random Markov jump multi-agent systems with full state constraints. Each agent is described by the fractional-order random nonlinear uncertain system driven by random differential equations, where the random noise is the second-order stationary stochastic process. First, in order to deal with the unknown functions with Markov jump parameters, a radial basis function neural network (RBFNN) structure is introduced to achieve approximation. Second, for the purpose of keeping the agents’ states from violating the constraint boundary, the tan-type barrier Lyapunov function is employed. By using the stochastic stability theory and adopting the backstepping technique, a novel adaptive neural control design method is presented. Furthermore, to cope with the differential explosion problem in the design course, the extended state observer (ESO) is developed instead of neural network (NN) approximation or command filtering techniques. Finally, the exponentially noise-to-state stability in the mean square is analyzed rigorously by the Lyapunov method, which guarantees the consensus of the considered multi-agent systems and all the agents’ outputs are bounded in probability. Two simulation examples are provided to verify the effectiveness of the suggested control strategy.
- 7 May 2024
- Published by MDPI AG in Fractal and Fractional
- Vol. 8 (5) , 278
Open AccessAdaptive Neural Control for a Class of Random Fractional-Order Multi-Agent Systems with Markov Jump Parameters and Full State Constraints
Based on an adaptive neural control scheme, this paper investigates the consensus problem of random Markov jump multi-agent systems with full state constraints. Each agent is described by the fractional-order random nonlinear uncertain system driven by random differential equations, where the random noise is the second-order stationary stochastic process. First, in order to deal with the unknown functions with Markov jump parameters, a radial basis function neural network (RBFNN) structure is introduced to achieve approximation. Second, for the purpose of keeping the agents’ states from violating the constraint boundary, the tan-type barrier Lyapunov function is employed. By using the stochastic stability theory and adopting the backstepping technique, a novel adaptive neural control design method is presented. Furthermore, to cope with the differential explosion problem in the design course, the extended state observer (ESO) is developed instead of neural network (NN) approximation or command filtering techniques. Finally, the exponentially noise-to-state stability in the mean square is analyzed rigorously by the Lyapunov method, which guarantees the consensus of the considered multi-agent systems and all the agents’ outputs are bounded in probability. Two simulation examples are provided to verify the effectiveness of the suggested control strategy.
- 12 December 2023
- Published by Springer Nature
- p. 503-542
Numerical Simulation of Aerodynamic Features with Ice Shapes via High-Fidelity CFD Method
- 12 December 2023
- Published by Springer Nature
- p. 503-542
Numerical Simulation of Aerodynamic Features with Ice Shapes via High-Fidelity CFD Method
- 25 June 2023
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 48 (94) , 36972-36986
Influence of flow control schemes of the three-strut injector on the mixing efficiency and total pressure loss for a scramjet combustor
- 25 June 2023
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 48 (94) , 36972-36986
Influence of flow control schemes of the three-strut injector on the mixing efficiency and total pressure loss for a scramjet combustor
- 1 June 2023
- Published by Springer Nature
- p. 1-40
Numerical Simulation of Aerodynamic Features with Ice Shapes via High-Fidelity CFD Method
- 1 June 2023
- Published by Springer Nature
- p. 1-40
Numerical Simulation of Aerodynamic Features with Ice Shapes via High-Fidelity CFD Method
- Open Access
- 3 November 2022
- Published by MDPI AG in Fractal and Fractional
- Vol. 6 (11) , 642
Distributed Optimization for Fractional-Order Multi-Agent Systems Based on Adaptive Backstepping Dynamic Surface Control Technology
In this article, the distributed optimization problem is studied for a class of fractional-order nonlinear uncertain multi-agent systems (MASs) with unmeasured states. Each agent is represented through a system with unknown nonlinearities, unmeasurable states and a local objective function described by a quadratic polynomial function. A penalty function is constructed by a sum of local objective functions and integrating consensus conditions of the MASs. Radial basis function Neural-networks (RBFNNs) and Neural networks (NN) state observer are applied to approximate the unknown nonlinear dynamics and estimate unmeasured states, respectively. By combining the NN state observer and the penalty function, and the stability theory of the Lyapunov function, the distributed observer-based adaptive optimized backstepping dynamic surface control protocol is proposed to ensure the outputs of all agents asymptotically reach consensus to the optimal solution of the global objective function. Simulations demonstrate the effectiveness of the proposed control scheme.
- 3 November 2022
- Published by MDPI AG in Fractal and Fractional
- Vol. 6 (11) , 642
Open AccessDistributed Optimization for Fractional-Order Multi-Agent Systems Based on Adaptive Backstepping Dynamic Surface Control Technology
In this article, the distributed optimization problem is studied for a class of fractional-order nonlinear uncertain multi-agent systems (MASs) with unmeasured states. Each agent is represented through a system with unknown nonlinearities, unmeasurable states and a local objective function described by a quadratic polynomial function. A penalty function is constructed by a sum of local objective functions and integrating consensus conditions of the MASs. Radial basis function Neural-networks (RBFNNs) and Neural networks (NN) state observer are applied to approximate the unknown nonlinear dynamics and estimate unmeasured states, respectively. By combining the NN state observer and the penalty function, and the stability theory of the Lyapunov function, the distributed observer-based adaptive optimized backstepping dynamic surface control protocol is proposed to ensure the outputs of all agents asymptotically reach consensus to the optimal solution of the global objective function. Simulations demonstrate the effectiveness of the proposed control scheme.
- 23 March 2022
- Published by Elsevier in Computers & Fluids
- Vol. 240
High-fidelity modeling of turbulent shear flow downstream of a 3-D airfoil with spanwise ice contamination leading stall
- 23 March 2022
- Published by Elsevier in Computers & Fluids
- Vol. 240
High-fidelity modeling of turbulent shear flow downstream of a 3-D airfoil with spanwise ice contamination leading stall