Xiaojian Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xiaojian Li has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Zhengxian Liu
Department of Mechanics, School of Mechanical Engineering, Tianjin University, Tianjin, China
12 shared publications Zhengxian Liu
10 shared publicationsYijia Zhao
School of Mechanical Engineering, Tianjin University of Commerce, Tianjin, China
9 shared publicationsMing Zhao
School of Mechanical Engineering, Tianjin University, Tianjin, China
5 shared publications- Ming Zhao
Tianjin University
5 shared publications - Lianchao Xu
Tianjin Univ, Sch Mech Engn, Dept Mech, Tianjin 300350, Peoples R China
4 shared publications
Latest Publications
- Open Access
- 2 June 2026
- Published by Elsevier in SSRN Electronic Journal
Optimization of Closure Coefficients for the Spalart-Allmaras Turbulence Model in Turbomachinery Flows
Spalart-Allmaras turbulence model is widely used in turbomachinery internal flow simulations. However, the baseline model often exhibits delayed response when confronted with extreme non-equilibrium flows such as shock wave and boundary-layer interaction, leading to insufficient accuracy in flow field prediction. This study proposes a multi-objective optimization method for turbulence model closure coefficients, driven by the prediction accuracy improvements of overall aerodynamic performances and local flow fields. Firstly, using the Rotor 37 transonic compressor as the test case, the prediction accuracy of the baseline Spalart-Allmaras model is evaluated. It is found that due to insufficient turbulence kinetic energy production, the baseline model triggers non-physical large-scale separation bubbles that induce channel blockage in the compressor passage, resulting in aerodynamic performances prediction errors. Secondly, to pursue better prediction accuracies of both aerodynamic performances and local flow fields, a multi-objective optimization is performed on six key closure coefficients of the Spalart-Allmaras model, by coupling optimal Latin hypercube sampling, a random forest surrogate model, and the non-dominated sorting genetic algorithm. After optimization, the total pressure ratio prediction error at the near-stall condition is reduced from 3.17% to 1.35%, and the trailing-edge relative Mach number error is reduced from 4.35% to 1.47%. Finally, the cross-speed-range and cross-configuration validations of the optimized turbulence model are conducted. It is found that the optimized model not only corrects the excessive dissipation problem in high-speed flows but also maintains compatibility with conventional subsonic flows, exhibiting favorable generalization capability and robustness.
- 2 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessOptimization of Closure Coefficients for the Spalart-Allmaras Turbulence Model in Turbomachinery Flows
Spalart-Allmaras turbulence model is widely used in turbomachinery internal flow simulations. However, the baseline model often exhibits delayed response when confronted with extreme non-equilibrium flows such as shock wave and boundary-layer interaction, leading to insufficient accuracy in flow field prediction. This study proposes a multi-objective optimization method for turbulence model closure coefficients, driven by the prediction accuracy improvements of overall aerodynamic performances and local flow fields. Firstly, using the Rotor 37 transonic compressor as the test case, the prediction accuracy of the baseline Spalart-Allmaras model is evaluated. It is found that due to insufficient turbulence kinetic energy production, the baseline model triggers non-physical large-scale separation bubbles that induce channel blockage in the compressor passage, resulting in aerodynamic performances prediction errors. Secondly, to pursue better prediction accuracies of both aerodynamic performances and local flow fields, a multi-objective optimization is performed on six key closure coefficients of the Spalart-Allmaras model, by coupling optimal Latin hypercube sampling, a random forest surrogate model, and the non-dominated sorting genetic algorithm. After optimization, the total pressure ratio prediction error at the near-stall condition is reduced from 3.17% to 1.35%, and the trailing-edge relative Mach number error is reduced from 4.35% to 1.47%. Finally, the cross-speed-range and cross-configuration validations of the optimized turbulence model are conducted. It is found that the optimized model not only corrects the excessive dissipation problem in high-speed flows but also maintains compatibility with conventional subsonic flows, exhibiting favorable generalization capability and robustness.
- 26 March 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 177
Energy loss analyses of the flow-dynamic modes in a compressor impeller
- 26 March 2026
- Published by Elsevier in Aerospace Science and Technology
- Vol. 177
Energy loss analyses of the flow-dynamic modes in a compressor impeller
- 2 February 2026
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science
- Vol. 240 (10) , 3469-3484
Design of a variable geometry turbine constrained by geometric and aerodynamic parameters to improve engine low-speed characteristics at high-altitude
High-altitude environment with low atmospheric pressure will significantly reduce the torque and power of diesel engines at low-speeds. To enhance the low-speed characteristics of an existing diesel engine at high-altitude of 4500 m, a variable nozzle is designed and integrated with the fixed-geometry turbine. Based on the mean-line thermodynamic calculations assisted by genetic algorithm optimization, the three-dimensional modeling and simulations, and the one-dimensional engine simulation, a design method for variable geometry turbine (VGT) under the constraints of geometric and aerodynamic parameters is established. Under the constraints of maintaining the prototype turbine volute geometry and achieving equivalent flow capacity and efficiency at the design rotating speed, the fixed-geometry turbine is modified to be a VGT. The newly designed VGT has increased the turbine efficiency by 2% at the design rotating speed, and expanded the flow adjustment range by 31%. Based on the calibrated one-dimensional simulation model of the diesel engine, the VGT is evaluated at high-altitude of 4500 m. Compared to the diesel engine with the fixed-geometry turbine, the engine with the VGT shows improvements in torque and power by 10.41%, 19.28%, and 23.28% at engine speeds of 800, 1000, and 1200 rpm respectively.
- 2 February 2026
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science
- Vol. 240 (10) , 3469-3484
Design of a variable geometry turbine constrained by geometric and aerodynamic parameters to improve engine low-speed characteristics at high-altitude
High-altitude environment with low atmospheric pressure will significantly reduce the torque and power of diesel engines at low-speeds. To enhance the low-speed characteristics of an existing diesel engine at high-altitude of 4500 m, a variable nozzle is designed and integrated with the fixed-geometry turbine. Based on the mean-line thermodynamic calculations assisted by genetic algorithm optimization, the three-dimensional modeling and simulations, and the one-dimensional engine simulation, a design method for variable geometry turbine (VGT) under the constraints of geometric and aerodynamic parameters is established. Under the constraints of maintaining the prototype turbine volute geometry and achieving equivalent flow capacity and efficiency at the design rotating speed, the fixed-geometry turbine is modified to be a VGT. The newly designed VGT has increased the turbine efficiency by 2% at the design rotating speed, and expanded the flow adjustment range by 31%. Based on the calibrated one-dimensional simulation model of the diesel engine, the VGT is evaluated at high-altitude of 4500 m. Compared to the diesel engine with the fixed-geometry turbine, the engine with the VGT shows improvements in torque and power by 10.41%, 19.28%, and 23.28% at engine speeds of 800, 1000, and 1200 rpm respectively.
- 18 August 2025
- Published by Springer Nature in International Journal of Aeronautical and Space Sciences
- Vol. 27 (2) , 1112-1129
A New Local-Loss-Coefficient Model for Energy Loss Analyses of Compressors Under Different Altitudes
- 18 August 2025
- Published by Springer Nature in International Journal of Aeronautical and Space Sciences
- Vol. 27 (2) , 1112-1129
A New Local-Loss-Coefficient Model for Energy Loss Analyses of Compressors Under Different Altitudes
- 9 August 2025
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering
- Vol. 240 (6) , 3407-3420
Structure optimization of turbocharger air-to-air intercooler for an air-cooled engine based on porous-media-model
Enhancing the flow and heat-transfer performances of the turbocharger intercooler can improve the overall performance of vehicle engines. Given the immense quantity of grids involved, numerically simulating the full-scale model of intercoolers becomes infeasible. Currently, most of the existing simulations and optimizations of intercoolers are based on the local fin model. A few studies try to employ the porous-media-model to investigate full-scale intercoolers, but they usually focus on water-to-air intercoolers. The applicability of the porous-media-model for full-scale air-to-air intercoolers simulation and optimization still needs further verification. Hence, it is urgent to explore efficient numerical simulation and optimization method for the full-scale air-to-air intercooler. Based on the porous-media-model, this study proposes an efficient numerical simulation method of the full-scale air-to-air intercooler. And it is validated by the air-cooled diesel engine bench test. The mean relative errors of the hot-side pressure-drop and the outlet temperature are 4.28% and 1.52% respectively. Then, on the basis of the efficient numerical simulation method, the optimal-Latin-hypercube sampling method, the general regression neural network (GRNN) surrogate model, and the multi-objective genetic algorithm (GA) are further combined to construct an optimization method of the full-scale intercooler. The optimization method is able to decrease the hot-side pressure-drop of the intercooler by 6.83% and maintain the heat-transfer performance. Finally, the flow mechanisms of performances improvement of the optimized intercooler are analyzed.
- 9 August 2025
- Published by SAGE Publications in Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering
- Vol. 240 (6) , 3407-3420
Structure optimization of turbocharger air-to-air intercooler for an air-cooled engine based on porous-media-model
Enhancing the flow and heat-transfer performances of the turbocharger intercooler can improve the overall performance of vehicle engines. Given the immense quantity of grids involved, numerically simulating the full-scale model of intercoolers becomes infeasible. Currently, most of the existing simulations and optimizations of intercoolers are based on the local fin model. A few studies try to employ the porous-media-model to investigate full-scale intercoolers, but they usually focus on water-to-air intercoolers. The applicability of the porous-media-model for full-scale air-to-air intercoolers simulation and optimization still needs further verification. Hence, it is urgent to explore efficient numerical simulation and optimization method for the full-scale air-to-air intercooler. Based on the porous-media-model, this study proposes an efficient numerical simulation method of the full-scale air-to-air intercooler. And it is validated by the air-cooled diesel engine bench test. The mean relative errors of the hot-side pressure-drop and the outlet temperature are 4.28% and 1.52% respectively. Then, on the basis of the efficient numerical simulation method, the optimal-Latin-hypercube sampling method, the general regression neural network (GRNN) surrogate model, and the multi-objective genetic algorithm (GA) are further combined to construct an optimization method of the full-scale intercooler. The optimization method is able to decrease the hot-side pressure-drop of the intercooler by 6.83% and maintain the heat-transfer performance. Finally, the flow mechanisms of performances improvement of the optimized intercooler are analyzed.
- 15 July 2025
- Published by Springer Nature in International Journal of Automotive Technology
- Vol. 27 (2) , 499-514
Parametric Modeling Method of Double-Channel Turbine Volute with Arbitrary Area Ratio
- 15 July 2025
- Published by Springer Nature in International Journal of Automotive Technology
- Vol. 27 (2) , 499-514
Parametric Modeling Method of Double-Channel Turbine Volute with Arbitrary Area Ratio
- 1 June 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (6)
Unsteady mechanisms of flow stability improvement for a compressor with self-circulating casing treatment
Self-circulating casing treatment (SRCT) is usually employed to expand the flow range of compressors, while the influencing mechanisms of SRCT on compressors are usually discussed from local and steady perspectives. In this study, the role of SRCT for a compressor is investigated in the perspective of global and unsteady features. The dynamic-mode-decomposition (DMD) method is employed to analyze the global unsteady features of the compressor with and without SRCT. The temporal and spatial characteristics of the DMD modes are explored to elucidate their physical meanings. The evolutions of the mode structures and mode energies during the compressor throttling process are studied, to explore the unsteady mechanisms of the stability enhancement of the SRCT. The following results are obtained. Under the design condition, due to the unsteady and non-uniform suction of the SRCT, the flow structure of the tip leakage vortex (TLV) in the SRCT model is altered, the circumferential uniformity is enhanced, and the unsteady pulsation intensity is weakened. Under the near-stall condition, the SRCT hinders the propagation and interference of upstream and downstream disturbances through unsteady suction, preventing the fragmentation of the TLV and maintaining the stability of the blade leading edge and the entire flow field. This study helps to obtain a deeper comprehension of the unsteady mechanisms of SRCT and provides new ideas for flow stability optimization of compressors by unsteady flow structures regulation.
- 1 June 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (6)
Unsteady mechanisms of flow stability improvement for a compressor with self-circulating casing treatment
Self-circulating casing treatment (SRCT) is usually employed to expand the flow range of compressors, while the influencing mechanisms of SRCT on compressors are usually discussed from local and steady perspectives. In this study, the role of SRCT for a compressor is investigated in the perspective of global and unsteady features. The dynamic-mode-decomposition (DMD) method is employed to analyze the global unsteady features of the compressor with and without SRCT. The temporal and spatial characteristics of the DMD modes are explored to elucidate their physical meanings. The evolutions of the mode structures and mode energies during the compressor throttling process are studied, to explore the unsteady mechanisms of the stability enhancement of the SRCT. The following results are obtained. Under the design condition, due to the unsteady and non-uniform suction of the SRCT, the flow structure of the tip leakage vortex (TLV) in the SRCT model is altered, the circumferential uniformity is enhanced, and the unsteady pulsation intensity is weakened. Under the near-stall condition, the SRCT hinders the propagation and interference of upstream and downstream disturbances through unsteady suction, preventing the fragmentation of the TLV and maintaining the stability of the blade leading edge and the entire flow field. This study helps to obtain a deeper comprehension of the unsteady mechanisms of SRCT and provides new ideas for flow stability optimization of compressors by unsteady flow structures regulation.
- 1 February 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (2)
A new composite neural network with spatiotemporal features extraction capability for unsteady flow fields predictions
Artificial intelligence based on neural network technology has provided innovative methods for predicting unsteady flow fields. However, both purely data-driven and single physics-driven methods can only perform short-term predictions for unsteady flow fields and are unable to achieve medium- to long-term predictions. A composite neural network CNN-GRU-PINN (CGPINN) is proposed by combining convolutional neural network (CNN), gated recurrent unit (GRU), and physics-informed neural network (PINN). CNN and GRU are used to learn the spatial and temporal characteristics of unsteady flows, respectively. PINN is adopted to constrain flow field prediction data according to physical laws. The flow around a circular cylinder is employed to verify the medium- and long-term prediction performances of the CGPINN. The test results show that compared to PINN, the reconstruction accuracy of the CGPINN is improved by about 86.10% on average, and the prediction accuracy is improved by about 96.18%. Compared to pure data-driven approaches, the prediction accuracy of the CGPINN is improved by an average of 65.71%. Additionally, CGPINN exhibits better robustness, demonstrating insensitivity to variations in sample size and noise levels, thereby ensuring stable and reliable performances across diverse data conditions. This study has provided a more accurate and robust method for the reconstruction and prediction of unsteady flow fields.
- 1 February 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (2)
A new composite neural network with spatiotemporal features extraction capability for unsteady flow fields predictions
Artificial intelligence based on neural network technology has provided innovative methods for predicting unsteady flow fields. However, both purely data-driven and single physics-driven methods can only perform short-term predictions for unsteady flow fields and are unable to achieve medium- to long-term predictions. A composite neural network CNN-GRU-PINN (CGPINN) is proposed by combining convolutional neural network (CNN), gated recurrent unit (GRU), and physics-informed neural network (PINN). CNN and GRU are used to learn the spatial and temporal characteristics of unsteady flows, respectively. PINN is adopted to constrain flow field prediction data according to physical laws. The flow around a circular cylinder is employed to verify the medium- and long-term prediction performances of the CGPINN. The test results show that compared to PINN, the reconstruction accuracy of the CGPINN is improved by about 86.10% on average, and the prediction accuracy is improved by about 96.18%. Compared to pure data-driven approaches, the prediction accuracy of the CGPINN is improved by an average of 65.71%. Additionally, CGPINN exhibits better robustness, demonstrating insensitivity to variations in sample size and noise levels, thereby ensuring stable and reliable performances across diverse data conditions. This study has provided a more accurate and robust method for the reconstruction and prediction of unsteady flow fields.
- 4 November 2024
- Published by Springer Nature in International Journal of Automotive Technology
- Vol. 26 (3) , 813-827
A Coupling Optimisation Strategy for Improving Flow Stability and Aerodynamic Performances of Turbocharger Centrifugal Compressors
- 4 November 2024
- Published by Springer Nature in International Journal of Automotive Technology
- Vol. 26 (3) , 813-827
A Coupling Optimisation Strategy for Improving Flow Stability and Aerodynamic Performances of Turbocharger Centrifugal Compressors