1d-3d centripetal turbine collaborative optimization method

Abstract

The application discloses a 1D-3D coupling centripetal turbine multi-objective collaborative optimization method based on a proxy model, which comprises the following steps: step one: based on heat source conditions and operating conditions, determining an optimization objective function and constraint conditions, the optimization objective function involved including efficiency, wheel power and flow stability; step two: generating an initial parameter set covering thermodynamic parameters, aerodynamic parameters and structural parameters through a one-dimensional aerodynamic design model and a multi-objective optimization design function, and the application relates to the field of centripetal turbine optimization technology in the utilization of medium and low temperature heat energy sources. The 1D-3D centripetal turbine collaborative optimization method deeply fuses one-dimensional aerodynamic design and three-dimensional flow characteristic analysis, introduces an experimental design idea to establish a high-precision proxy model, realizes efficient collaborative optimization of thermodynamic parameters, kinetic parameters and structural parameters, and can shorten the optimization cycle by 40% and improve the design point efficiency by more than 2% in the design of ORC centripetal turbines.

Description

Technical Field

[0001] This invention relates to the field of centripetal turbine optimization design technology in the utilization of medium and low temperature thermal energy, specifically a 1D-3D centripetal turbine collaborative optimization method. Background Technology

[0002] With the continuous advancement of energy conservation and emission reduction efforts in my country, the utilization of low-temperature waste heat has gained attention. The organic Rankine cycle is one of the effective ways to generate electricity using low-temperature waste heat, possessing advantages such as simple and compact structure and low operation and maintenance costs. Abroad, this technology has been applied in the utilization of low-grade energy sources such as geothermal, solar, and biomass energy, generating significant economic benefits. Under the pressure of energy shortages, organic Rankine cycle power generation technology has received increasing attention from domestic scholars and industries.

[0003] Referring to the patent publication number "CN114398832B", a hierarchical collaborative optimization design method for an organic Rankine cycle centripetal turbine is disclosed, which includes the following steps: (1) According to the actual working conditions of the organic Rankine cycle, the heat source temperature, heat source flow rate, ambient temperature and circulating working fluid are input; (2) According to the one-dimensional aerodynamic design method of centripetal turbine, six parameters are selected as the design variables of turbine performance: turbine inlet pressure (Pt), superheat (ΔT), pressure ratio (ΠT), rotational speed (Ns), speed ratio (U), and reaction degree (Ω), and a mathematical model of turbine shaft efficiency and power is established.

[0004] As shown in the above technology, the connection between one-dimensional design and three-dimensional modeling in the existing technical solutions has limitations. It fails to fully realize the dynamic synergistic optimization of aerodynamic parameters and blade parameters, which may affect the adaptability and robustness of the optimization results in practical applications. In addition, the simulation of three-dimensional flow characteristics by the existing methods depends on the prediction accuracy of the BP neural network. Errors may be introduced due to insufficient data samples or insufficient model training, thereby affecting the final optimization effect. The existing centripetal turbine optimization design method still has room for improvement in terms of synergistic optimization of one-dimensional and three-dimensional models, dynamic adaptability, and prediction accuracy. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a 1D-3D approach that solves the problem that existing centripetal turbine co-optimization methods fail to fully realize the dynamic co-optimization of aerodynamic parameters and airfoil parameters.

[0006] To achieve the above objectives, the present invention provides a 1D-3D centripetal turbine collaborative optimization method comprising the following steps:

[0007] Step 1: Based on the heat source conditions and operating conditions, determine the optimization objective function and its constraints. The optimization objectives include efficiency, wheel power, and flow stability.

[0008] Step 2: Generate an initial parameter set covering thermodynamic, aerodynamic, and structural parameters through a one-dimensional aerodynamic design model and a multi-objective optimization design function. The thermodynamic parameters include inlet and outlet total pressure and static pressure and temperature; the aerodynamic parameters include reaction degree, specific speed, speed ratio and airflow angle; and the structural parameters include meridional channel profile and structural dimensions, blade installation angle and number of blades.

[0009] Step 3: Based on the initial aerodynamic parameter set, construct a three-dimensional numerical model, use computational fluid dynamics tools to perform full three-dimensional viscous flow numerical simulation, analyze velocity distribution, secondary flow and boundary layer separation law, evaluate the static pressure load distribution on the blade surface and the location of shock wave occurrence, and quantify the dissipation characteristics of tip leakage vortex and channel vortex loss source.

[0010] Step 4: Iteratively correct the one-dimensional structural parameters and dynamic parameter set based on the three-dimensional flow characteristic parameters to form the aerodynamic parameter set after 1D-3D collaborative design optimization;

[0011] Step 5: Input the optimized structural parameter set from the first round into the 3D airfoil parameter optimization module to generate an initial airfoil parameter set, which includes the meridional flow channel profile, blade mounting angle, blade curvature distribution, and thickness distribution.

[0012] Step 6: Reconstruct the three-dimensional numerical model based on the airfoil parameter set obtained from the 1D-3D co-optimization design, and perform full flow field numerical simulation again to extract new flow characteristic parameters;

[0013] Step 7: Based on the Latin hypercube experimental design, the airfoil parameter space is sampled, high-precision CFD numerical simulation is carried out and a sample database is constructed. The Kriging response surface is used to establish a 3D surrogate model for airfoil aerodynamic design, and the airfoil parameter set is iteratively optimized by combining optimization algorithms.

[0014] Step 8: Finally, output the aerodynamic parameter set and airfoil parameter set after multiple rounds of iteration and convergence, as the optimization design result.

[0015] Preferably, in the one-dimensional aerodynamic design model, the inlet and outlet boundary conditions of the centripetal turbine are defined, including total temperature, total pressure and mass flow rate. Based on the center streamline and loss model, and by integrating the Refprop real property database, the structural and dynamic parameters of the moving blades, guide vanes and volute are calculated.

[0016] Preferably, when constructing the three-dimensional numerical model, a three-dimensional geometric model is generated based on the structural parameter set output by the one-dimensional aerodynamic design model, including the blade passage, hub, and casing. The three-dimensional geometric model is meshed, with the blade surface area using a structured mesh and the remaining area using an unstructured mesh. Boundary conditions of the three-dimensional numerical model are set, including inlet total pressure, outlet static pressure, and wall no-slip condition. A suitable turbulence model is selected, and the k-ε two-equation model or the SST k-ω model is used for turbulence simulation.

[0017] In the mesh generation process of the three-dimensional numerical model, the cell size of the structured mesh and the cell size of the unstructured mesh are adaptively adjusted according to the changes in the flow field gradient to ensure the computational accuracy and efficiency of complex flow regions.

[0018] Preferably, in the post-processing stage of the calculation results of the three-dimensional numerical model, key flow characteristic parameters are extracted, including blade surface pressure distribution, secondary flow and flow separation location. Evaluation indicators of flow characteristic parameters are defined, including pressure distribution, airfoil loss and flow uniformity index. The rationality of the current aerodynamic parameter set is judged based on the changing trend of the evaluation indicators, and it is used as a feedback signal to be input into the one-dimensional aerodynamic design model for parameter correction.

[0019] Preferably, in the blade profile parameter optimization module, an initial search space for blade profile parameters is defined based on the aerodynamic parameter set output by the one-dimensional aerodynamic design model, including the meridional flow channel profile, the range of the leading and trailing edge mounting angles of the blade, and the blade profile thickness distribution. A genetic algorithm or particle swarm optimization algorithm is used to generate a candidate blade profile parameter set in the initial search space. The candidate blade profile parameter set is screened, and parameter combinations that do not meet the geometric constraints are eliminated. The screened blade profile parameter set is then input into a three-dimensional numerical model for numerical simulation to evaluate its flow characteristic parameters.

[0020] In the blade profile parameter optimization module, during the generation of the candidate blade profile parameter set, the Latin hypercube sampling method is used to uniformly sample within the initial search space and establish a three-dimensional optimization surrogate model to improve the coverage of the parameter space and the optimization efficiency.

[0021] Preferably, during the iterative optimization process, an optimization objective function is defined, including isentropic efficiency, flow stability index, and circumferential power. Optimization termination conditions are set, including the objective function's rate of change being less than a preset threshold or reaching the maximum number of iterations. In each iteration, the objective function is calculated based on the flow characteristic parameters of the current airfoil parameter set. The search direction of the airfoil parameter set is updated based on the objective function, and a new round of candidate airfoil parameter sets is generated.

[0022] Preferably, in the selection and implementation of the optimization algorithm, the adaptive weighted particle swarm optimization algorithm is adopted, which dynamically adjusts the inertia weight and acceleration factor to improve the convergence speed. In each iteration, the current optimal solution and its corresponding flow characteristic parameters are recorded. A local search strategy is introduced to conduct a fine search near the global optimal solution to avoid getting trapped in local optima. Parallel computing technology is used to accelerate the optimization process and reduce the overall computation time.

[0023] Preferably, during the verification and evaluation of the optimization results, the final optimized aerodynamic parameter set and airfoil parameter set are input into an independent three-dimensional numerical model for verification calculation. The verification calculation results are compared with the flow characteristic parameters in the optimization process to evaluate the consistency of the calculation results. The optimization results are further verified through experimental tests, including efficiency tests, pressure distribution tests, and vibration characteristic tests. The optimization method is improved based on the experimental test results to form a closed-loop optimization process. Beneficial effects

[0024] This invention provides a 1D-3D centripetal turbine collaborative optimization method. Compared with existing technologies, it has the following advantages:

[0025] 1. The 1D-3D centripetal turbine collaborative optimization method achieves efficient collaborative optimization of thermodynamic parameters, structural parameters and dynamic parameters by deeply integrating one-dimensional aerodynamic design with three-dimensional flow characteristic analysis. In the optimization process, a multi-scale modeling mechanism and an adaptive boundary condition adjustment strategy are introduced, which effectively improves the adaptability and robustness of the optimization results.

[0026] 2. This 1D-3D centripetal turbine collaborative optimization method establishes Kriging response surface fitting based on Latin hypercube sampling. Through efficient space-filling sampling and the construction of a high-precision surrogate model, it significantly improves the computational efficiency and design accuracy of complex numerical simulation optimization such as turbines. While reducing simulation costs, it supports multi-objective, high-dimensional parameter and uncertainty quantification analysis.

[0027] 3. This 1D-3D centripetal turbine collaborative optimization method ensures the efficiency and accuracy of the optimization process through the selection and implementation of optimization algorithms, providing strong technical support for the design of centripetal turbines.

[0028] 4. The 1D-3D centripetal turbine collaborative optimization method uses an adaptive weighted particle swarm optimization algorithm to dynamically adjust the inertia weight and acceleration factor to improve the convergence speed. In each iteration, the current optimal solution and its corresponding flow characteristic parameters are recorded. A local search strategy is introduced to perform a fine search near the global optimal solution to avoid getting trapped in local optima. Parallel computing technology is used to accelerate the optimization process and reduce the overall computation time. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the overall process of the method of the present invention;

[0030] Figure 2 This is a schematic diagram of the mesh generation for the three-dimensional numerical model of the present invention;

[0031] Figure 3 This is a flowchart of the airfoil parameter optimization module of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] Please see Figure 1-3 This invention provides a 1D-3D centripetal turbine collaborative optimization method. In actual implementation, it is first necessary to clarify the connection relationship and cooperation mode between the modules. The entire optimization process consists of a one-dimensional aerodynamic design model, a three-dimensional numerical model, and an airfoil parameter optimization module. These three modules achieve dynamic collaborative optimization through data transmission and feedback mechanisms. The initial aerodynamic parameter set generated by the one-dimensional aerodynamic design model is passed as input to the three-dimensional numerical model. The three-dimensional numerical model extracts flow characteristic parameters through numerical simulation and then feeds them back to the one-dimensional aerodynamic design model for correction. Finally, the airfoil parameter optimization module completes the iterative optimization.

[0034] First, based on the heat source conditions and operating conditions, the optimization objective function and its constraints are determined. The optimization objective function includes efficiency, power density, and flow stability. These objectives are defined as a global objective function through weighted synthesis. The constraints involve key parameters such as total pressure ratio, flow coefficient, and speed range. Based on this, a one-dimensional aerodynamic design model is used to generate an initial parameter set covering thermodynamic, aerodynamic, and structural parameters. Its core lies in solving the one-dimensional flow equations using the finite difference method. Specifically, the inlet and outlet boundary conditions of the centripetal turbine are defined, including total temperature, total pressure, and mass flow rate. A one-dimensional flow equation set including continuity, momentum, and energy equations is established. By solving this equation set, the pressure, velocity, and temperature distribution along the flow direction are obtained, and preliminary calculations are performed. The axial, radial, and tangential velocity distributions were determined. Based on the center streamline and loss model, and integrating the Refprop real property database, the structural and dynamic parameters of the blade, guide vane, and volute were calculated. Based on the initial aerodynamic parameter set, a three-dimensional numerical model was constructed. Computational fluid dynamics tools were used to perform full three-dimensional viscous flow numerical simulation, analyzing velocity distribution, secondary flow, and boundary layer separation. The distribution of hydrostatic load on the blade surface and the location of shock waves were evaluated. The dissipation characteristics of loss sources such as tip leakage vortices and channel vortices were quantified. The one-dimensional structural and dynamic parameter sets were iteratively corrected based on the three-dimensional flow characteristic parameters to form a 1D-3D co-design optimized aerodynamic parameter set. These parameters constitute the initial aerodynamic parameter set, providing the foundation for the subsequent construction of the three-dimensional numerical model.

[0035] The construction process of the three-dimensional numerical model is based on the aerodynamic parameter set output by the one-dimensional aerodynamic design model to generate a three-dimensional geometric model, including the blade passage, hub, and casing. In the mesh generation stage, a hybrid mesh strategy is adopted to ensure computational accuracy and efficiency. Specifically, the blade surface area uses a structured mesh, while the remaining areas use an unstructured mesh. This mesh generation method can effectively capture the complex flow characteristics of the blade surface while reducing the consumption of computational resources. Subsequently, the boundary conditions of the three-dimensional numerical model are set, including the inlet total pressure, outlet static pressure, and wall no-slip condition. An appropriate turbulence model is selected for turbulence simulation. In this embodiment, the k-ε two-equation model or the SST k-ω model is used for turbulence simulation to accurately describe the turbulence characteristics in the flow field. After completing the above steps, a full flow field numerical simulation is performed using computational fluid dynamics tools to extract key flow characteristic parameters, including velocity distribution, pressure distribution, and turbulent kinetic energy distribution. These parameters are further used to evaluate the rationality of the current aerodynamic parameter set and are used as feedback signals to input into the one-dimensional aerodynamic design model for parameter correction.

[0036] After the first round of correction, the aerodynamic parameter set is input into the airfoil parameter optimization module to generate the initial airfoil parameter set. The airfoil parameter set includes the meridional channel profile, blade mounting angle, blade curvature distribution, and thickness distribution. The workflow of the airfoil parameter optimization module includes the generation, screening, and iterative optimization of candidate airfoil parameter sets. Specifically, based on the aerodynamic parameter set output by the one-dimensional aerodynamic design model, the initial search space for airfoil parameters is defined, including the range of blade leading edge angle, trailing edge angle, and maximum thickness location, as well as the range of meridional channel profile, blade leading and trailing edge mounting angles, and airfoil thickness distribution. A genetic algorithm or particle swarm optimization algorithm is used to generate candidate airfoil parameter sets within the initial search space, and the candidate airfoil parameter sets are screened to remove parameter combinations that do not meet the geometric constraints. The screened airfoil parameter set is input into a three-dimensional numerical model for numerical simulation to evaluate the changing trends of its flow characteristic parameters. By comparing the changing trends of the old and new flow characteristic parameters, the adaptability of the current airfoil parameter set is determined, and iterative optimization is performed using the optimization algorithm.

[0037] The specific process of iterative optimization involves defining an optimization objective function, including weighted overall efficiency, flow stability index, and mechanical strength margin. Optimization termination conditions are set, including the objective function's rate of change being less than a preset threshold or reaching the maximum number of iterations. In each iteration, the objective function value is calculated based on the flow characteristic parameters of the current airfoil parameter set, and the search direction of the airfoil parameter set is updated based on the objective function value, generating a new round of candidate airfoil parameter sets. In this embodiment, an adaptive weighted particle swarm optimization algorithm is used to dynamically adjust the inertia weight and acceleration factor to improve convergence speed. Simultaneously, a local search strategy is introduced to perform a fine search near the global optimum to avoid getting trapped in local optima. Parallel computing technology is used to accelerate the optimization process and reduce overall computation time.

[0038] The verification and evaluation process of the optimization results is a key step in ensuring the reliability and practicality of the optimization method. The final optimized aerodynamic parameter set and airfoil parameter set are input into an independent three-dimensional numerical model for verification calculation. The verification calculation results are compared with the flow characteristic parameters in the optimization process to evaluate the consistency of the optimization results. The optimization results are further verified through experimental tests, including efficiency tests, pressure distribution tests, and vibration characteristic tests. The experimental test results show that the optimized centripetal turbine has achieved the expected goals in terms of efficiency, flow stability, and mechanical strength. Based on the experimental test results, the optimization method is improved to form a closed-loop optimization process.

[0039] Throughout the implementation process, by deeply integrating one-dimensional aerodynamic design with three-dimensional flow characteristic analysis, efficient collaborative optimization of thermodynamic, structural, and kinetic parameters was achieved. A multi-scale modeling mechanism and adaptive boundary condition adjustment strategy were introduced during the optimization process, effectively improving the adaptability and robustness of the optimization results. Kriging response surface fitting was established based on Latin hypercube sampling. Efficient space-filling sampling and the construction of high-precision surrogate models significantly improved the computational efficiency and design accuracy of complex numerical simulation optimizations such as those for turbines. While reducing simulation costs, it supports multi-objective, high-dimensional parameter, and uncertainty quantification analysis. The selection and implementation of optimization algorithms ensured the efficiency and accuracy of the optimization process, providing strong technical support for the design of centripetal turbines. An adaptive weighted particle swarm optimization algorithm dynamically adjusted inertia weights and acceleration factors to improve convergence speed. During each iteration, the current optimal solution and its corresponding flow characteristic parameters were recorded. A local search strategy was introduced to perform a fine search near the global optimum to avoid getting trapped in local optima. Parallel computing technology accelerated the optimization process and reduced overall computation time.

[0040] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0041] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A 1D-3D centripetal turbine collaborative optimization method, characterized in that: Includes the following steps: Step 1: Based on the heat source conditions and operating conditions, determine the optimization objective function and its constraints. The optimization objectives include efficiency, wheel power, and flow stability. Step 2: Generate an initial parameter set covering thermodynamic, aerodynamic, and structural parameters using a one-dimensional aerodynamic design model and a multi-objective optimization design function. The thermodynamic parameters are the total pressure and static pressure at the inlet and outlet, and the temperature. The aerodynamic parameters are the reaction degree, specific speed, speed ratio, and airflow angle. The structural parameters are the meridional channel profile and structural dimensions, blade installation angle, and number of blades. Step 3: Based on the initial aerodynamic parameter set, construct a three-dimensional numerical model, use computational fluid dynamics tools to perform full three-dimensional viscous flow numerical simulation, analyze velocity distribution, secondary flow and boundary layer separation law, evaluate the static pressure load distribution on the blade surface and the location of shock wave occurrence, and quantify the dissipation characteristics of tip leakage vortex and channel vortex loss source. Step 4: Iteratively correct the one-dimensional structural parameters and dynamic parameter set based on the three-dimensional flow characteristic parameters to form the aerodynamic parameter set after 1D-3D collaborative design optimization; Step 5: Input the optimized structural parameter set from the first round into the 3D airfoil parameter optimization module to generate an initial airfoil parameter set, which includes the meridional flow channel profile, blade mounting angle, blade curvature distribution, and thickness distribution. Step 6: Reconstruct the three-dimensional numerical model based on the airfoil parameter set obtained from the 1D-3D co-optimization design, and perform full flow field numerical simulation again to extract new flow characteristic parameters; Step 7: Based on the Latin hypercube experimental design, the airfoil parameter space is sampled, high-precision CFD numerical simulation is carried out and a sample database is constructed. The Kriging response surface is used to establish a 3D surrogate model for airfoil aerodynamic design, and the airfoil parameter set is iteratively optimized by combining optimization algorithms. Step 8: Finally, output the aerodynamic parameter set and airfoil parameter set after multiple rounds of iteration and convergence, as the optimization design result.

2. The 1D-3D centripetal turbine collaborative optimization method according to claim 1, characterized in that: In the one-dimensional aerodynamic design model, the inlet and outlet boundary conditions of the centripetal turbine are defined, including total temperature, total pressure and mass flow rate. Based on the center streamline and loss model, and fused with the Refprop real property database, the structural and dynamic parameters of the moving blades, guide vanes and volute are calculated.

3. The 1D-3D centripetal turbine collaborative optimization method according to claim 1, characterized in that: When constructing the three-dimensional numerical model, a three-dimensional geometric model is generated based on the structural parameter set output by the one-dimensional aerodynamic design model, including the blade passage, hub and casing. The three-dimensional geometric model is meshed, with the blade surface area using a structured mesh and the remaining area using an unstructured mesh. Boundary conditions of the three-dimensional numerical model are set, including inlet total pressure, outlet static pressure and wall no-slip condition. A suitable turbulence model is selected, and the k-ε two-equation model or the SST k-ω model is used for turbulence simulation. In the mesh generation process of the three-dimensional numerical model, the cell size of the structured mesh and the cell size of the unstructured mesh are adaptively adjusted according to the changes in the flow field gradient to ensure the computational accuracy and efficiency of complex flow regions.

4. The 1D-3D centripetal turbine collaborative optimization method according to claim 1, characterized in that: In the post-processing stage of the calculation results of the three-dimensional numerical model, key flow characteristic parameters are extracted, including blade surface pressure distribution, secondary flow and flow separation location. Evaluation indicators of flow characteristic parameters are defined, including pressure distribution, airfoil loss and flow uniformity index. The rationality of the current aerodynamic parameter set is judged based on the changing trend of the evaluation indicators, and it is used as a feedback signal to be input into the one-dimensional aerodynamic design model for parameter correction.

5. The 1D-3D centripetal turbine collaborative optimization method according to claim 1, characterized in that: In the airfoil parameter optimization module, the initial search space for airfoil parameters is defined based on the aerodynamic parameter set output by the one-dimensional aerodynamic design model. This includes the meridional flow channel profile, the range of the leading and trailing edge mounting angles of the blade, and the airfoil thickness distribution. A genetic algorithm or particle swarm optimization algorithm is used to generate a candidate airfoil parameter set within the initial search space. The candidate airfoil parameter set is then screened, and parameter combinations that do not meet the geometric constraints are eliminated. The screened airfoil parameter set is then input into the three-dimensional numerical model for numerical simulation to evaluate its flow characteristic parameters. In the blade profile parameter optimization module, during the generation of the candidate blade profile parameter set, the Latin hypercube sampling method is used to uniformly sample within the initial search space and establish a three-dimensional optimization surrogate model to improve the coverage of the parameter space and the optimization efficiency.

6. The 1D-3D centripetal turbine collaborative optimization method according to claim 1, characterized in that: During the iterative optimization process, an optimization objective function is defined, including isentropic efficiency, flow stability index, and circumferential power. Optimization termination conditions are set, including the rate of change of the objective function being less than a preset threshold or reaching the maximum number of iterations. In each iteration, the objective function is calculated based on the flow characteristic parameters of the current airfoil parameter set. The search direction of the airfoil parameter set is updated based on the objective function, and a new round of candidate airfoil parameter sets is generated.

7. The 1D-3D centripetal turbine collaborative optimization method according to claim 6, characterized in that: In the selection and implementation of the optimization algorithm, the adaptive weighted particle swarm optimization algorithm is adopted. The inertia weight and acceleration factor are dynamically adjusted to improve the convergence speed. In each iteration, the current optimal solution and its corresponding flow characteristic parameters are recorded. A local search strategy is introduced to conduct a fine search in the vicinity of the global optimal solution to avoid getting trapped in local optima. Parallel computing technology is used to accelerate the optimization process and reduce the overall computation time.

8. The 1D-3D centripetal turbine collaborative optimization method according to claim 1, characterized in that: In the process of verifying and evaluating the optimization results, the final optimized aerodynamic parameter set is input into an independent three-dimensional numerical model for verification calculation. The verification calculation results are compared with the flow characteristic parameters in the optimization process to evaluate the consistency of the calculation results. The optimization results are further verified through experimental tests, including efficiency tests, pressure distribution tests, and vibration characteristic tests. The optimization method is improved based on the experimental test results to form a closed-loop optimization process.

Images