A high-pressure cylinder flow field analysis method based on computational fluid dynamics

Abstract

The application discloses a high-pressure cylinder flow field analysis method based on computational fluid dynamics, and particularly relates to the high-pressure cylinder flow field analysis field, and comprises the following steps: constructing a flow field model through a cross-factory design knowledge base, identifying a sensitive area according to a flow loss criterion and gridizing, solving a fluid equation set in parallel in a partitioned manner, dynamically terminating iteration based on entropy production rate variation, and finally encapsulating analysis results. The high-pressure cylinder flow field analysis method based on computational fluid dynamics realizes the intellectualization and light weight of grid generation by automatically identifying a secondary flow loss sensitive area and combining gridization processing, reduces manual intervention time, and improves grid generation efficiency; and the utilization rate of calculation resources is improved by dividing a calculation domain into a main flow area, a boundary layer area and a vortex core area and matching corresponding fluid control equation sets.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of high-pressure cylinder flow field analysis, and more particularly to a high-pressure cylinder flow field analysis method based on computational fluid dynamics. BACKGROUND

[0002] With the increasing demand for efficiency improvement of thermal power units, the fine simulation of the through-flow design of the high-pressure cylinder of the steam turbine is particularly important. The traditional design relies on the experience formula system of a single manufacturer to realize the stepwise distribution of enthalpy drop through the simplified flow field equation. Although the traditional technology can complete the basic design, it has a fundamental defect: due to the simplified mechanism of the two-dimensional model and the manual grid discretization method, it is difficult to accurately capture the three-dimensional secondary flow vortex structure and boundary layer separation effect.

[0003] To overcome the defects of the traditional design, the existing technology introduces a three-dimensional full-channel simulation method of computational fluid dynamics, uses a parameterized modeling tool to construct a million-level grid model, optimizes the blade profile through a mixed loading flow pattern algorithm and controllable vortex design, and uses a multi-condition transient solver to realize dynamic distribution of enthalpy drop, so as to improve the prediction accuracy of the high-pressure cylinder efficiency.

[0004] However, in actual use, it still has some disadvantages, such as low grid generation efficiency, long-time manual intervention for single-model million-level grid construction, dramatic increase in computing resource consumption, long time-consuming for single-condition simulation caused by full three-dimensional transient solving, and high optimization iteration cost, which requires a week of calculation period for multiple parameter iterations, and finally leads to delayed engineering response and uncontrolled hardware cost. SUMMARY

[0005] To overcome the above-mentioned defects of the prior art, the present application provides a high-pressure cylinder flow field analysis method based on computational fluid dynamics, which solves the problems in the background art by the following scheme.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0007] A high-pressure cylinder flow field analysis method based on computational fluid dynamics, comprising:

[0008] S1: generating an inter-stage constraint matrix and a blade cascade geometric topology based on the enthalpy drop distribution rules and blade profile parameters in the pre-set cross-manufacturer design knowledge base, to construct a first flow field analysis model of the high-pressure cylinder;

[0009] S2: identifying a secondary flow loss sensitive area in the first flow field analysis model in response to a pre-set flow loss design criterion, and generating a second flow field analysis model through grid processing;

[0010] S3: Based on the second flow field analysis model, perform region division, and match the corresponding fluid control equation set for partitioned parallel solving to obtain the first flow field data set of the high-pressure cylinder;

[0011] S4: Real-time monitoring of entropy production rate change in the rotor-stator blade row interface area is performed using the first flow field data set, and when the entropy production rate change is lower than the convergence threshold, the iterative calculation is terminated, and the second flow field data set is identified based on the reverse;

[0012] S5: The analysis process including S1 to S4 and the second flow field data set output by S4 are packaged into a lightweight container and deployed to a distributed computing node.

[0013] Preferably, S1, constructing the first flow field analysis model of the high-pressure cylinder, specifically includes:

[0014] Quantifying low-level load as an inter-stage constraint matrix, and the low-level load specifically represents small enthalpy drop and multiple stages;

[0015] Obtaining the blade row geometry topology of the blade row channel based on the characteristics of low hub ratio and high aspect ratio , wherein, is represented as the blade root diameter, is represented as the relative blade height, and the characteristics of low hub ratio and high aspect ratio specifically represent the characteristics of low root diameter and large relative blade height;

[0016] Outputting the full-flow passage geometry topology model structure of the high-pressure cylinder, i.e., the first flow field analysis model.

[0017] Preferably, S2, the identification of the secondary flow loss sensitive area, specifically includes:

[0018] Quantitative analysis of vortex core intensity to calculate the secondary flow intensity factor , specifically represented as:

[0019] ,

[0020] wherein, is represented as the tangential velocity component of the fluid micro-cluster, is represented as the axial velocity component of the fluid micro-cluster, is represented as the radial position of the fluid micro-cluster in the flow passage;

[0021] Preferably, S2, the identification of the secondary flow loss sensitive area, specifically further includes:

[0022] Quantitative analysis of the entropy increase risk area by calculating the entropy increase risk index , specifically represented as:

[0023] ,

[0024] wherein, is expressed as the blade root diameter, is expressed as the relative blade height, is expressed as the proportionality coefficient, is expressed as the flow passage curvature function.

[0025] Preferably, the specific steps of the S2, the meshing processing, include:

[0026] Based on the secondary flow loss sensitive area identified by the secondary flow intensity factor, the following is performed:

[0027] In the secondary flow intensity factor > 5% vortex core area is encrypted to 0.1 mm grid resolution, and the boundary layer boundary layer grid meeting the wall distance < 1 is generated near the wall surface;

[0028] In the entropy increase risk index < 0.1% non-critical area is sparse to 2 mm grid.

[0029] Preferably, the S3, obtaining the first flow field data set, specifically includes:

[0030] The calculation domain of the second flow field analysis model is divided into a main flow area, a boundary layer area and a vortex core area;

[0031] The vortex core area is determined by the secondary flow intensity factor > 5%, the boundary layer area is determined according to the wall distance of the near-wall grid, and the remaining part is divided into a main flow area;

[0032] The Euler equation is solved in the main flow area, the Prandtl boundary layer equation is solved in the boundary layer area, and the vorticity transport equation is solved in the vortex core area;

[0033] Real-time data exchange of the three areas is realized by constructing a data interface.

[0034] Preferably, the S4, the convergence threshold is defined as the entropy generation rate change rate < 0.1%.

[0035] Preferably, the S4, the reverse identification is specifically: according to the steam extraction regulation accuracy ± 2%, the high-pressure cylinder flow field is reversely deduced to the key design parameter set for directional optimization.

[0036] Technical effects and advantages of the present application:

[0037] 1、The present application automatically identifies the secondary flow loss sensitive area through S2, realizes the intelligentization and light weight of grid generation by combining with the meshing processing, reduces the manual intervention time, and improves the grid generation efficiency;

[0038] 2、The application divides the calculation domain into mainstream area, boundary layer area and vortex core area through S3, matches corresponding fluid control equation set, improves calculation resource utilization rate, shortens full three-dimensional transient solution time consumption, and relieves problem of rapid increase of calculation resource consumption;

[0039] 3、The application dynamically terminates iteration through S4 real-time monitoring of entropy production rate change of rotor-stator interface area, greatly reduces iteration number and optimization parameter dimension, and reduces calculation period and cost of optimization iteration. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 A step block diagram of a high-pressure cylinder flow field analysis method based on computational fluid dynamics is provided according to an embodiment of the application. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the application.

[0042] The terms used in the following embodiments of the application are only for the purpose of describing the specific embodiments, and are not intended to be limiting on the application. As used in the specification, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "and / or" as used herein refer to and encompass any or all possible combinations of one or more of the associated listed items.

[0043] Hereinafter, the terms "first" and "second" are only for the purpose of description, and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more of the features, and in the description of the embodiments of the application, the meaning of "a plurality of" is two or more, unless otherwise specified.

[0044] As shown in the accompanying drawings Figure 1 A high-pressure cylinder flow field analysis method based on computational fluid dynamics constructs a flow field model through a cross-manufacturer design knowledge base, identifies sensitive areas according to flow loss criteria and gridding, solves fluid equation sets in parallel in zones, dynamically terminates iteration based on entropy production rate change, and finally encapsulates the analysis results; specifically including the following steps:

[0045] S1: generating an inter-stage constraint matrix and a cascade geometry topology based on an enthalpy drop distribution rule in a preset cross-manufacturer design knowledge base and a blade profile parameter to construct a first flow field analysis model of the high-pressure cylinder;

[0046] S2: identifying a secondary flow loss sensitive area in the first flow field analysis model in response to a preset flow loss design criterion, and generating a second flow field analysis model through a meshing process;

[0047] S3: performing region division based on the second flow field analysis model, matching a corresponding fluid control equation set for partitioned parallel solving to obtain a first flow field data set of the high-pressure cylinder;

[0048] S4: monitoring an entropy production rate change of a rotor-stator cascade junction area in real time by using the first flow field data set, terminating iteration calculation when the entropy production rate change is lower than a convergence threshold, and identifying a second flow field data set based on a reverse identification;

[0049] S5: encapsulating an analysis process including S1 to S4 and the second flow field data set output by S4 into a lightweight container and deploying to a distributed computing node.

[0050] It should be noted that the embodiment takes a 330 MW imported Alstom steam turbine as an example, and aims to comprehensively improve the overall performance of the steam turbine without changing the size of the steam turbine body, by applying the whole-process algorithm of S1 to S5, comprehensively considering the performance optimization of each component of the steam turbine, through the modification of the high-pressure cylinder through-flow and the middle-pressure cylinder rotating partition plate, and the improvement of the condenser related technology, to comprehensively improve the overall performance of the steam turbine, so as to realize the dual improvement of the unit through-flow efficiency and the industrial steam supply capacity, and finally verify the advancement and effectiveness of the algorithm by using actual operation data.

[0051] Specifically, in S1, the structured processing of the input data is realized by pre-constructing a cross-manufacturer design knowledge base.

[0052] It should be noted that the cross-manufacturer design knowledge base is not a simple data set, but is constructed by deeply integrating and fusing the proprietary design rules and parameter sets of multiple manufacturers.

[0053] In the embodiment, the cross-manufacturer design knowledge base integrates an enthalpy drop distribution rule library of A factory and a blade profile parameter data set of B factory, the enthalpy drop distribution rule library defines the recommended value range, constraint conditions and adjustment strategy of the enthalpy drop ratio between each stage in the high-pressure cylinder under different working conditions, and the blade profile parameter data set contains more than 20 dimensions of key parameters, including but not limited to blade installation angle, chord length, inlet and outlet flow angle, etc.

[0054] In a possible implementation, generating the inter-stage constraint matrix and the cascade geometry topology includes: quantifying a low-stage load into an inter-stage constraint matrix [M]n×n ; obtain the cascade channel of the cascade geometry topology based on the low hub ratio and high aspect ratio characteristics , wherein, is expressed as the blade root diameter, is expressed as the relative blade height; the full passage geometry topology model structure of the high-pressure cylinder, that is, the first flow field analysis model, is output.

[0055] In the embodiment, the first flow field analysis model adopts a B-rep+NURBS hybrid representation in the STEP P242 format, and only stores the control point coordinates and constraint matrix [M] n×n , thereby reducing the file size; and the design variable dimension is compressed from the traditional full-parameter optimization to only adjusting the non-zero elements of the matrix through the inter-stage constraint matrix [M] n×n , thereby reducing the upper limit of the iteration number of S4.

[0056] Further, the low load specifically represents the classic layout of "small enthalpy drop and multiple stages", the purpose of which is to make each through-flow stage operate near its highest efficiency point and make it a calculable constraint condition, which is specifically implemented by quantifying the inter-stage constraint matrix [M] n×n , and is specifically expressed as:

[0057] ,

[0058] , wherein, and are respectively expressed as the enthalpy drop values of the first stage and the second stage, and are expressed as the constraint boundaries of the enthalpy drop values, and the values are dynamically loaded according to the specific design working condition by the enthalpy drop distribution rule library in the cross-manufacturer design knowledge base.

[0059] Further, the low hub ratio and high aspect ratio characteristics are implemented by realizing the geometric topology of "low root diameter and large relative blade height", thereby significantly reducing the axial through-flow velocity, effectively suppressing the secondary flow loss, and finally realizing the improvement of the through-flow efficiency. In the embodiment, the NURBS control points of the parameterized template are constructed according to the aforementioned cascade geometry topology , and the construction mode of the control points is specifically expressed as:

[0060] ,

[0061] , wherein, , , , , is expressed as the blade profile chord length, is expressed as the blade root diameter, denoted as relative blade height; it is to be noted that the denoted as leading edge point, embodied as low root diameter constraint; the denoted as blade midspan control point; the denoted as trailing edge point, embodied as low root diameter constraint. denoted as trailing edge point, embodied as low root diameter constraint.

[0062] Specifically, in S2, the full passage geometry topology model structure output by S1 is received, and a preset design criterion, i.e., a flow loss design criterion, containing expert knowledge is integrated to establish an intelligent identification mechanism for a secondary flow loss sensitive region, thereby constructing a light-weight grid model, i.e., a second flow field analysis model.

[0063] In the embodiment, the flow loss design criterion is derived from a mixed loading blade shaping design criterion and controllable vortex design parameters in an external configuration file.

[0064] In a possible implementation, the intelligent identification of the secondary flow loss sensitive region includes: quantitatively analyzing vortex core strength, calculating a secondary flow intensity factor to evaluate the secondary flow loss risk of each point in the flow passage, specifically denoted as:

[0065] ,

[0066] wherein, denotes a tangential velocity component of the fluid micro-cluster, denotes an axial velocity component of the fluid micro-cluster, denotes a radial position of the fluid micro-cluster in the flow passage; in the embodiment, when the calculated value exceeds a preset threshold of 5%, the region is automatically marked as a secondary flow loss sensitive region.

[0067] In a possible implementation, the intelligent identification of the secondary flow loss sensitive region further includes: constructing an entropy increase prediction model based on the mixed loading blade shaping design criterion, and quantitatively analyzing and locating an entropy increase risk region by calculating an entropy increase risk index , specifically denoted as:

[0068] ,

[0069] wherein, denotes a root diameter, denotes a relative blade height, denotes a proportional coefficient, determined according to the mixed loading blade shaping design criterion, denotes a flow passage curvature function, used to evaluate the entropy increase risk caused by the bending of the flow passage.

[0070] In a possible implementation, the meshing process includes: based on the secondary flow loss sensitive area identified by the secondary flow intensity factor, performing; in the secondary flow intensity factor >5% of the vortex core area is encrypted to a 0.1 mm grid resolution, and a boundary layer boundary layer grid that meets a wall distance <1 is generated near the wall surface to ensure the solution accuracy of the boundary layer flow; in the entropy increase risk index <0.1% of the non-critical area is sparse to 2 mm grid.

[0071] In this embodiment, in order to realize global light weight, an Octree spatial segmentation topology optimization algorithm is also used to maximize the reduction of the total number of grid nodes under the premise of ensuring solution accuracy.

[0072] Specifically, in S3, the light-weight mesh model generated by S2 is received, and the preset optimization target is integrated to divide the calculation domain and match the corresponding fluid control equation set for partitioned parallel solving, and real-time data exchange is performed between regions through the preset flux matching interface, so as to obtain the transient flow field data set of the high-pressure cylinder, that is, the first flow field data set.

[0073] In this embodiment, the optimization target is derived from the dynamic-static matching optimization parameters loaded in the configuration file and the enthalpy drop distribution data inherited from the original design scheme.

[0074] In a possible implementation, the region division and matching of the fluid control equation set include: dividing the calculation domain of the second flow field analysis model into a main flow region, a boundary layer region and a vortex core region, wherein the vortex core region is determined by the secondary flow intensity factor >5%, the boundary layer region is determined according to the wall distance of the near-wall grid

[0075] It should be noted that the region division is automatically performed based on the physical markers carried by each grid in the second flow field analysis model generated by S2; in this embodiment, the grid with a wall distance <5 is identified as the boundary layer region; the Euler equation directly loads the enthalpy drop distribution data of the original design scheme as the initial condition for solving; in the boundary layer region with a steep flow gradient but an extremely thin scale, the Prandtl boundary layer equation is used for accurate calculation, and the key parameters of the turbulence model are inherited from S1. The parameters are associated; in the most complex vortex core structure of the flow structure, the vorticity transport equation capturing the vortex motion is used, and its boundary condition is coupled with the design criterion of the mixed loading blade shaping inherited from the S2 input.

[0076] Further, in order to realize regional load balancing optimization, the computing resources are dynamically allocated according to the computing complexity of different regions, that is, according to the grid proportion and estimated computing weight of each region, the number of GPU cores is intelligently allocated; in the embodiment, the vortex core region with a grid proportion of 12% and a computing weight of 58% is allocated to the 6 GPU cores with the strongest performance, and the mainstream region with a grid proportion of 65% and a computing weight of 10% with the smallest computing amount is only allocated to 1 GPU core.

[0077] In the embodiment, in the mainstream region, due to the relatively weak viscous effect, the Euler equation simplified model is used, and the control equation is specifically represented as:

[0078] ,

[0079] wherein, is represented as the current time , at the center of the grid unit, the mass of the fluid in the unit volume, is represented as the transient solving time step, is represented as the absolute velocity vector at the center of the grid unit; is represented as the unit tensor, is represented as the static pressure.

[0080] It should be noted that, , , The initial value is directly inherited from S1, without additional calibration; in the mainstream region, the viscous dissipation is ignored, The change is only determined by the mass conservation; the transient solving time step is synchronized with the monitoring frequency of the entropy production rate in S4; the absolute velocity vector is determined by and , as the initial condition of the Euler equation; the static pressure is used to calculate the total enthalpy of the mainstream region.

[0081] In the embodiment, in the boundary layer region, the viscous effect and shear force are the dominant factors, then the Prandtl boundary layer equation is used for accurate calculation, and is specifically represented as:

[0082] ,

[0083] wherein, is represented as the time-averaged velocity component along the tangential direction of the blade wall surface, The time-averaged velocity component along the wall normal, The streamwise coordinate along the wall, The wall normal coordinate, The static pressure at the current time step, The mass of fluid per unit volume at the center of the grid cell, The static pressure, The kinematic viscosity.

[0084] It should be noted that in the boundary layer of the high-pressure cylinder, Directly related to the large relative blade height The blade surface load distribution is determined; under the assumption of thin boundary layer, ≪ , and The order of magnitude of y⁺<5 is generated by S2 near-wall grid analysis; the origin of the streamwise coordinate along the wall corresponds to the blade leading edge derived from S1, and the end is the trailing edge, which is used to integrate the entire blade friction loss; the wall normal coordinate =0 is the blade surface, = is the outer edge of the boundary layer, is determined by and the Reynolds number; if >0, boundary layer separation will occur, which directly affects the entropy production monitoring of S4; the kinematic viscosity is automatically adjusted according to the "large relative blade height" to adjust the turbulence model parameters, so that the viscosity changes with the blade geometry, thereby improving the accuracy of the boundary layer solution.

[0085] In this embodiment, in the vortex core region, the vortex transport equation is used for solving, and the boundary conditions of the design criteria of mixing and loading blade shaping in S2 are coupled, which is specifically expressed as:

[0086] ,

[0087] Wherein, The rotation intensity and direction of the micro-cluster in the vortex core region, The transient solution time step, The absolute velocity vector at the center of the grid cell, The kinematic viscosity.

[0088] It should be noted that, The diffusion rate of vorticity in the vortex core region due to viscosity to the surrounding; The stretching and bending effect of vortex lines in the velocity gradient field; The instantaneous rate of change of vorticity vector with time when the fluid micro-cluster moves together in the vortex core region.

[0089] Specifically, in S4, the first flow field data set output by S3 is loaded, and at the same time, a preset stability requirement for the steam supply pressure fluctuation is read, and a converged entropy production rate change threshold is calculated, so as to reversely identify the key design parameter set that has the greatest impact on the target based on the flow field data in the converged state, that is, the second flow field data set.

[0090] In a possible implementation, the convergence threshold is the entropy production rate change rate <0.1%, and the reverse identification specifically comprises: according to the steam extraction regulation accuracy ± 2%, the key design parameter set of the high-pressure cylinder flow field is reversely deduced and optimized.

[0091] It should be noted that the entropy production rate field is extracted from the first flow field data set , and a derivative thereof in the time dimension is calculated , the derivative is a key index for measuring the stability and convergence of the flow field; further, the entropy production monitoring function is continuously called in an iteration cycle, and once the monitored average value is lower than the preset convergence threshold 0.1%, the convergence flag is triggered, indicating that the flow field has tended to be stable and does not need to be further iterated.

[0092] In the embodiment, when the steam supply pressure fluctuation is greater than 1.5%, the convergence threshold is automatically relaxed to 0.2% to avoid invalid fine iteration under unstable working conditions; otherwise, a more stringent threshold of 0.1% is used.

[0093] Further, after the convergence flag is triggered, that is, after the iteration is terminated, the sensitive parameter reverse identification is started, based on the finally converged flow field data, combined with the performance index of the steam extraction regulation accuracy ± 2%, the sensitivity of each design parameter to the final performance is analyzed by calculating the Jacobian matrix, and through the preset sensitivity threshold, the sensitivity threshold of the embodiment is 0.15, the 3 to 5 key parameters with the greatest impact on the performance are automatically selected as the key design parameter set, that is, the second flow field data set.

[0094] In the embodiment, the steam extraction pressure PG is an observation quantity, and the key design parameters , the sensitivity matrix is , if > 0.15, indicates the index of the key design parameter, and the key design parameter ; otherwise, it is removed.

[0095] The second flow field data set comprises a velocity field, a vorticity distribution of all time steps, and boundary layer data.

[0096] Specifically, in S5, the lightweight container is a Docker image, and the local re-optimization of the high-pressure cylinder is triggered by the condenser water supply variable condition signal after deployment.

[0097] In the embodiment, the modification of the high-pressure cylinder of the unit is directly derived from the output of the algorithm of the application; the aerodynamic optimization of the blade is based on the second flow field data set output by S4, that is, only the sensitive parameter set containing 3-5 dimensions is used; the physical modification strictly implements the small enthalpy drop and multi-stage layout scheme defined by the constraint matrix of S1; the specific profile design of the blade follows the flow loss criterion of S2 and adopts an advanced mixed loading flow pattern; the geometric shape is accurately realized by reducing the blade root diameter and increasing the relative blade height, thereby realizing the optimization of the low root diameter and large relative blade height defined by S1; the sensitive parameter dimension reduction of S4 and the topology and flow pattern optimization of S1 and S2 accurately guide the physical modification and produce significant performance improvement.

[0098] Further, the modification of the rotating partition of the medium-pressure cylinder is driven by the algorithm data of the application; the implementation of the sealing strengthening process is based on the axial thrust balance parameters in the second flow field data set output by S4 and the material durability constraint, and nitriding treatment is performed on the sealing surface to improve the surface hardness; to ensure the stability of the unit under variable conditions, the cooling area reserved is calculated according to the requirements of the lightweight container of S5 for the response of the working condition;

[0099] Further, the pipeline configuration scheme for the large-capacity water supply of the condenser directly implements the dynamic deployment scheme output by the lightweight container in S5, and under the base load condition, a 100t / h pipeline is enabled, while under the variable condition of a sharp increase in water supply demand, three 150t / h pipelines are automatically switched to run in parallel, ensuring efficient operation of the water supply under any condition, and finally stabilizing the oxygen content of the condensed water, successfully achieving the engineering goal set in the optimization iteration of S4, and verifying the effectiveness of the edge deployment and real-time response of S5.

[0100] Secondly: the drawings in the disclosed embodiment of the application only involve the structures involved in the disclosed embodiment of the application, other structures can be referred to the general design, and under the condition of no conflict, the same embodiment and different embodiments of the application can be combined with each other;

[0101] Finally: the above only describes the preferred embodiments of the application and is not used to limit the application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. A high-pressure cylinder flow field analysis method based on computational fluid dynamics, characterized by, The method comprises the following steps: S1: Based on the enthalpy drop allocation rule in the preset cross-factory design knowledge base and the blade type parameter, an inter-stage constraint matrix and a cascade geometry topology are generated to construct a first flow field analysis model of the high-pressure cylinder; S2: In response to the preset flow loss design criterion, a secondary flow loss sensitive area in the first flow field analysis model is identified, and a second flow field analysis model is generated through meshing processing; S3: Based on the second flow field analysis model, regional division is performed, and corresponding fluid control equation sets are matched for partitioned parallel solving to obtain a first flow field data set of the high-pressure cylinder, comprising: The calculation domain of the second flow field analysis model is divided into a main flow area, a boundary layer area and a vortex core area; Vortex core region, determined by the secondary flow intensity factor > 5% decision, boundary layer region, determined by the wall distance of the near-wall grid Values determine, the rest is divided into the main flow region; Euler equations are solved in the main flow area, Prandtl boundary layer equations are solved in the boundary layer area, and vorticity transport equations are solved in the vortex core area; A data interface is constructed to transfer data exchange of the three areas in real time; S4: Using the first flow field data set, the entropy production rate change of the rotor-stator cascade junction area is monitored in real time, and when the entropy production rate change is lower than the convergence threshold, the iteration calculation is terminated, and the second flow field data set is identified based on the reverse; S5: The analysis process comprising S1 to S4 and the second flow field data set output by S4 are packaged into a lightweight container and deployed to a distributed computing node.

2. The method of claim 1, wherein: The S1, constructing the first flow field analysis model of the high-pressure cylinder, specifically comprises: Quantifying the low-level load into an inter-stage constraint matrix, and the low-level load specifically represents small enthalpy drop and multiple stages; Acquiring a cascade geometry topology of a cascade passage based on a low hub ratio and a high aspect ratio , wherein is expressed as a blade root diameter, is expressed as a relative blade height, and the low hub ratio and the high aspect ratio are specifically embodied as a low root diameter and a large relative blade height. Outputting a full-flow passage geometry topology model structure, i.e. the first flow field analysis model of the high-pressure cylinder.

3. The method of claim 1, wherein: The S2, the identification of the secondary flow loss sensitive area, specifically comprises: The vortex core strength is quantitatively analyzed, and the secondary flow intensity factor is calculated , specifically expressed as: ， wherein, Vt represents the velocity component of the fluid element in the tangential direction, Va represents the velocity component of the fluid element in the axial direction, r represents the radial position of the fluid element in the flow passage.

4. The method of claim 1, wherein: The S2, the identification of the secondary flow loss sensitive area, specifically further comprises: By calculating the entropy increase risk index Quantitative analysis positioning entropy increase risk area, specifically expressed as: ， wherein, is expressed as a blade root diameter, is expressed as a relative blade height, is expressed as a proportionality coefficient, is expressed as a flow passage curvature function.

5. The computational fluid dynamics-based flow field analysis method for a high-pressure cylinder according to claim 4, characterized by: The specific steps of the meshing processing of the S2 comprise: Based on the secondary flow loss sensitive area identified by the secondary flow intensity factor, execution is performed; In the secondary flow intensity factor The vortex core region of 5% is encrypted to 0.1mm grid resolution, while the boundary layer <1 boundary layer boundary layer grid; In the entropy increase risk indicator Non-critical zone sparseness to 2mm grid of <0.1%.

6. The computational fluid dynamics-based flow field analysis method for a high-pressure cylinder according to Claim 1, characterized by: The S4, convergence threshold is defined as the rate of change of entropy production <0.1%.

7. The computational fluid dynamics-based flow field analysis method for a high-pressure cylinder according to Claim 1, characterized by: The S4, the reverse identification specifically comprises: according to the steam extraction regulation accuracy ± 2%, the key design parameter set of the high-pressure cylinder flow field is reversely pushed to perform directional optimization.

Images