CFD-based method and apparatus for topology optimization design of slurry valve anti-erosion flow channels

Abstract

The present application relates to the technical field of slurry valve design, and discloses a slurry valve anti-erosion flow channel topology optimization design method and device based on CFD, which establishes a particle size grouping identification system by acquiring slurry particle size distribution data, and constructs a three-dimensional particle enrichment index distribution atlas; a flow channel geometric feature parameter sequence is extracted from a slurry valve three-dimensional CAD model to generate a full-flow channel particle concentration distribution state sequence; a layered erosion energy density distribution is generated based on the sequence, three erosion modes are identified and comprehensive erosion intensity is calculated, and a three-dimensional erosion intensity distribution field is constructed; then, erosion hotspots are identified, differentiated geometric adjustment strategies are formulated, and finally an anti-erosion optimized flow channel geometric model is generated; the present application solves the problem of local erosion prediction distortion caused by the assumption of uniform particle distribution in traditional CFD simulation, improves the service life and reliability of the slurry valve under the working condition of multi-particle size particle slurry, and reduces the risk of sudden shutdown and maintenance cost caused by local erosion.

Description

Technical Field

[0001] This invention relates to the field of slurry valve design technology, and more specifically, to a method and apparatus for CFD-based anti-scouring flow channel topology optimization design of slurry valves. Background Technology

[0002] In industries such as mining, chemical engineering, and environmental protection, slurry valves are widely used to control the flow of slurries containing particles of various sizes. However, the slurries in these industries typically contain particles with a wide range of sizes. When flowing through the slurry valve, due to gravity, large particles tend to settle at the bottom of the valve body or in low-lying flow channels, forming a high-concentration particle layer. Traditional CFD simulations are usually based on the assumption of "uniform particle distribution," which cannot accurately simulate the stratified flow characteristics of particles and the resulting local high concentrations and intensified scouring. This leads to severely distorted predictions of scouring intensity in areas such as the bottom and corners of the valve body.

[0003] Chinese Patent CN119558145B discloses a design method and apparatus for the sealing structure of a large-diameter slurry valve suitable for high-wear conditions. By acquiring key geometric parameters of the sealing structure and constructing a three-dimensional topological model and a mechanical performance simulation model, a digital twin model is formed. Simultaneously, a time-series dataset of wear morphology is constructed, and three-level early warning thresholds are set for key operating parameters, establishing rules for classifying operating risk levels. Through these measures, this prior art can achieve accurate prediction and optimized design of the wear state of the sealing structure of a large-diameter slurry valve, significantly improving the service life and reliability of the sealing structure under high-wear conditions.

[0004] However, existing technologies cannot accurately capture the gravitational stratification behavior of particles and the resulting localized high concentrations when processing slurries with multi-size particles. This leads to inaccurate predictions of scouring intensity in areas such as the bottom and corners of the slurry valve. Consequently, the slurry valve may experience bottom wear-through or leakage within a few months (e.g., 2-4 months) after commissioning, resulting in sudden production shutdowns, high repair costs, and severe disruption to production continuity. Summary of the Invention

[0005] To overcome the aforementioned shortcomings of existing technologies, this invention provides a CFD-based method and apparatus for optimizing the topology of slurry valve anti-scour flow channels. By accurately simulating the stratified flow characteristics of multi-size particle slurry within the slurry valve, it identifies and optimizes local scour hotspots. This invention can significantly extend the service life of slurry valves, reduce maintenance costs, and improve production continuity.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A CFD-based method for optimizing the topology of slurry valve anti-erosion channels includes:

[0008] Obtain slurry particle size distribution data, and establish a particle size grouping and identification system based on the slurry particle size distribution data; construct a three-dimensional particle enrichment index distribution map based on the slurry particle size distribution data and the particle size grouping and identification system.

[0009] Based on the 3D CAD model of the slurry valve, the sequence of geometric feature parameters of the flow channel is extracted; based on the sequence of geometric feature parameters of the flow channel and the 3D particle enrichment index distribution map, the sequence of particle concentration distribution state of the entire flow channel is generated.

[0010] Based on the state sequence of particle concentration distribution throughout the flow channel, a layered scouring energy density distribution is generated; according to the layered scouring energy density distribution, three scouring modes, namely impact, abrasion and corrosion, are identified, the comprehensive scouring intensity is calculated and a three-dimensional scouring intensity distribution field is generated.

[0011] Based on the three-dimensional scour intensity distribution field, layered scour hotspots are identified and differentiated geometric adjustment strategies are formulated. Topology optimization iterative feedback is performed to generate a scour-resistant optimized flow channel geometric model.

[0012] Furthermore, the method for establishing a particle size grouping and identification system includes: dividing particles into N particle size groups based on particle size distribution data using a logarithmic interval principle, and assigning a unique digital identification code to each particle size group.

[0013] Furthermore, the method for constructing the three-dimensional particle enrichment index distribution map includes:

[0014] Calculate the theoretical settling velocity for each particle size group, establish a flow velocity-stratification critical determination model, and generate quantitative indicators of the stratification degree for each particle size group.

[0015] Based on the stratification index, the vertical direction of the flow channel is discretized into M computational layers. The particle enrichment index of each particle size group in each computational layer is calculated, and a three-dimensional particle enrichment index distribution map is constructed.

[0016] Furthermore, the flow velocity-stratification critical determination model is as follows:

[0017] Obtain the local flow velocity v at the current position within the flow channel. flow Define the critical ratio R for the i-th particle size group. i =v flow / v settle,i Where i is the index of the particle size group, v settle,i This represents the theoretical settling velocity of the i-th particle size group;

[0018] When R i <R critical When R is significantly stratified, it is considered a significant stratification. i >R mix When R is fully mixed, it is considered fully mixed; when R is fully mixed, it is considered fully mixed. critical ≤Ri ≤R mix At that time, the quantification index D was used to determine the degree of stratification. i Describe the transition state, D i D represents the quantitative index of the stratification degree of the i-th particle size group. i The value range is from 0 to 1, where 0 corresponds to complete mixing and 1 corresponds to complete stratification. R critical R is the critical threshold for stratification. mix This is the mixed critical threshold.

[0019] Furthermore, the method for calculating the particle enrichment index of each particle size group in each computational layer is as follows:

[0020] Let k be the index variable of the computation layer. For the k-th computation layer L k and the i-th particle size group G i Obtain the computation layer L k Inner particle size group G i Local concentration C ik and the particle size group G throughout the flow channel i average concentration C i,avg Through C ik Divide by C i,avg Obtain the i-th particle size group G i In the k-th computational layer L k The particle enrichment index.

[0021] Furthermore, in the three-dimensional particle enrichment index distribution map, the horizontal axis represents the particle size group number, the vertical axis represents the computational layer number, and the value represents the particle enrichment index E. ik .

[0022] Furthermore, the method for extracting the sequence of geometric feature parameters of the flow channel based on the 3D CAD model of the slurry valve includes:

[0023] Establish a curved coordinate system along the centerline of the flow channel; from the inlet to the outlet of the slurry valve, set a geometric feature extraction section at preset distances, for a total of P geometric feature extraction sections;

[0024] Based on the 3D CAD model of the slurry valve, the geometric feature parameters of each geometric feature section are extracted. The geometric feature parameters include the centroid coordinates of the section, the area of ​​the section, and the equivalent diameter of the section.

[0025] Geometric feature parameters are extracted from each cross section based on each geometric feature, and geometric change parameters between adjacent cross sections are calculated. The geometric change parameters include radius of curvature, contraction ratio, and expansion angle. The geometric feature parameters and geometric change parameters constitute a geometric feature parameter sequence.

[0026] Furthermore, the geometric types include straight pipe sections, curved sections, contraction sections, and expansion sections;

[0027] The method for constructing the full-channel particle concentration distribution state sequence includes:

[0028] Based on the sequence of geometric feature parameters, the labeled geometric type, and the three-dimensional particle enrichment index distribution map, a correction model based on geometric type is constructed.

[0029] Based on the geometry-based correction model, a multi-segment flow channel particle distribution state transfer algorithm is established to generate a full flow channel particle concentration distribution state sequence containing the concentration distribution of each particle size group at each location.

[0030] Furthermore, the method for establishing the multi-segment flow channel particle distribution state transfer algorithm includes:

[0031] The entire flow channel is divided into Q geometric segments, and the particle distribution state vector of the upstream geometric segment is constructed based on the particle enrichment index of each particle size group in each calculation layer.

[0032] Based on the geometry-type-based modified model, a distribution state transfer matrix is ​​constructed between geometric segments to describe the evolution of particle distribution from the upstream geometric segment to the downstream geometric segment;

[0033] The particle distribution state vector of the downstream geometric segment is calculated by performing matrix operations between the distribution state transfer matrix and the particle distribution state vector of the upstream geometric segment.

[0034] The particle distribution state vector of the downstream geometric segment is used as the particle distribution state vector of the new upstream geometric segment. The particle distribution state vector and distribution state transfer matrix of the upstream geometric segment are applied sequentially along the flow channel direction to complete the full flow channel particle distribution state transfer calculation from the inlet to the outlet, including all geometric segments.

[0035] A CFD-based slurry valve anti-erosion channel topology optimization design device is used to implement the aforementioned CFD-based slurry valve anti-erosion channel topology optimization design method. The device includes:

[0036] Enrichment map construction module: used to acquire slurry particle size distribution data, establish a particle size grouping identification system based on the slurry particle size distribution data; and construct a three-dimensional particle enrichment index distribution map based on the slurry particle size distribution data and the particle size grouping identification system.

[0037] The full-channel concentration state calculation module extracts the geometric feature parameter sequence of the channel based on the 3D CAD model of the slurry valve; and generates the full-channel particle concentration distribution state sequence based on the geometric feature parameter sequence of the channel and the 3D particle enrichment index distribution map.

[0038] Scour intensity distribution field construction module: Based on the particle concentration distribution state sequence of the entire flow channel, a layered scour energy density distribution is generated; according to the layered scour energy density distribution, three scour modes, namely impact, abrasion and corrosion, are identified, the comprehensive scour intensity is calculated and a three-dimensional scour intensity distribution field is generated.

[0039] Topology optimization module: Based on the three-dimensional scour intensity distribution field, it identifies layered scour hotspots and formulates differentiated geometric adjustment strategies, performs topology optimization iterative feedback, and generates a scour-resistant optimized flow channel geometric model.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] This invention provides a method for accurately capturing the stratified flow characteristics and enrichment patterns of multi-size particles within the slurry valve channel by constructing a particle size grouping identification system and a three-dimensional particle enrichment index distribution map. This overcomes the prediction limitations caused by the assumption of uniform particle distribution in traditional CFD simulations. By combining the extraction of channel geometric feature parameter sequences and the generation of a full-channel particle concentration distribution state sequence, this invention achieves quantitative analysis of particle concentration distribution and scouring energy density distribution within the channel. It can effectively identify three scouring modes—impact, abrasion, and corrosion—and their coupling effects, providing data support for the formulation of differentiated geometric adjustment strategies. Through a topology optimization iterative feedback mechanism, this invention specifically optimizes the channel geometry, reducing the scouring intensity of high-concentration particle layers in areas such as the valve body bottom and corners. This significantly extends the service life of the slurry valve under harsh operating conditions in industries such as mining, chemical, and environmental protection, improves production continuity, and reduces maintenance costs. It effectively solves the problem of early failure caused by inaccurate prediction of local scouring hotspots, enhancing the overall reliability and economy of the slurry valve. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of the CFD-based slurry valve anti-erosion flow channel topology optimization design method in this invention.

[0044] Figure 2 A flowchart of a method for establishing a sequence of geometric feature parameters and marking geometric types provided in an embodiment of the present invention;

[0045] Figure 3 A schematic diagram illustrating the principle of constructing a geometry-based correction model as provided in an embodiment of the present invention;

[0046] Figure 4 This is a functional block diagram of the CFD-based slurry valve anti-erosion flow channel topology optimization design device in this invention. Detailed Implementation

[0047] 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.

[0048] Example 1:

[0049] Please see Figure 1 As shown, this embodiment provides a CFD-based method for optimizing the topology of a slurry valve's anti-erosion flow channel, including:

[0050] Step S10: Obtain slurry particle size distribution data, establish a particle size grouping identification system based on the slurry particle size distribution data; construct a three-dimensional particle enrichment index distribution map based on the slurry particle size distribution data and the particle size grouping identification system.

[0051] Further, step S10 includes:

[0052] Step S11: Obtain particle size distribution data at the slurry valve inlet; based on the particle size distribution data, divide the particles into N particle size groups, calculate the sedimentation sensitivity value of each particle size group, and assign a unique digital identifier to each particle size group.

[0053] Step S12: Calculate the theoretical settling velocity for each particle size group, establish a flow velocity-stratification critical determination model, and generate quantitative indicators of the stratification degree for each particle size group.

[0054] Step S13: Discretize the vertical direction of the flow channel into M computational layers according to the stratification index, calculate the particle enrichment index of each particle size group in each computational layer, and construct a three-dimensional particle enrichment index distribution map.

[0055] Step S10 involves the digital extraction of gravity stratification features based on particle size gradient. By constructing a digital grouping model for multi-size particles, a critical judgment model, and an enrichment index map, it achieves accurate quantification of particle stratification behavior in complex flow channels, providing fundamental data support for subsequent flow channel topology optimization. Step S11 addresses the limitation of existing technologies in converting continuous particle size distribution into traceable units. It first acquires particle size distribution data at the slurry valve inlet. This data is obtained through measurements using equipment such as a laser particle size analyzer and includes information such as particle diameter and corresponding particle density. Considering that particle sizes in slurry often span multiple orders of magnitude, such as from 0.1 micrometers to 50 millimeters, linear interval grouping would result in overly dense grouping of small-size particles and sparse grouping of large-size particles, failing to evenly cover the characteristics of particles of different sizes. Therefore, a logarithmic interval principle is used for grouping: after sorting the particle sizes from smallest to largest, the logarithm is taken to base 10, ensuring that the particle size range of each group is evenly spaced on a logarithmic scale. For example, for particles with a size range of 0.1-1 micrometers, the logarithm is represented as log... 10 0.1 = -1 to log 10 1 = 0; for particles with a diameter range of 1-10 micrometers, its logarithm is expressed as log 10 1 = 0 to log 10 10 = 1, ultimately dividing the particles into N size groups. In this example, the number of size groups from 0.1 micrometers to 50 millimeters is N = 5. To enable dynamic tracking of each group, a unique numerical identifier "G" is assigned. i T j ", where "G" i "Identify the particle size group number, where i is the index of the particle size group, i = 1 to N, "T j "The system records the moment when particles enter the flow channel, with j being the timestamp number. This identification system ensures accurate location and state of a specific particle size group at a given moment in CFD simulations, resolving the ambiguity problem in particle mixing tracking using traditional methods. Furthermore, to quantify the settling tendency of each particle size group, the settling sensitivity value S is calculated." i , representing the sedimentation sensitivity value of particles in the i-th particle size group, is calculated using the following formula: , where d i It is the equivalent representative diameter of the particles in the i-th particle size group, calculated by the arithmetic mean or weighted average of the particle diameters within the group. For fluid density, Particle density, The fluid dynamic viscosity is obtained by measuring fluid density and particle density using a densitometer. The fluid dynamic viscosity is determined experimentally or by consulting a physical property handbook, depending on the fluid type, such as water or mud, and the temperature. The numerical characteristics directly reflect the particle settling ability: The larger the particle size, the easier it is for the particles to settle and stratify under gravity. For example, the S5 value of the ultra-coarse particle group G5 is the largest. This quantitative indicator provides a key parameter for subsequent stratification critical determination. Without this step, it is impossible to establish the correlation between particle size and sedimentation behavior, resulting in a lack of support for stratification judgment based on the characteristics of the particles themselves, and the stratification differences of different particle size groups cannot be reflected.

[0056] Step S12, building upon the grouping in S11, addresses the limitation of traditional methods in quantitatively determining particle stratification at different flow velocities by constructing a flow velocity-stratification critical determination digital model. First, the theoretical settling velocity v for each particle size group is calculated. settle,i , representing the theoretical settling velocity of the i-th particle size group. For small-diameter particles, Stokes' law is used for calculation, and the formula is: Where g is the acceleration due to gravity; for large-diameter particles, since they may deviate from the laminar settling state, Allen's formula is used for correction, and the coefficient is adjusted according to the Reynolds number range. Small-diameter particles correspond to particles with a Reynolds number Re ≤ 1, and large-diameter particles correspond to particles with a Reynolds number Re > 1.

[0057] Obtain the local flow velocity v at the current position within the flow channel. flow By extracting velocity field data from CFD simulations, the critical ratio R for the i-th particle size group is defined. i =v flow / v settle,i The critical ratio reflects the balance between fluid entrainment force and particle gravity settling force: R i The smaller the value, the more significant the effect of gravity, and the easier it is for the particles to stratify; R i The larger the particle size, the stronger the fluid entrainment and the easier it is for particles to mix. Based on multiple sets of experimental data, the actual stratification state of each particle size group was measured at different flow velocities to determine the critical stratification threshold R. critical and the mixed critical threshold R mix R critical Example is 0.8, R mix Example is 3.0; as shown in Table 1, when R i <R critical When R is significantly stratified, it is considered a significant stratification. i >R mix When R is fully mixed, it is considered fully mixed; when R is fully mixed, it is considered fully mixed. critical ≤R i ≤R mix At that time, the quantification index D was used to determine the degree of stratification. i Describe the transition state, D i The stratification index for the i-th particle size group is represented by the formula D. i =(R mix -R i ) / (R mix -R criticalThe value range is 0 to 1, where 0 corresponds to complete mixing and 1 corresponds to complete stratification. This formula achieves continuous quantification of the transition state through linear interpolation, avoiding the either-or problem in traditional qualitative description.

[0058] Table 1. Critical Determination Table for Flow Velocity-Stratification

[0059]

[0060] Step S13, based on the stratification index of S12, addresses the problem that traditional methods cannot intuitively display the differences in vertical particle distribution, and constructs a three-dimensional particle enrichment index distribution map in the vertical direction. The method for constructing the three-dimensional particle enrichment index distribution map includes: discretizing the flow channel vertically into M computational layers, numbered L1 to L2 from top to bottom. M The thickness of each layer is determined based on the maximum particle diameter, for example, by taking 10 times the maximum particle diameter, to ensure that each layer can accommodate at least a number of particles and avoid distortion in particle distribution calculations due to excessively thin layers. Let k be the index variable of the computational layer, 1≤k≤M, for each computational layer L k and particle size group G i Calculate the particle enrichment index E ik =C ik / C i,avg C ik For computation layer L k Inner particle size group G i The local concentration was extracted from the concentration field data of CFD simulation. i,avg For the particle size group G throughout the flow channel i The average concentration, E ik >1 indicates the computation layer L k The particle concentration is higher than average, i.e., enrichment, E ik A value less than 1 indicates below-average levels, i.e., scarcity. To reinforce the impact of stratification on the enrichment index, when the stratification quantification index D... i When it approaches 1, the enrichment index E of the lower layers iM =1+D i ×i, reflecting the strong enrichment characteristics of large-particle groups at the bottom layer, D i A value close to 1 indicates near-complete stratification, with larger particle sizes corresponding to higher i values; the top enrichment index E i1 =1-D i ×i / N reflects the scarcity of large particle size groups at the top layer. This formula, by introducing the particle size group number and the total number of groups, dynamically correlates the enrichment degree with particle size and stratification, solving the problem of neglecting particle size differences in traditional concentration distribution descriptions. Finally, a three-dimensional particle enrichment index distribution map is constructed, with the horizontal axis representing the particle size group number, the vertical axis representing the calculation layer number, and the value representing the particle enrichment index E. ikThe color coding visualization visually displays the distribution pattern of particles of different sizes in the vertical direction.

[0061] The synergistic effect of each sub-step in step S10 forms a complete hierarchical feature extraction system: the digital grouping in S11 transforms continuous particle size into discrete traceable units, providing a basic data structure for the critical determination in S12; the quantitative model in S12 correlates flow velocity with the degree of stratification, providing stratification intensity parameters for the enrichment index calculation in S13; and the three-dimensional particle enrichment index distribution map in S13 transforms the quantitative results of the first two steps into intuitive spatial distribution data, providing the input basis for the geometric constraint modeling in subsequent S20. This synergy not only achieves a complete mapping from particle characteristics to spatial distribution, but also allows logarithmic interval grouping to capture 10-100 micrometer particles, i.e., particle size group G3 in the example, and the abrupt stratification phenomenon at medium flow velocities. This is a key feature missed by traditional linear grouping due to its coarse grouping. The combination of sedimentation sensitivity and enrichment index reveals that although ultrafine particles have small S1 values ​​at high flow velocities, such as G1 in the example, they may still form enrichment in local areas due to fluid turbulence, breaking the traditional understanding that "ultrafine particles are always uniformly mixed". The absence of any sub-step in S10 will cause the entire technology chain to break: without S11, accurate particle grouping and tracking cannot be achieved, and the critical judgment in S12 will lose its target; without S12, the impact of flow rate on stratification cannot be quantified, and the enrichment index in S13 will lack a basis for dynamic adjustment; without S13, stratified data cannot be transformed into spatial distribution, and the subsequent geometric coupling modeling in S20 will lose the input of particle concentration field, ultimately causing the flow channel optimization to fail due to the lack of support from microscopic particle behavior.

[0062] Step S20: Based on the three-dimensional CAD model of the slurry valve, extract the sequence of geometric feature parameters of the flow channel; based on the sequence of geometric feature parameters of the flow channel and the three-dimensional particle enrichment index distribution map, generate the particle concentration distribution state sequence of the entire flow channel.

[0063] Further, step S20 includes:

[0064] Step S21: From the inlet to the outlet of the slurry valve, set P geometric feature extraction sections along the center line of the flow channel; extract the geometric feature parameters of each geometric feature extraction section and establish a geometric feature parameter sequence; mark the geometric type of the geometric feature extraction sections according to the geometric feature parameter sequence.

[0065] Please see Figure 2 As shown, step S21 further includes:

[0066] Step S211: Establish a curved coordinate system along the centerline of the flow channel;

[0067] Step S212: From the inlet to the outlet of the slurry valve, a geometric feature extraction section is set at a preset distance, for a total of P geometric feature extraction sections;

[0068] Step S213: Based on the 3D CAD model of the slurry valve, extract the geometric feature parameters of each geometric feature section. The geometric feature parameters include the centroid coordinates of the section, the area of ​​the section, and the equivalent diameter of the section.

[0069] Step S214: Extract the geometric feature parameters of the cross section based on each geometric feature, and calculate the geometric change parameters between adjacent cross sections, including the radius of curvature, shrinkage ratio, and expansion angle; the geometric feature parameters and geometric change parameters constitute a geometric feature parameter sequence;

[0070] Step S215: Mark the geometric type of the extracted cross section according to the geometric feature parameter sequence. The geometric type includes straight pipe section, curved section, contraction section and expansion section.

[0071] Step S21 serves as the foundation for dynamic coupling modeling of particle distribution under flow channel geometric constraints. It aims to address the problem that traditional methods cannot accurately quantify complex flow channel geometries into particle behavior analysis parameters. By extracting flow channel geometric features and transforming them into a sequence of computable parameters, it provides a quantitative basis for subsequent correlation analysis between geometry and particle behavior. Specifically, S211 establishes a curvilinear coordinate system along the flow channel centerline. The origin of the curvilinear coordinate system is set as the centroid of the flow channel inlet section. The axial coordinates increase from the inlet to the outlet along the centerline, while the radial coordinates are perpendicular to the centerline and point towards the flow channel wall. Compared to the traditional rectangular coordinate system, this coordinate system better reflects the natural flow channel direction and can more accurately reflect the influence of geometric changes such as flow channel bending and contraction on particle motion. For example, when analyzing a 90-degree bend, the axial direction of the curvilinear coordinate system can naturally follow the curvature of the bend, avoiding geometric feature distortion caused by coordinate transformation in the rectangular coordinate system. S212 sets geometric feature extraction sections at preset distances from the inlet to the outlet. The selection of preset distances must balance calculation accuracy and efficiency. If there are obvious geometric abrupt changes in the flow channel, such as at the valve opening and closing points, the sections are densified before and after the abrupt changes to ensure that local geometric details can be captured. A total of P sections are set to form a uniform coverage of the entire flow channel. S213 Based on the 3D CAD model of the slurry valve, the geometric feature parameters of each section are extracted. The geometric feature parameters include the centroid coordinates of the section, the section area, and the equivalent diameter of the section. Among them, the centroid coordinates of the section are directly obtained through the section analysis function of the CAD software and are used to locate the position of the section in the curvilinear coordinate system. The section area is obtained by calculating the area of ​​the region enclosed by the section outline. The equivalent diameter of the section is calculated by dividing 4 times the section area by the wetted perimeter. The wetted perimeter is the circumference of the section in contact with the flow channel wall, which can equivalently reflect the flow capacity of the flow channel. S214 calculates the geometric change parameters between adjacent cross-sections. The radius of curvature is obtained by fitting a circular arc to the centroid coordinates of three adjacent cross-sections. These three points determine a unique arc, and its radius is the radius of curvature at that location, used to determine the degree of channel curvature. The contraction ratio is the ratio of the current cross-sectional area to the previous cross-sectional area; a value less than 1 indicates contraction, and a value greater than 1 indicates expansion. The expansion angle is calculated by taking the tangent of the ratio of the change in the equivalent diameter of adjacent cross-sections to the distance between cross-sections and converting it to an angle, used to quantify the severity of expansion. Based on these parameters, S214 marks each cross-section with a geometric type, as shown in Table 2. When the radius of curvature is less than 5 times the equivalent diameter, it is marked as a curved section, where centrifugal force significantly affects the particles. When the contraction ratio is less than 0.8, it is marked as a contraction section, where the flow velocity increases significantly. When the contraction ratio is greater than 1.2, it is marked as an expansion section, where the flow velocity decreases significantly. The rest are straight pipe sections. The sequence of geometric characteristic parameters formed by the geometric characteristic parameters and geometric change parameters completely records the details of the geometric changes in the channel from the inlet to the outlet.

[0072] Table 2 Criteria for Judging Flow Channel Geometry Type

[0073]

[0074] Step S21 establishes a curvilinear coordinate system, achieving precise matching between geometric parameters and the actual flow path, avoiding calculation errors when analyzing curved flow channels in traditional rectangular coordinate systems. The dynamic adjustment mechanism of cross-sectional spacing ensures analytical accuracy in areas of geometric abrupt changes while reducing computation in smoother regions, achieving a balance between accuracy and efficiency. The extraction of multi-dimensional geometric parameters comprehensively characterizes the geometric properties of the flow channel, including parameters of the cross-section itself and parameters of changes between adjacent cross-sections. Clear labeling of geometric types provides a basis for subsequently adopting differentiated particle behavior models for different geometric segments. This parameterized extraction method can identify "hidden geometric abrupt changes" in the flow channel. Some cross-sections, viewed individually as straight pipe sections, exhibit a large rate of change in the radius of curvature of adjacent cross-sections, indicating a slight bending trend. Such subtle changes are often overlooked in traditional methods but can lead to local anomalies in particle distribution. The continuous parameter sequence extracted in S21 captures this feature, providing crucial clues for subsequent accurate analysis of particle behavior. Without S21, the flow channel geometry can only exist as a qualitative description, such as "there is a bend in the pipe," and cannot be converted into quantitative parameters. As a result, S22 cannot establish a correlation model between geometry and particle behavior, and the dynamic coupling modeling of the entire S20 will lose its foundation. Consequently, the subsequent scouring analysis will be distorted due to the lack of accurate particle distribution data.

[0075] Step S22: Based on the geometric feature parameter sequence, the labeled geometric type, and the three-dimensional particle enrichment index distribution map, construct a geometric type-based correction model;

[0076] Please see Figure 3 As shown, step S22 further includes:

[0077] Step S221: For the straight pipe section, establish a gravity settlement cumulative correction model for the straight pipe section based on the settlement sensitivity value and particle enrichment index.

[0078] Step S222: For the curved section, based on the radius of curvature, the value of sedimentation sensitivity and the particle enrichment index, a collaborative correction model of centrifugal force offset and enrichment for the curved section is established.

[0079] Step S223: For the shrinkage segment, based on the shrinkage ratio, sedimentation sensitivity value, and particle enrichment index, establish a synergistic correction model for the acceleration effect and stratification inhibition of the shrinkage segment.

[0080] Step S224: For the expansion section, based on the expansion angle, sedimentation sensitivity value, and particle enrichment index, establish a collaborative correction model for the deceleration effect and stratification enhancement of the expansion section.

[0081] Specifically, step S22, based on the extracted geometric feature parameter sequence and the three-dimensional particle enrichment index distribution map, addresses the problem that traditional methods cannot quantitatively describe the influence of geometry on particle distribution. By constructing correction models for different geometric types and integrating them into a response matrix, a dynamic correlation between geometric features and particle behavior is achieved. Specifically, step S221, by constructing a gravity settlement accumulation correction model for straight pipe sections, dynamically quantifies the gravity stratification effect of particles in straight pipe sections. This model is based on settlement sensitivity numerical values. Particle enrichment index Straight pipe section length Equivalent diameter of straight pipe section Constructing the correction formula .in, This is a correction value for the degree of stratification in straight pipe sections. The initial stratification degree at the inlet is a quantification index derived from step S12. 0.02 is an empirical coefficient fitted through multiple sets of straight pipe section particle settling experiments, reflecting the cumulative stratification rate per unit length. The equivalent diameter of the straight pipe section is... The equivalent diameter of the cross-section extracted in step S21 is obtained by averaging. The cumulative correction model for gravity settlement in straight pipe sections quantifies the cumulative effect of gravity on the flow distance in straight pipe sections. For example, when... Larger and At higher levels, the degree of stratification It will increase significantly, while larger This will weaken the effect. This correction solves the problem of inaccurate particle distribution prediction in straight pipe sections in traditional methods, enabling subsequent scour analysis to more accurately capture the scour contribution of high-concentration large particles in the bottom layer.

[0082] For the curved segment, step S222 achieves a quantitative description of the influence of the curved segment's geometry on particle distribution by constructing a collaborative correction model of centrifugal force offset and enrichment in the curved segment. This model is based on the radius of curvature of the curved segment. Local flow velocity v in the curved section p Particle size group G i average diameter d i Settling sensitivity and particle enrichment index Construct a collaborative correction model for centrifugal force offset and enrichment in curved sections. ,in, The radial offset is given by g, where g is the acceleration due to gravity, and the radius of curvature of the curved segment is given by g. The sequence of geometric feature parameters from step S21, and the local flow velocity v in the curved section. pExtracted through CFD simulation. This model addresses the shortcomings of traditional methods that rely solely on qualitative analysis for predicting particle distribution in curved sections by quantifying the synergistic effect of centrifugal force, particle characteristics, and initial distribution. The synergistic correction model of centrifugal force migration and enrichment in curved sections couples the centrifugal force effect with the particle's intrinsic characteristics and initial distribution state. The centrifugal force effect refers to... and Proportional to, and Inversely proportional, the inherent characteristics of the particles are reflected in their sedimentation sensitivity, and the initial distribution state is determined by... This is reflected, for example, when the radius of curvature... When the particles are small, such as in a 90-degree bend, the centrifugal force is significantly enhanced, while for large-diameter particles, the centrifugal force is significantly enhanced. Larger size and higher settlement sensitivity, i.e. Larger particles have greater inertia and therefore greater radial offset. It will increase further. Meanwhile, the particle enrichment index... A value greater than 1 indicates that the particles in this layer are dense, and the concentration increase on the outer side after shifting is achieved through... Equivalent diameter D of straight pipe section p The ratio is multiplied by an enrichment amplification factor of 1.5, which is determined by fitting multiple sets of particle distribution experiments in the curved section, reflecting the nonlinear effect of particle aggregation on the outer side. A multi-parameter coupled model of centrifugal force shift and enrichment synergistic correction in the curved section quantifies the synergistic effect of centrifugal force on particle characteristics and initial distribution. For example, under conditions of small radius of curvature and high flow velocity, large-diameter particles will significantly shift outwards, leading to a substantial increase in outer wall concentration—an effect often underestimated in traditional methods. This correction significantly improves the accuracy of particle distribution prediction in the curved section, providing more accurate particle concentration distribution data for scour analysis.

[0083] Step S223 constructs a synergistic correction model for acceleration effect and stratification suppression in the contraction section. This aims to address the limitation of traditional methods that only qualitatively describe the impact of the contraction section on particle distribution using geometric parameters (such as the contraction ratio), failing to quantify the synergistic effect of particle characteristics (sedimentation sensitivity) and initial distribution state (enrichment index) on stratification suppression. In the contraction section, the reduced cross-sectional area leads to a significant increase in local flow velocity and enhanced fluid entrainment, which weakens the stratification tendency of particles caused by gravity. However, the degree of influence varies significantly among particles of different sizes and with different initial concentrations. Traditional methods ignore this difference, leading to biased particle distribution predictions. This step introduces the contraction ratio. Based on sedimentation sensitivity values ​​and particle enrichment index, a synergistic correction model for acceleration effect and stratification inhibition was constructed. .in, D represents the degree of stratification after correction of the contraction segment. i For particle size group G iThe degree of stratification before shrinkage is derived from the stratification quantification index in step S12, with 0.3 representing the inhibition coefficient of the experimental fit. It comes from the sequence of geometric feature parameters. This term is used to quantify the basic inhibitory effect of the acceleration effect on stratification. When the contraction ratio is 0.5, this term is 1.5, indicating that the acceleration effect enhances the stratification inhibition effect by 50%. When the contraction ratio is close to 1 (i.e. close to the non-contraction state), this term is close to 1, and the inhibition effect approaches zero, which is consistent with the influence of flow velocity change on entrainment force. This term quantifies the synergistic effect of particle characteristics and initial distribution, where 0.3 is the inhibition coefficient. This inhibition coefficient is obtained by fitting multiple sets of experimental data from the contraction section: flow channels with different contraction ratios are selected, and the changes in stratification degree of different particle size groups under different initial enrichment states are measured. The coefficient is obtained by fitting using the least squares method, ensuring that the deviation between the calculated result and the experimental measurement value is controlled within 5%. When the particle settling sensitivity is high and the particle is in an enriched state, A decrease in the numerical value indicates a reduction in the degree of stratification after correction, suggesting that the stratification trend of these particles is more significantly suppressed. This is because large-diameter particles are inherently prone to sedimentation, and in the enriched state, the interaction between particles is enhanced, making it easier for the accelerated fluid to simultaneously entrain these particles, thus weakening stratification. When the particle sedimentation sensitivity is low or the enrichment index is close to 1, i.e., the concentration is uniform, When the value is close to 1, the degree of stratification after correction is mainly dominated by the shrinkage ratio, which is consistent with the characteristics that small-diameter particles are not easily stratified and are more uniformly affected by the acceleration effect when evenly distributed. The synergistic correction model of acceleration effect and stratification suppression achieves the synergistic coupling of geometric parameter shrinkage ratio, particle characteristics and initial distribution state. Compared with the traditional model that only relies on shrinkage ratio, the prediction accuracy is improved. It reveals the nonlinear characteristics of particle stratification suppression in the shrinkage segment that the traditional linear model cannot capture. When the particle enrichment index is less than 1, that is, in a scarce state, The value is greater than 1, meaning that the degree of stratification after correction will be slightly higher than that before contraction. This indicates that in the contraction section, the sparse particles may exhibit slight enhanced stratification due to insufficient entrainment by the surrounding fluid. This is different from the traditional understanding that the contraction section necessarily inhibits stratification, and is of great significance for the scouring analysis of low-concentration particles.

[0084] Step S224 constructs a synergistic correction model for the deceleration effect and stratification enhancement in the expansion section. This aims to address the technical challenge of traditional methods failing to quantify the enhancing effect of velocity deceleration in the expansion section on particle stratification, and the synergistic influence of particle characteristics and initial distribution on this process. In the expansion section, the increased cross-sectional area leads to a decrease in local velocity and a weakening of fluid entrainment, thus reinforcing the stratification tendency of particles due to gravity. However, the degree of influence varies significantly among particles of different sizes and with different initial concentrations. Traditional methods only qualitatively describe this effect through geometric parameters, ignoring the differences in particle characteristics and initial distribution, leading to biased particle distribution predictions. This step introduces the expansion angle. (Calculated from the geometric characteristic parameter sequence in step S21 using the equivalent diameter change of adjacent sections and the section spacing), settlement sensitivity, and particle enrichment index, to construct a synergistic correction model for deceleration effect and stratification enhancement. .in, To enhance the degree of post-layering, the expansion angle From the sequence of geometric characteristic parameters, in the actual flow channel Typically ≤30° to avoid errors caused by flow separation. The value range of the term is 1 to 1.33. When = 30°, A value of 1.33 indicates that the deceleration effect enhances stratification by 33%; when Approaching 0° When the value approaches 1, the enhancing effect approaches zero. For the particle size group G before expansion i The degree of stratification before shrinkage, with 0.2 being the enhancement coefficient for experimental fitting.

[0085] in, This term is used to quantify the basic enhancement effect of deceleration on stratification. When the expansion angle is 30°, the factor is 1.33, indicating that the deceleration effect enhances stratification by 33%. When the expansion angle is close to 0°, that is, close to the non-expansion state, the term is close to 1, and the enhancement effect approaches zero, which is consistent with the influence of flow velocity change on entrainment force. This term quantifies the synergistic effect between particle characteristics and initial distribution, where 0.2 is the enhancement coefficient. This enhancement coefficient is obtained by fitting multiple sets of experimental data from the expansion section: selecting flow channels with different expansion angles, measuring the changes in stratification degree of different particle size groups under different initial enrichment states, and obtaining the coefficient through least squares fitting to ensure that the deviation between the calculated result and the experimental measurement value is controlled within 5%. When the particle settling sensitivity is high and the particle is in an enriched state, the value of this factor increases, and the stratification degree after enhancement is significantly improved. This is because large-diameter particles are inherently easy to settle, and the interaction between particles is enhanced in the enriched state, making it easier for the decelerated fluid to release its settling tendency, leading to intensified stratification. When the particle settling sensitivity is low or the enrichment index is close to 1, the factor is close to 1, and the stratification degree after enhancement is mainly dominated by the expansion angle, which is consistent with the characteristics that small-diameter particles are not easy to stratify and are more uniformly affected by the deceleration effect when evenly distributed. The synergistic correction model of expansion section deceleration effect and stratification enhancement quantifies the synergistic effect between the expansion section deceleration effect and particle characteristics and initial distribution, solving the problem of vague description of stratification enhancement effect in traditional methods, and enabling scour analysis to more accurately reflect the changes in particle distribution in the expansion section.

[0086] Step S22, through the collaborative modeling of multi-dimensional geometric parameters and particle characteristics, solves the problem of ambiguity in the relationship between geometry and particle behavior in traditional methods, significantly improving the accuracy of particle distribution prediction. Differentiated correction models can be applied to different geometric segments; for example, centrifugal force offset is emphasized in curved segments, while acceleration effect suppression is emphasized in contraction segments. This differentiated approach avoids the "one-size-fits-all" flaw of traditional methods. Furthermore, the particle size distribution, sedimentation sensitivity, and enrichment index provided in step S10 serve as inputs to the correction model, while the geometric feature parameters provided in step S21 form the geometric basis of the model. The combination of these three elements achieves a complete mapping from particle characteristics to geometric constraints. Without step S22, the influence of channel geometry on particle distribution cannot be quantified, subsequent scouring analysis will be distorted due to a lack of accurate particle distribution data, and the entire optimization process will lose crucial support.

[0087] Step S23: Based on the geometry-based correction model, establish a multi-segment flow channel particle distribution state transfer algorithm to generate a full flow channel particle concentration distribution state sequence containing the concentration distribution of each particle size group at each location.

[0088] Further, step S23 includes:

[0089] Step S231: Divide the entire flow channel into Q geometric segments, and construct the particle distribution state vector of the upstream geometric segment based on the particle enrichment index of each particle size group in each calculation layer in each geometric segment.

[0090] Step S232: Construct a distribution state transfer matrix between geometric segments based on the geometry type-based correction model to describe the evolution of particle distribution from the upstream geometric segment to the downstream geometric segment;

[0091] Step S233: Calculate the particle distribution state vector of the downstream geometric segment through matrix operation between the distribution state transfer matrix and the particle distribution state vector of the upstream geometric segment.

[0092] Step S234: The particle distribution state vector of the downstream geometric segment is used as the particle distribution state vector of the new upstream geometric segment. The particle distribution state vector and the distribution state transfer matrix of the upstream geometric segment are applied sequentially along the flow channel direction to complete the full flow channel particle distribution state transfer calculation from the inlet to the outlet, which includes all geometric segments.

[0093] Step S235: Integrate the full-channel particle distribution state transfer calculation results of all geometric segments to generate a full-channel particle concentration distribution state sequence containing the concentration distribution of each particle size group at each location.

[0094] Step S23 establishes a multi-segment flow channel particle distribution state transfer algorithm based on the geometry-type-based correction model. This aims to address the problem that traditional methods cannot quantify the continuous evolution of particle distribution across different geometric segments, particularly the issue of distorted particle distribution predictions across the entire flow channel due to the neglect of transition effects at geometric abrupt changes. Existing techniques often treat the flow channel as an isolated geometric segment for analysis, ignoring the correlation between particle distribution segments and failing to consider response lag caused by particle inertia, resulting in significant deviations between predicted and actual distributions.

[0095] Specifically, step S231 divides the entire flow channel into Q geometric segments, based on the abrupt change points of the geometric types marked in step S21. These abrupt change points are the boundaries between different geometric types, such as the boundary between a straight pipe segment and a curved segment. Each geometric segment contains a continuous geometric feature extraction section. For the inlet of each geometric segment, a particle distribution state vector is constructed. The dimension of the particle distribution state vector is the product of the number of particle size groups N and the number of computational layers M. The vector elements are the particle enrichment indices of each particle size group in each computational layer. q is defined as the index variable of the geometric segment. For the first geometric segment at the flow channel inlet, q=1, its initial particle distribution state vector is set based on the three-dimensional particle enrichment index distribution map in step S13. It is assumed that the particle distribution at the inlet is relatively uniform, and the initial value of the particle enrichment index is close to 1, because the particles are not affected by geometric constraints when they first enter the flow channel. For subsequent geometric segments (q>1), their inlet particle distribution state vector is the outlet particle distribution state vector of the previous geometric segment, ensuring the continuity of distribution evolution.

[0096] The dimension of the distribution state transfer matrix is ​​the same as that of the particle distribution state vector. Each element in the matrix corresponds to a particle distribution correction coefficient for a specific particle size group in a specific computational layer. These correction coefficients are derived from the correction model. For example, for a curved section, the element value corresponding to the large particle size group in the distribution state transfer matrix in the outer computational layer will be greater than 1. This is because the centrifugal force in the curved section causes this particle size group to be enriched on the outside, while the element value corresponding to the inner computational layer will be less than 1, reflecting the decrease in concentration on the inside after the particles shift outward. For a contracting section, the element value corresponding to the large particle size group in the bottom computational layer in the distribution state transfer matrix will be less than 1, because the acceleration effect of the contracting section inhibits stratification and reduces the enrichment degree of the bottom layer. The construction of the distribution state transfer matrix needs to be combined with the specific parameters of the geometric segment. For example, the smaller the radius of curvature of the curved section, the larger the element value corresponding to the outer computational layer, reflecting a stronger centrifugal force. The particle distribution state vector of the downstream geometric segment is the matrix product of the upstream particle distribution state vector and the distribution state transfer matrix.

[0097] Step S234 applies the above matrix operations sequentially along the flow channel direction to complete the calculation of the particle distribution state transfer throughout the flow channel. The key lies in introducing a "memory effect" correction, meaning that the particle distribution's response to geometric changes has a lag, with the lag distance... The calculation formula is: ,in, For localized flow velocities, the larger particle size group exhibits greater inertia. Larger, small particle size group The hysteresis distance is relatively small. In the calculation, the hysteresis distance is realized through the effective position of the delayed distribution state transfer matrix. For example, the centrifugal force effect at the starting point of a certain bend section only fully reflects the influence of G5 after the hysteresis distance, while the first half still retains some distribution characteristics of the straight pipe section. The particle concentration distribution state sequence of the entire flow channel is arranged in the order from the inlet to the outlet of the flow channel, containing the particle distribution state vector at the outlet section of each geometric segment. By multiplying the particle enrichment index in the vector by the average concentration of that particle size group in the entire flow channel, the local concentration of each calculation layer is obtained, thus forming concentration distribution data containing three-dimensional information of location, particle size group, and calculation layer.

[0098] Step S23 divides the flow channel into geometric segments and constructs a distribution state transfer matrix to realize the continuous evolution simulation of particle distribution under different geometric conditions. This overcomes the shortcomings of traditional methods that treat the flow channel as a whole and cannot capture local geometric influences. The introduction of the memory effect takes into account the hysteresis of particle inertia to geometric response, making the prediction more in line with the actual flow process. For example, in a short curved section, large particles enter the next geometric segment before completing radial offset. Traditional methods will overestimate their offset, while this step can accurately reflect this incomplete response.

[0099] Step S20 addresses the difficulty of quantifying the constraint effect of channel geometry on particle stratification distribution. Through the synergistic effect of S21-S23, a precise mapping from three-dimensional geometry to particle distribution is achieved. Specifically, S21 extracts a sequence of geometric feature parameters, discretizing the continuous channel geometry into a set of computable parameters. This overcomes the limitation of traditional methods in converting complex geometry into quantitative indicators. For example, continuous recording of the radius of curvature and contraction ratio captures minute geometric changes in the channel. The modified model constructed in S22 quantifies the correlation between different geometric types and particle behavior, upgrading the influence of geometric parameters on particle distribution from qualitative description to quantitative calculation. The state transfer algorithm in S23 connects the results of the preceding steps, simulating the continuous evolution of particle distribution along the channel, solving the problem of the inability to simulate the transition effect of particle distribution between geometric segments. The combination of the modified model and the transfer algorithm allows the effects of different geometric types to be superimposed and calculated. For example, the gravity settling of the straight pipe section and the centrifugal force of the subsequent curved section can act synergistically, making the enrichment of large particles in the outer bottom layer of the curved section far exceed the sum of the individual effects. If step S20 is missing, the influence of channel geometry on particle distribution cannot be quantified, the layered feature data obtained in step S10 cannot be applied to actual channel analysis due to the lack of geometric constraints, and the scouring calculation in step S30 will also become meaningless due to the distortion of particle distribution data, resulting in a break in the technical chain of the entire optimization scheme.

[0100] Step S30: Based on the full-channel particle concentration distribution state sequence, generate a layered scouring energy density distribution; based on the layered scouring energy density distribution, identify three scouring modes: impact, abrasion, and corrosion, calculate the comprehensive scouring intensity, and generate a three-dimensional scouring intensity distribution field.

[0101] Further, step S30 includes:

[0102] Step S31: Based on the particle concentration distribution state sequence of the entire flow channel, calculate the single particle kinetic energy and impact frequency of each particle size group in each calculation layer to generate the layered scouring energy density distribution.

[0103] Step S32: Based on the layered scour energy density distribution, construct the scour contribution step function of the vertical calculation layer, introduce the interface enhancement effect correction, and generate a continuous scour contribution distribution function.

[0104] Step S33: Based on the scour contribution distribution function, identify the three scour modes of impact, abrasion, and corrosion, calculate the comprehensive scour intensity, and generate a three-dimensional scour intensity distribution field.

[0105] Step S30, based on the full-channel particle concentration distribution state sequence, quantifies the scouring contribution of particles at different levels and with different sizes. This solves the problem that traditional methods cannot distinguish the differences in scouring mechanisms between high-concentration large particles at the bottom layer and low-concentration small particles at the top layer, and it is difficult to accurately quantify the contribution of each level of particles to wall scouring. Specifically, steps S31 to S33 achieve a precise mapping from particle distribution to scouring intensity through synergistic effects. Step S31 addresses the deficiency of traditional scouring energy calculations that ignore differences in particle size and concentration levels. Based on the full-channel particle concentration distribution state sequence generated in step S23, it calculates the kinetic energy and impact frequency of single particles in each calculation layer for each particle size group, thereby generating a layered scouring energy density distribution. The calculation of single particle kinetic energy requires consideration of particle physical properties and flow state. For particle size group G... i In computation layer L k The kinetic energy of a single particle is half the product of its mass and the square of its velocity. The particle mass is calculated from the particle density and volume, measured by a particle density meter, and the particle volume is based on the particle size group G. i The equivalent representative diameter d i Calculated using the formula for the volume of a sphere; the particle velocity is taken as the local flow velocity v at that location in the flow channel. flow With particle settling velocity v settle,i The vector composite value, because particles have a settling velocity in the vertical direction due to gravity and a flow velocity in the horizontal direction with the fluid, more accurately reflects the actual velocity of particles impacting the wall. Impact frequency represents the number of impacts per unit area of ​​the wall per unit time by particles of that size group, and it is related to the calculated layer L. k Particle concentration C ik The particle velocity and diameter are related, and the calculation formula is the product of particle concentration, particle velocity, and the square of particle diameter, divided by particle volume. This formula is derived from the spatial distribution density and trajectory of particles within the flow channel. Higher concentrations, faster velocities, and larger particles result in more impacts on the wall per unit time. The stratified scouring energy density is the product of a single particle's kinetic energy and its impact frequency. Its physical meaning is the energy received per unit area of ​​the wall per unit time from particles of size G. i In computation layer L k The scouring energy and stratified scouring energy density, by distinguishing different computational layers and particle size groups, achieve a quantitative differentiation between the bottom layer of large particles, such as particle size group G5, which experience high kinetic energy and high frequency impacts, and the upper layer of small particles, such as particle size group G1, which experience low kinetic energy and low frequency impacts. For example, the bottom layer G5 particles, due to d i Big, C ikThe high density of the stratified scouring energy in the upper G1 layer provides an energy basis for distinguishing scouring modes. Without step S31, it is impossible to establish the correlation between particle distribution and scouring energy, resulting in a lack of quantitative basis for subsequent scouring intensity calculations and failing to reflect the energy differences between different particle levels, causing scouring analysis to fall into the "averaging" trap of traditional methods.

[0106] Step S32 addresses the discontinuity caused by vertical hierarchical division in traditional scour contribution calculations. Based on the layered scour energy density distribution generated in step S31, it constructs a step function for the scour contribution of the vertical calculation layers and introduces interface enhancement effect correction to generate a continuous scour contribution distribution function. The step function is constructed on a per-layer basis, defining the scour contribution of each calculation layer as the ratio of the layered scour energy density of that layer to the average scour energy density of the entire flow channel. For the bottom calculation layer, such as L... M Due to the enrichment of large particles, their initial contribution is usually greater than 1. For the top computational layer, such as L1, where small particles dominate, the initial contribution is usually less than 1. Although this step-like division can reflect the differences between layers, it ignores the scouring enhancement phenomenon caused by particle interaction at the interface of adjacent computational layers. The interface enhancement effect originates from the difference in the direction of particle movement between adjacent layers. For example, in the middle computational layers, such as the interface between L3 and L4, the particles have both horizontal velocity with the fluid and vertical velocity components due to the velocity gradient between the upper and lower layers, causing the particles to collide obliquely at the interface. The scouring energy is higher than that of the forward collision within a single computational layer. The correction method is to introduce an enhancement coefficient into the interface region of adjacent computational layers based on the step function. The value of the enhancement coefficient is positively correlated with the difference in energy density between the two layers. The larger the difference, the more intense the particle interaction at the interface, and the larger the enhancement coefficient. For example, when the energy density of L3 is 1.5 times that of L4, the enhancement coefficient is 1.2. By calculating the contribution of the interface region as the product of the weighted average of the contributions of the two layers and the enhancement coefficient, the continuity of the step function is achieved. The continuous scour contribution distribution function solves the problem of abrupt changes between layers in the traditional step function, making the scour contribution in the vertical direction present a smooth transition. For example, at the middle layer interface on the outer side of the curved section, this function can capture the local scour energy peak caused by the combined effect of centrifugal force and gravity, a feature that traditional methods would miss due to layer division. The synergy between steps S32 and S31 transforms the layered scour energy density from discrete interlayer data into a continuous spatial distribution, providing a continuous foundation for the subsequent generation of the three-dimensional scour intensity field. Without step S32, the scour contribution at the interlayer interface would be underestimated, leading to deviations in scour hotspot identification, especially at geometric abrupt changes where the interface effect is significant. Ignoring correction would cause subsequent optimization strategies to lose accuracy.

[0107] Step S33, based on a continuous scour contribution distribution function, identifies three scour modes: impact, abrasion, and corrosion. It calculates the comprehensive scour intensity and generates a three-dimensional scour intensity distribution field, solving the problems of fuzzy scour mode identification and lack of theoretical basis for comprehensive intensity calculation in traditional methods. The impact mode is identified based on whether the product of single-particle kinetic energy and impact frequency exceeds a first threshold. This first threshold is determined experimentally. When particle kinetic energy is high and impact frequency is high, such as large particles at high speed impacting the wall surface, the wall material undergoes plastic deformation or spalling due to the instantaneous impact force, and this is determined to be an impact mode.

[0108] The determination of the erosion mode requires that the kinetic energy of a single particle be within a moderate range and the impact frequency be higher than a second threshold. The definition of the moderate kinetic energy range is related to the first threshold of the impact mode, and is typically set to 30% to 70% of the single-particle kinetic energy threshold of the impact mode. This range is determined based on the differences in the mechanism of particle interaction with the material surface. When the kinetic energy is below 30%, the particles are unlikely to generate significant cutting force, while above 70%, they are more likely to induce impact deformation. At moderate kinetic energy, particles are more likely to slide or roll along the wall surface, generating a cutting effect through continuous friction. The determination of the corrosion mode uses a dual standard of extremely low particle concentration and critically high chemical activity of the fluid medium. The quantitative index for extremely low particle concentration is based on the particle enrichment index E. ik , when E ik When the value is ≤0.3, it indicates that the particle concentration in the calculation layer is much lower than the average level of the entire flow channel. The physical shielding effect of the particles on the wall is weak and cannot prevent the direct contact between the fluid medium and the wall.

[0109] The abrasion mode corresponds to situations where the kinetic energy of a single particle is moderate but the impact frequency is high. For example, when small particles in the middle layer continuously rub against the wall, the cutting action of the particles on the wall dominates the scouring process. The threshold for this mode is when the kinetic energy is lower than that of the impact mode but the frequency is higher than the second threshold. The first threshold for the impact mode is set as the energy density value that causes obvious pitting on the low-carbon steel specimen within 1 hour. The second threshold for the abrasion mode is set as the frequency value that increases the surface roughness of the specimen by 0.5 μm / h. The specific threshold can be adjusted according to the characteristics of the flow channel material. The calculation of comprehensive erosion intensity requires weighting the contributions of the three modes. The weights are determined based on the energy proportion of each mode at that location. The impact mode, due to its most severe destructive effect on the material, has the highest weight, followed by the abrasion mode, and the corrosion mode has the lowest weight. The calculation formula is the sum of the products of impact intensity, abrasion intensity, and corrosion intensity with their respective weights. Impact intensity is the ratio of energy density under the impact mode to the material's impact resistance coefficient; abrasion intensity is the ratio of energy density under the abrasion mode to the material's wear resistance coefficient; and corrosion intensity is the ratio of the medium concentration under the corrosion mode to the material's corrosion resistance coefficient. The material's impact resistance coefficient, wear resistance coefficient, and corrosion resistance coefficient are obtained through material testing. For example, stainless steel has a higher impact resistance coefficient than cast iron, therefore its impact intensity is lower at the same energy density. The three-dimensional erosion intensity distribution field is generated by linking the comprehensive erosion intensity to a curve coordinate system. Color coding visually displays the intensity levels. For example, the outer bottom layer of the curved section, dominated by the impact mode, presents a red high-intensity area, while the top layer of the straight pipe section, dominated by the corrosion mode, presents a blue low-intensity area. This step, in conjunction with steps S31 and S32, achieves a complete mapping from particle energy to material damage. Step S31 provides layered energy data, laying the foundation for pattern recognition; step S32's continuousization function ensures the spatial continuity of the three-dimensional field. The combination of these three steps reveals a "hybrid erosion effect" undetected by traditional methods. In certain geometric segments, impact and abrasion modes coexist, and their combined erosion intensity is not simply additive but significantly enhanced by synergistic effects. This stems from the alternating effects of the two modes on the material surface: the pits created by impact provide more cutting points for abrasion, while the rough surface formed by abrasion exacerbates the concentrated stress of impact. Without step S33, it would be impossible to distinguish the different erosion mechanisms, leading to distorted calculations of the combined erosion intensity. For example, a corrosion-dominated region might be misclassified as abrasion, causing subsequent optimization strategies to use incorrect wear-resistant materials, ultimately affecting the erosion resistance. Simultaneously, the lack of a three-dimensional distribution field would blur the location of erosion hotspots, failing to provide precise targeting for topology optimization in step S40.

[0110] Step S40: Based on the three-dimensional scour intensity distribution field, identify layered scour hotspots and formulate differentiated geometric adjustment strategies, perform topology optimization iterative feedback, and generate a scour-resistant optimized flow channel geometric model.

[0111] Further, step S40 includes:

[0112] Step S41: Based on the three-dimensional scour intensity distribution field, identify and mark severe, moderate and mild scour hotspots, extract the feature information of scour hotspots, and establish a graded scour hotspot database;

[0113] Step S42: Based on the hotspot feature information in the hierarchical scour hotspot database, formulate differentiated geometric adjustment strategies for different scour modes;

[0114] Step S43: Apply the differentiated geometry adjustment strategy to generate the modified flow channel geometry model, re-execute S10 to S30 to obtain a new three-dimensional scour intensity distribution field, calculate the optimization effect index and perform convergence judgment, and output the final optimized flow channel geometry model.

[0115] Step S40, based on the three-dimensional scour intensity distribution field, identifies stratified scour hotspots and formulates targeted geometric adjustment strategies to achieve precise optimization of the flow channel topology. This solves the problem that traditional optimization methods cannot make fine adjustments for localized stratified scour. Specifically, S41 to S43 work together to construct a complete closed loop from scour analysis to geometric optimization. S41 addresses the lack of quantitative standards and a grading system in traditional hotspot identification. Based on the three-dimensional scour intensity distribution field generated in S33, it identifies and marks severe, moderate, and mild scour hotspots by setting multi-level thresholds, extracting hotspot feature information to establish a grading hotspot database. In the three-dimensional scour intensity distribution field, the comprehensive scour intensity of each spatial point has been quantified. Based on this, three threshold levels are set: severe scour hotspots correspond to areas with comprehensive scour intensity significantly higher than the flow channel average intensity; moderate scour hotspots are those with a moderately higher intensity than the average; and mild scour hotspots are those with a slightly higher intensity than the average. These thresholds are determined through statistical analysis of multiple failure cases. For example, scour failure analysis of slurry valves after service shows that areas with significantly higher intensity than the average will inevitably leak within a short service life; therefore, this is used as the severe threshold. During the identification process, it is necessary to extract the feature information of each hotspot, including its spatial coordinates in the curve coordinate system, scour pattern, maximum scour intensity value, and the particle size group and computational layer involved. A graded scour hotspot database stores this information according to severity, with each record associated with corresponding geometric feature parameters, providing multi-dimensional basis for subsequent strategy formulation. This step, in conjunction with S33, upgrades scour hotspot identification from a vague qualitative description to a quantitative spatial location. Traditional methods can only indicate that "the outer side of the bend is prone to scour," while S41 can pinpoint the specific axial position and circumferential angle, and distinguish the scour differences between different computational layers at the same location. This refined identification lays the foundation for targeted adjustments. Without S41, subsequent optimization will lack a clear target, leading to blind adjustments, such as treating mild hotspots and severe hotspots the same, resulting in over-optimization or under-optimization.

[0116] Based on the characteristic information of scour hotspots in a graded scour hotspot database, S42 formulates differentiated geometric adjustment strategies for different scour modes, solving the problem of poor local effects caused by the "one-size-fits-all" approach of traditional optimization. For hotspots dominated by impact mode, whose formation is related to high-kinetic-energy, high-frequency impacts, the adjustment strategy focuses on reducing the kinetic energy and frequency of particle impacts, specifically by increasing the local radius of curvature or increasing the flow guide radius. For hotspots dominated by abrasion mode, due to the significant cutting effect of medium-kinetic-energy particles, the strategy focuses on suppressing the relative velocity between particles and the wall surface, achieved by adjusting the contraction ratio or expansion angle to optimize the velocity gradient. For hotspots dominated by corrosion mode, due to the prominent chemical effects of the medium, the strategy combines material adjustment and geometric optimization, such as increasing the surface roughness of the flow channel in this region to promote the formation of a passivation film, while simultaneously reducing the medium residence time by fine-tuning the expansion angle. For hotspots with mixed modes, such as regions where impact and abrasion coexist, a composite strategy is adopted. For example, for a hotspot in the middle layer of a certain expansion section, both the radius of curvature is increased to suppress impact and the velocity gradient is optimized to suppress abrasion. Certain geometric adjustments can produce cross-mode suppression effects. For example, optimizing the radius of curvature of a bend not only reduces the impact intensity but also disperses the particle concentration in the abrasion zone due to a more uniform flow velocity distribution. This synergistic effect cannot be predicted by traditional methods.

[0117] S43 verifies the optimization effect through an iterative feedback mechanism, addressing potential issues like local overtuning or the generation of new hotspots in a single optimization. Based on the differentiated geometry adjustment strategy of S42, a modified 3D CAD model is generated, requiring re-execution of S10 to S30: In S10, changes in the flow channel boundaries of the new geometry model affect the particle enrichment index distribution; for example, increased curvature reduces the enrichment index of large-diameter particles on the outer side. In S20, new geometric feature parameters update the input of the corrected model, making the particle concentration distribution throughout the flow channel more closely match the optimized flow field. S30 calculates the 3D scour intensity based on the new particle distribution and compares it with the pre-optimization result. The optimization effect metric is the maximum intensity reduction of severe hotspots. The convergence criterion is that the absolute value of the difference between the maximum intensity reduction of severe hotspots in two consecutive iterations is less than a preset reduction threshold, and no new hotspots are generated. The threshold for reducing the erosion rate needs to be determined through multiple sets of experiments, taking into account the stable range of the material's erosion resistance and measurement accuracy. This involves selecting specimens of the same material as the flow channel (such as stainless steel or wear-resistant cast iron) and conducting erosion experiments under simulated conditions. The maximum strength reduction of severe hotspots at different iteration stages is measured. It is found that when the absolute value of the difference between two reduction values ​​is less than a certain threshold, further iterations reduce the specimen's service life beyond the range of natural aging. For example, experiments on stainless steel specimens show that when this difference is less than 5%, further iterations only extend the expected service life of the specimen by less than 1%, while the optimization cost exceeds the benefit. Therefore, 5% is used as an example for the reduction threshold. The upper limit of the number of iterations is determined through multiple sets of experiments; exceeding this number significantly reduces marginal benefits. Without S43, the chain reaction of geometric adjustments may not be considered, potentially leading to new hotspots. For example, simply increasing the curvature of a bend may increase the particle concentration at the bottom of the downstream straight pipe section, triggering new impact hotspots.

[0118] Step S40, through the synergy of hierarchical identification, differentiated adjustment, and iterative verification, achieves precise optimization of the flow channel topology. Geometric adjustments not only reduce scouring intensity but also, due to a more uniform flow field, disperse particle distribution, reducing local energy concentration. Step S40, together with S10 to S30, forms a closed loop, creating a complete mapping between particle stratification characteristics, geometric constraints, scouring mechanisms, and optimization strategies. Without S40, the analytical data from the preceding steps cannot be translated into actual geometric improvements, the entire methodological chain breaks, and the flow channel's scouring resistance cannot be substantially improved. Through this step, flow channel design is upgraded from experience-driven to data-driven, extending service life and reducing maintenance costs; its value far exceeds simple scouring suppression.

[0119] Example 2:

[0120] This embodiment, based on Embodiment 1, provides a CFD-based slurry valve anti-erosion flow channel topology optimization design device, such as... Figure 4 As shown, it includes:

[0121] Enrichment map construction module: used to acquire slurry particle size distribution data, establish a particle size grouping identification system based on the slurry particle size distribution data; and construct a three-dimensional particle enrichment index distribution map based on the slurry particle size distribution data and the particle size grouping identification system.

[0122] The full-channel concentration state calculation module extracts the geometric feature parameter sequence of the channel based on the 3D CAD model of the slurry valve; and generates the full-channel particle concentration distribution state sequence based on the geometric feature parameter sequence of the channel and the 3D particle enrichment index distribution map.

[0123] Scour intensity distribution field construction module: Based on the particle concentration distribution state sequence of the entire flow channel, a layered scour energy density distribution is generated; according to the layered scour energy density distribution, three scour modes, namely impact, abrasion and corrosion, are identified, the comprehensive scour intensity is calculated and a three-dimensional scour intensity distribution field is generated.

[0124] Topology optimization module: Based on the three-dimensional scour intensity distribution field, it identifies layered scour hotspots and formulates differentiated geometric adjustment strategies, performs topology optimization iterative feedback, and generates a scour-resistant optimized flow channel geometric model.

[0125] Furthermore, in the enrichment map construction module, the method for constructing a three-dimensional particle enrichment index distribution map includes: Let k be the index variable of the computational layer, for the k-th computational layer L... k and the i-th particle size group G i Obtain the computation layer L k Inner particle size group G i Local concentration C ik and the particle size group G throughout the flow channel i average concentration C i,avg Through C ik Divide by C i,avg Obtain the i-th particle size group G i In the k-th computational layer L k The particle enrichment index.

[0126] Furthermore, in the full-channel concentration state calculation module, the method for extracting the sequence of geometric feature parameters of the channel based on the 3D CAD model of the slurry valve includes:

[0127] Establish a curved coordinate system along the centerline of the flow channel; from the inlet to the outlet of the slurry valve, set a geometric feature extraction section at preset distances, for a total of P geometric feature extraction sections;

[0128] Based on the 3D CAD model of the slurry valve, the geometric feature parameters of each geometric feature section are extracted. The geometric feature parameters include the centroid coordinates of the section, the area of ​​the section, and the equivalent diameter of the section.

[0129] Geometric feature parameters of the cross section are extracted based on each geometric feature, and geometric change parameters between adjacent cross sections are calculated, including radius of curvature, shrinkage ratio, and expansion angle; the geometric feature parameters and geometric change parameters constitute a geometric feature parameter sequence;

[0130] Furthermore, in the full-channel concentration state calculation module, the method for generating the full-channel particle concentration distribution state sequence includes:

[0131] Step S231: Divide the entire flow channel into Q geometric segments, and construct the particle distribution state vector of the upstream geometric segment based on the particle enrichment index of each particle size group in each calculation layer in each geometric segment.

[0132] Step S232: Construct a distribution state transfer matrix between geometric segments based on the geometry type-based correction model to describe the evolution of particle distribution from the upstream geometric segment to the downstream geometric segment;

[0133] Step S233: Calculate the particle distribution state vector of the downstream geometric segment through matrix operation between the distribution state transfer matrix and the particle distribution state vector of the upstream geometric segment.

[0134] Step S234: The particle distribution state vector of the downstream geometric segment is used as the particle distribution state vector of the new upstream geometric segment. The particle distribution state vector and the distribution state transfer matrix of the upstream geometric segment are applied sequentially along the flow channel direction to complete the full flow channel particle distribution state transfer calculation from the inlet to the outlet, which includes all geometric segments.

[0135] Step S235: Integrate the full-channel particle distribution state transfer calculation results of all geometric segments to generate a full-channel particle concentration distribution state sequence containing the concentration distribution of each particle size group at each location.

[0136] The methods and apparatus of this application may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the method is for illustrative purposes only, and the steps of the method of this application are not limited to the order specifically described above, unless otherwise specifically stated.

[0137] In addition, the parts of the technical solutions provided in the embodiments of this application that are consistent with the implementation principles of the corresponding technical solutions in the prior art have not been described in detail, so as to avoid excessive elaboration.

[0138] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A CFD-based slurry valve anti-erosion flow passage topology optimization design method, characterized in that, The method comprises: obtaining slurry particle size distribution data, dividing the particles into N particle size groups based on the particle size distribution data using the logarithmic interval principle, and assigning a unique digital identification code to each particle size group; calculating the theoretical settling velocity of each particle size group, establishing a flow rate-lamination criticality determination model, and generating a lamination degree quantization index for each particle size group; discretizing the vertical direction of the flow channel into M calculation layers according to the lamination degree quantization index, calculating the particle enrichment index of each particle size group in each calculation layer, and constructing a three-dimensional particle enrichment index distribution map; the horizontal axis of the three-dimensional particle enrichment index distribution map is the particle size group number, the vertical axis is the calculation layer number, and the numerical value is the particle enrichment index; based on the three-dimensional CAD model of the slurry valve, extracting the geometric feature parameter sequence of the flow channel; based on the geometric feature parameter sequence of the flow channel and the three-dimensional particle enrichment index distribution map, generating a full-flow channel particle concentration distribution state sequence; based on the full-flow channel particle concentration distribution state sequence, generating a lamination scouring energy density distribution; according to the lamination scouring energy density distribution, identifying three scouring modes of impact, abrasion and corrosion, calculating the comprehensive scouring intensity and generating a three-dimensional scouring intensity distribution field; based on the three-dimensional scouring intensity distribution field, identifying the lamination scouring hotspots and formulating a differentiated geometric adjustment strategy, performing topological optimization iteration feedback, and generating an anti-scouring optimized flow channel geometric model.

2. The CFD-based slurry valve anti-washdown flow passage topology optimization design method of claim 1, wherein, The flow rate-lamination criticality determination model is: Obtaining the local flow velocity v at the current position in the flow channel flow , defining the critical ratio R of the i-th particle size group i =v flow / v settle,i , wherein i is the index of the particle size group, v settle,i represents the theoretical settling velocity of the i-th particle size group; When R i <R critical When R is significantly stratified, it is considered a significant stratification. i >R mix When R is fully mixed, it is considered fully mixed; when R is fully mixed, it is considered fully mixed. critical ≤R i ≤R mix At that time, the quantification index D was used to determine the degree of stratification. i Describe the transition state, D i D represents the quantitative index of the stratification degree of the i-th particle size group. i The value range is from 0 to 1, where 0 corresponds to complete mixing and 1 corresponds to complete stratification. R critical R is the critical threshold for stratification. mix This is the mixed critical threshold.

3. The CFD-based slurry valve anti-washdown flow passage topology optimization design method of claim 2, wherein, The method for calculating the particle enrichment index of each particle size group in each calculation layer is: Let k be the index variable of the calculation layer, for the kth calculation layer L k and the ith particle size group G i , the local concentration C k of the particle size group G i in the calculation layer L ik , and the average concentration C i of the particle size group G i,avg in the entire flow channel, the particle enrichment index of the ith particle size group G i in the kth calculation layer L k is obtained by dividing C ik by C i,avg .

4. The CFD-based slurry valve anti-washdown flow passage topology optimization design method of claim 3, wherein, The method for extracting the geometric feature parameter sequence of the flow channel based on the three-dimensional CAD model of the slurry valve comprises: establishing a curve coordinate system along the center line of the flow channel; from the inlet to the outlet of the slurry valve, a geometric feature extraction section is set every predetermined distance, and a total of P geometric feature extraction sections are set; based on the three-dimensional CAD model of the slurry valve, extracting the geometric feature parameters of each geometric feature extraction section, the geometric feature parameters including the section centroid coordinates, the section area and the section equivalent diameter; according to the geometric feature parameters of each geometric feature extraction section, calculating the geometric change parameters between adjacent sections, the geometric change parameters including the curvature radius, the contraction ratio and the expansion angle; the geometric feature parameters and the geometric change parameters constitute the geometric feature parameter sequence.

5. The CFD-based slurry valve anti-washdown flow passage topology optimization design method of claim 4, wherein, The construction method of the full-flow channel particle concentration distribution state sequence comprises: according to the geometric feature parameter sequence, marking the geometric types of the geometric feature extraction sections, the geometric types including straight pipe sections, curved sections, contraction sections and expansion sections; based on the geometric feature parameter sequence, the marked geometric types and the three-dimensional particle enrichment index distribution map, constructing a correction model based on the geometric types; according to the correction model based on the geometric types, establishing a multi-section flow channel particle distribution state transfer algorithm to generate a full-flow channel particle concentration distribution state sequence containing the concentration distribution of each particle size group at each position.

6. The CFD-based slurry valve anti-washdown flow passage topology optimization design method of claim 5, wherein, The method for establishing the multi-section flow channel particle distribution state transfer algorithm comprises: dividing the entire flow channel into Q geometric sections, and constructing a particle distribution state vector of an upstream geometric section according to the particle enrichment index of each particle size group in each calculation layer in each geometric section; According to the correction model based on the geometric type, a distribution state transfer matrix for describing the evolution rule of the particle distribution from an upstream geometric section to a downstream geometric section is constructed between geometric sections; A particle distribution state vector of the downstream geometric section is calculated through matrix operation of the distribution state transfer matrix and the particle distribution state vector of the upstream geometric section; The particle distribution state vector of the downstream geometric section is taken as a particle distribution state vector of a new upstream geometric section, and the particle distribution state vector and the distribution state transfer matrix of the upstream geometric section are sequentially applied along the flow passage direction to complete the particle distribution state transfer calculation of the whole flow passage containing all geometric sections from the inlet to the outlet.

7. A CFD-based slurry valve anti-erosion flow passage topology optimization design apparatus for implementing the CFD-based slurry valve anti-erosion flow passage topology optimization design method of any one of claims 1-6, characterized in that, The device comprises: An enrichment map construction module is configured to obtain slurry particle size distribution data, establish a particle size grouping identification system according to the slurry particle size distribution data, and construct a three-dimensional particle enrichment index distribution map based on the slurry particle size distribution data and the particle size grouping identification system; A full-flow passage concentration state calculation module is configured to extract geometric characteristic parameter sequences of a flow passage based on a three-dimensional CAD model of a slurry valve, and generate a full-flow passage particle concentration distribution state sequence based on the geometric characteristic parameter sequences of the flow passage and the three-dimensional particle enrichment index distribution map; A scouring intensity distribution field construction module is configured to generate a layered scouring energy density distribution based on the full-flow passage particle concentration distribution state sequence, identify three scouring modes of impact, abrasion and corrosion according to the layered scouring energy density distribution, calculate a comprehensive scouring intensity, and generate a three-dimensional scouring intensity distribution field; A topological optimization module is configured to identify layered scouring hotspots based on the three-dimensional scouring intensity distribution field, formulate a differentiated geometric adjustment strategy, perform topological optimization iteration feedback, and generate an anti-scouring optimized flow passage geometric model.

Images