Traditional Chinese medicine crushing equipment mechanical design structure stress optimization analysis system

Abstract

This invention relates to the field of mechanical stress analysis and optimization technology, and discloses a structural stress optimization analysis system based on the mechanical design of a traditional Chinese medicine pulverizing equipment. The system includes: defining key load-bearing components of the entire machine as unified design variables for equipment comprising an upper pulverizing module, a side transmission module, and a lower sieving module; extracting macroscopic data through mechanical testing and converting it into discrete element contact parameters for fiber bundles to generate a load model; extracting contact torque and energy data based on the load model within the geometry corresponding to the design variables to construct an eccentricity index; mapping multi-source mechanical loads to a finite element mesh model, using the eccentricity index to perform time-domain weighting of the response to obtain the eccentric response; constructing a multidimensional normalized response based on the eccentric response, generating weights using the information entropy of a contemporary particle swarm optimization, and constructing a fitness function in conjunction with constraints; finally, using the fitness function to drive particle position updates to perform parameter optimization, and outputting the minimum solution as the final parameters.

Description

Technical Field

[0001] This invention relates to the field of mechanical stress analysis and optimization technology, and more specifically, to a structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment. Background Technology

[0002] In the processing system of traditional Chinese medicine, due to space constraints and process integration limitations, small processing units that integrate three-roller pulverization, belt and gear drives, and eccentric connecting rod screening are commonly used. However, during actual production testing, it was found that when processing low-moisture, high-fiber rhizomes, even with a herringbone feed ridge at the inlet, long fiber bundles are difficult to distribute evenly into the left and right roller gaps as expected, unlike brittle particles. These herbs are prone to bridging or entanglement within the narrow gap, leading to temporary blockage and retention on one side of the roller gap, followed by sudden breakage due to enormous transient shear force.

[0003] This asymmetric, alternating fracture release transforms the radial reactions, which should theoretically cancel each other out in a three-roller arrangement, into intense alternating bending and torsional pulsating loads. When these pulsating loads are transmitted to the bearing housing and the top beam of the frame via the drive roller and gears, and then superimposed on the inertial force of the reciprocating screening cycle at the bottom, the conventional load-bearing margin verification scheme based on vertical static pressure will fail. In practice, this manifests as a sharp increase in stress concentration in localized areas, slight tilting of the bearing housing, and eccentric meshing wear on the tooth sides.

[0004] To solve this problem, simply thickening the frame steel plate or blindly increasing the shaft diameter often leads to exceeding the curb weight limit and is prone to entering the dynamic resonance amplification zone. In engineering, there is an urgent need for a stress analysis scheme that can truly restore the material off-center load evolution process and guide the optimization of the whole machine parameters. Summary of the Invention

[0005] This invention provides a structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment, which solves the technical problems mentioned in the background art.

[0006] This invention provides a structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment. It is applied to equipment containing an upper three-roller pulverizing module, a side-mounted transmission module, and a lower-mounted sieving module, and is configured to execute: Extract the given initial baseline variables, write the key components as unified design variables, and establish a three-dimensional component stress model to define the physical boundaries of frame stiffness and roll gap space; Mechanical test data of rhizomes and other Chinese medicinal materials were converted into discrete element contact parameters of fiber bundles to construct a load model characterizing the anisotropic shear resistance of the materials. In the geometry corresponding to the unified design variables, contact torque and energy data are extracted based on the load model to construct an eccentricity index and quantify the load offset amplitude triggered by material retention. Multi-source mechanical loads are mapped to a finite element mesh model, and the response is weighted in the time domain using the eccentricity index to obtain an eccentric response that filters out steady-state load characteristics. Based on the eccentric response, a multidimensional normalized response is constructed. The information entropy of the contemporary particle swarm is used to generate response weights to correct dimensional bias. A fitness function is constructed in combination with the constraint conditions. The unified design variable is set as the particle position, and the fitness function is used to drive the particle position update execution parameter optimization, which converges to the target force structure that eliminates resonance interference. Extract the minimum fitness solution from the optimal solutions of each generation of population as the final parameter, and generate an optimization result set containing the final parameter.

[0007] The beneficial effects of this invention are as follows: Addressing the eccentric load pulse accumulation effect that easily occurs in the double-roller pulverizing zone of Chinese herbal root and rhizome materials, this invention integrates the anisotropic shear resistance of the medicinal materials with the asymmetric feeding geometric boundary on both sides, and introduces eccentric weighted response and contemporary particle swarm information entropy for multi-parameter dynamic mapping. The optimized results effectively weaken the alternating bending and torsional pulsations caused by long fiber retention and sudden breakage, alleviate the slight tilting of the active roller bearing seat and the eccentric load of the tooth side meshing, avoid the overall machine resonance frequency band, and significantly improve the equipment's eccentric load fatigue life and dynamic operational stability while taking into account the overall quality and manufacturing assembly tolerance. Attached Figure Description

[0008] Figure 1 This is a flowchart of the structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to the present invention. Detailed Implementation

[0009] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0010] like Figure 1 As shown, a structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment is applied to equipment containing an upper three-roller pulverizing module, a side-mounted transmission, and a lower-mounted sieving module, configured to execute: Extract the given initial baseline variables, write the key components as unified design variables, and establish a three-dimensional component stress model to define the physical boundaries of frame stiffness and roll gap space; Mechanical test data of rhizomes and other Chinese medicinal materials were converted into discrete element contact parameters of fiber bundles to construct a load model characterizing the anisotropic shear resistance of the materials. In the geometry corresponding to the unified design variables, contact torque and energy data are extracted based on the load model to construct an eccentricity index and quantify the load offset amplitude triggered by material retention. Multi-source mechanical loads are mapped to a finite element mesh model, and the response is weighted in the time domain using the eccentricity index to obtain an eccentric response that filters out steady-state load characteristics. Based on the eccentric response, a multidimensional normalized response is constructed. The information entropy of the contemporary particle swarm is used to generate response weights to correct dimensional bias. A fitness function is constructed in combination with the constraint conditions. The unified design variable is set as the particle position, and the fitness function is used to drive the particle position update execution parameter optimization, which converges to the target force structure that eliminates resonance interference. Extract the minimum fitness solution from the optimal solutions of each generation of population as the final parameter, and generate an optimization result set containing the final parameter.

[0011] The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment provided in this embodiment is applied to equipment that includes an upper three-roller pulverizing module, a side-mounted transmission module, and a lower-mounted screening module. By integrating material mechanical properties, structural geometric parameters, and dynamic response, the system achieves multi-objective stress optimization of key load-bearing components of the equipment, effectively alleviating the asymmetric off-center load impact problem during the pulverization process of root and rhizome traditional Chinese medicines.

[0012] S201: The active roller shaft diameter, bearing span, bearing housing thickness, bearing housing height, top beam thickness, top beam height, reinforcing rib thickness, reinforcing rib height, reinforcing rib inclination angle, material distribution ridge offset, material distribution ridge inclination angle, and roller surface pitch ratio are combined to form a unified design variable.

[0013] The unified design variables are represented in column vector form, covering all key geometric parameters that affect the stress performance of the equipment. The calculation formula is as follows: ; in, The diameter of the drive roller shaft ranges from 50mm to 150mm. The bearing span ranges from 200mm to 500mm. The thickness of the bearing housing is 20mm to 60mm. The bearing housing height ranges from 100mm to 300mm. The thickness of the top beam ranges from 15mm to 50mm. The height of the top beam ranges from 80mm to 250mm. The thickness of the reinforcing ribs is to be between 10mm and 30mm. The height of the reinforcing ribs is ranged from 50mm to 150mm. The angle of inclination of the stiffener is taken from 30 degrees to 90 degrees; For the material distribution ridge offset, the value range is -20mm to 20mm; The angle of the material distribution ridge ranges from 30 degrees to 60 degrees. The roller pitch ratio ranges from 0.8 to 1.2; superscript This is the matrix transpose operator.

[0014] Roller pitch ratio This is the ratio of the pitch of the active roller teeth to the pitch of the passive roller teeth, used to characterize the influence of roller meshing characteristics on material biting ability. At that time, the pitch of the two roller teeth is perfectly matched; when When the pitch of the active roller teeth is greater than that of the passive roller, the material shearing effect is enhanced; when When the passive roller pitch is greater than that of the active roller, the material conveying capacity is improved. The value range of 0.8 to 1.2 is an engineering experience value based on the ratio of fiber length to roller diameter of root and rhizome Chinese medicinal materials, and can be adjusted within this range according to the average fiber length of the material.

[0015] S202: Based on the center distance of the three rollers in the upper three-roller crushing module, position compensation is performed by combining the spatial vector deviation introduced by the offset of the material distribution ridge and the inclination angle of the material distribution ridge, and left roll-in distance and right roll-in distance representing the volume of the left and right material distribution channels are constructed respectively.

[0016] The material distribution ridge adopts an isosceles triangular cross section, and the foundation height is [missing information]. The width of the base is Offset of material distribution ridge This represents the horizontal offset of the apex of the material distribution ridge relative to the midpoint of the line connecting the centers of the three rollers; a leftward offset is positive, and a rightward offset is negative. Material distribution ridge inclination angle. This refers to the angle between the side of the material distribution ridge and the vertical direction. Spatial vector deviation. The calculation formula is determined by the offset and inclination angle of the material distribution ridge: ; The left and right roll-in distances reflect the effective distance at which material enters the corresponding roll gap, directly affecting the material holding capacity of the distribution channel. The calculation formula is as follows: ; ; in, Left roll-in distance, The right roll-in distance, This refers to the center distance between the three rollers, with a value ranging from 150mm to 300mm. The parameters are defined as functions with uniform design variables as independent variables.

[0017] S203: Perform spatial operations on the projected areas of the left and right distribution channels along the feeding direction to extract the left and right surface integrals that characterize the flow cross-section features.

[0018] The projection along the feeding direction is an orthographic projection, and the projection plane is perpendicular to the feeding direction. Left feed channel. The boundary is formed by the left side of the material distribution ridge, the left roller surface generatrix, the inner surface of the left side plate of the frame, and the lower surface of the top beam; the right material distribution channel The boundary is formed by the right side of the distribution ridge, the generatrix of the right roller surface, the inner surface of the right side plate of the frame, and the lower surface of the top beam. The flow cross-sectional characteristics directly determine the flow characteristics of the material through the distribution channel, which is the basis for quantifying the asymmetry of the flow distribution. The calculation formula is: ; ; in, For the surface integral on the left, For the projection integral domain of the left material distribution channel, For the right surface integral, This is the projection integral domain of the right material distribution channel. It is an area differential element.

[0019] S204: Based on the proportion of the left surface integral in the sum of the left and right surface integrals, a left feed coefficient is constructed to quantify the asymmetric characteristics of particle diversion.

[0020] The left feed coefficient ranges from 0 to 1. A value closer to 0.5 indicates more uniform material distribution to the left and right, while a value deviating more from 0.5 indicates stronger asymmetry in material distribution. The calculation formula is as follows: ; in, This is the left feed coefficient.

[0021] S301: Extract the relationship between cross-sectional resistance and deformation scale during mechanical deformation process, generate elastic modulus, shear modulus, fracture stress and shear strength to characterize the deformation resistance properties of material fiber bundles, and extract the spatial dimensions of discrete contact bridges.

[0022] Uniaxial tensile and shear tests were conducted on samples of rhizomes and other medicinal materials to obtain load-deformation curves, and then various mechanical parameters were calculated using the following formulas: ; ; ; in, For elastic modulus, For axial load increment, This represents the axial deformation increment. The initial length of the sample is 50 mm. The cross-sectional area of ​​the sample is 100 mm². For fracture stress, For tensile fracture load; Shear strength, For tangential shear load; The shear modulus is calculated from the elastic modulus and Poisson's ratio using the relations of mechanics of materials. The calculation formula is: ; in, The Poisson's ratio of the material is 0.2 to 0.4.

[0023] Discrete contact bridges refer to the connecting segments between adjacent fiber bundles. The spatial dimensions of a contact bridge include its area. and contact bridge length The microstructure of the fiber bundle was statistically obtained through scanning electron microscopy observation. The specific steps were as follows: 20 sample slices with a thickness of 50 μm were prepared. Five observation areas of 100 μm × 100 μm were selected from each slice. The area and length of all contact bridges in each area were measured, and the arithmetic mean was taken as the final contact bridge area and contact bridge length. The contact bridge area ranged from 10 μm² to 50 μm², and the contact bridge length ranged from 5 μm to 20 μm.

[0024] For rhizome-type medicinal materials, samples were taken from the middle of the main root, with the epidermis and xylem removed, and prepared into cylindrical specimens with a diameter of 10 mm and a length of 50 mm. The specimens were equilibrated for 24 hours at 25℃ and 60% relative humidity to stabilize the moisture content between 8% and 12%. Uniaxial tensile testing was performed using a universal testing machine with a loading rate of 1 mm / min; shear testing was performed using a shear clamp with a loading rate of 0.5 mm / min. At least 20 valid specimens were tested for each medicinal material, and the arithmetic mean was taken after removing outliers with a dispersion greater than 20%.

[0025] S302: Integrating elastic modulus, shear modulus, and discrete contact bridge spatial dimensions to construct normal and tangential stiffness to constrain relative particle displacement.

[0026] Normal stiffness and tangential stiffness describe the elastic contact characteristics between discrete particles in the normal and tangential directions, respectively. They are core parameters in discrete element simulation, and their calculation formulas are as follows: ; ; in, For normal stiffness, This refers to the tangential stiffness.

[0027] S303: Combining fracture stress, shear strength, and discrete contact bridge spatial dimensions to construct positive and tangential fracture forces, which are used to calibrate the critical mechanical threshold for fiber bundle fracture.

[0028] When the contact force between particles exceeds the corresponding breaking force, the fiber bundle breaks, and the contact relationship is released. The calculation formula is: ; ; in, For positive breaking force, It is a tangential breaking force.

[0029] S304: Based on the normal stiffness, tangential stiffness and material friction coefficient, the relative normal compression and tangential slip between discrete particles are converted into torque and quantified by frictional retardation to generate the instantaneous contact force that combines elasticity and frictional dissipation, thus forming a load model.

[0030] Instantaneous contact force comprises two parts: elastic restoring force and frictional force. It fully describes the interaction between particles, and the calculation formula is as follows: ; in, For instantaneous contact force, This is the relative normal compression. It is the normal unit vector, determined by the direction of the line connecting the centers of the two particles; This is the tangential slip. It is the tangential unit vector, perpendicular to the normal unit vector; The coefficient of dynamic friction of the material was measured under normal temperature and pressure, normal pressure of 1 MPa, and sliding speed of 0.5 m / s, with a value ranging from 0.3 to 0.5. This is the scalar value of the normal contact force. The slip direction is a unit vector, determined by the direction of the tangential component of the relative velocity between particles, and is calculated using the following formula: ; in, For the tangential component of the relative velocity between particles, when hour, Take the zero vector.

[0031] S401: Extract the left contact torque and right contact torque within the left and right roll gaps, and construct a deviation ratio based on the relative amplitude difference between the two to eliminate the interference of the foundation load sharing, thus characterizing the transient off-center load of the roll system.

[0032] Left contact torque Contact torque on the right To calculate the magnitude of the vector sum of the contact forces of all particles within the roll gap about the center of the roll shaft, considering only the radial moment and neglecting the axial moment, the calculation formula is as follows: ; ; in, This is the set of all contact points within the left roll gap. This is the collection of all contact points within the right roll gap. This is the position vector of the contact point relative to the center of the roller shaft.

[0033] The deviation ratio, after normalization to eliminate the influence of the total load magnitude, only reflects the relative difference in load between the left and right roll gaps. The calculation formula is as follows: ; in, The deviation ratio To prevent the torque from being a minimum constant with a denominator of zero, a value of [value to be specified] is taken. .

[0034] S402: Extracts the elastic energy stored due to particle deformation and the frictional energy generated by sliding work, and isolates it from the system kinetic energy introduced by the material mass and roller inertia, thus separating the hysteresis dissipation source inside the material.

[0035] Bonding points refer to the unbroken contact points between fiber bundles; the set of transiently activated bonding points. The activation condition is that the contact force is greater than zero and the fiber is not broken. (Set of transient contact points) This includes all particle-particle and particle-roller surface contact points within the left and right roll gaps, but excludes particle-frame contact points. Elastic energy storage is the energy stored by fiber bundle deformation, frictional energy is the energy consumed by particle-particle sliding, and the system kinetic energy is the sum of the kinetic energy of the roller system rotation and particle translation. The calculation formula is: ; ; ; in, For flexible energy storage, For the keypoint index variable, and For the first Normal and tangential compression at each bond point; Frictional energy For contact point index variables, For the first scalar of normal contact force at each contact point The relative slip velocity vector, The discrete time increment takes the value of ; As the system's kinetic energy, The equivalent rotational inertia of the roller system, with a value range of [value missing]. to , The active roller angular velocity ranges from 10 rad / s to 50 rad / s. For the aggregate of particles in the crushing zone, For granular index variables, For particle mass, the value range is: to , This is the vector of the particle's translational velocity.

[0036] Equivalent moment of inertia of roller system The sum of the moments of inertia of the driving roller, the driven roller, and the connecting shaft is calculated using the following formula: ; in, The moment of inertia of the driving roller. , The moment of inertia of the two passive rollers. The moment of inertia is the rotational inertia of the connecting shaft. The moments of inertia of each component are calculated using 3D modeling software or verified by actual measurement using the weighing method and the oscillation method.

[0037] S403: The accumulated amount of elastic stored energy and frictional energy along the working cycle of the active roller is collected in the time domain to obtain the hysteresis release energy, and proportionally mapped with the time domain integral of the system kinetic energy in the same period to construct a hysteresis multiplier characterizing the sudden fracture release effect after short-term energy storage of the fiber bundle.

[0038] The hysteresis multiplier reflects the relative intensity of energy accumulation and release within a material. A larger value indicates a higher proportion of energy released during sudden fracture, and a more significant impact effect. The calculation formula is: ; in, For lagging multipliers, The working cycle of the drive roller is calculated from the angular velocity of the drive roller. , It is a dummy variable during time integration. Let be the energy minimum constant, and take the value of . .

[0039] S404: Generate an eccentricity index based on the energy-weighted amplification of the deviation ratio of the lag multiplier pair in the execution time dimension.

[0040] The eccentricity index comprehensively considers the degree of load offset and the intensity of energy release, and can accurately quantify the load offset amplitude triggered by material retention. The calculation formula is as follows: ; in, It is an eccentric indicator.

[0041] S501: The load mapping matrix is ​​used to align the spatial coordinates of the multi-source mechanical loads, and the contact force between the left and right roller gaps, the gear meshing force, the belt drive pressure force and the screen body inertial force are reconstructed into nodal loads applied to the grid nodes, thus eliminating excitation spatial misalignment.

[0042] The load mapping matrix is ​​a rigid body transformation matrix, consisting of a rotation matrix and a translation vector, which realizes the transformation from local coordinate system load to global finite element mesh coordinate system. The coordinate mapping matrix of the roller system converts the resultant contact force of the left and right roller gaps in the local coordinate system into a nodal load in the global coordinate system, which is applied to the center node of the active roller journal. This is the gear coordinate mapping matrix, which converts the gear meshing force in the local coordinate system into a nodal load in the global coordinate system and applies it to the connection node between the gear pitch circle and the journal. The pulley coordinate mapping matrix converts the belt drive axial force in the local coordinate system of the pulley into a nodal load in the global coordinate system, which is then applied to the center node of the pulley. The screen body coordinate mapping matrix converts the inertial forces of the screen body in the local coordinate system into nodal loads in the global coordinate system, which are then applied to the connection nodes between the screen body and the frame. The nodal load calculation formula is as follows: ; in, Let be the column vector of nodal loads. , , , , These are the corresponding local coordinate system physical load components.

[0043] Left roller gap contact force Contact force with the right roll gap The formula for calculating the vector sum of the contact forces between all particles and the roll surface within the roll gap is as follows: ; ; in, This is the contact force vector between a single particle and the roller surface, containing only normal and tangential components, with the axial component being negligible.

[0044] S502: Based on the system physical topology, construct dynamic equations including component mass scalars, damping matrices, and stiffness matrices, and substitute nodal loads to obtain the time-domain nodal displacements and corresponding equivalent stresses of the structure.

[0045] The dynamic equations were established using the finite element method, and the damping matrix was obtained using the Rayleigh damping model, derived from a linear combination of the mass matrix and stiffness matrix. The calculation formula is as follows: ; in, This is the mass damping coefficient. The stiffness damping coefficient is calculated using the first two natural frequencies and the damping ratio. The calculation formula is as follows: ; ; in, , These are the first and second order natural frequencies of the rack, respectively. , These are the corresponding damping ratios, with values ​​ranging from 0.02 to 0.05.

[0046] The dynamic equations and equivalent stress calculation formulas are as follows: ; ; in, A mass matrix containing scalar attributes of component mass. Here is the damping matrix. Here is the stiffness matrix. , , These are column vectors representing the nodal displacements and their first-order velocity and second-order acceleration responses with respect to time, respectively. Spatial coordinate points Equivalent stress at the point, , , These are the corresponding principal stress components.

[0047] mass matrix and stiffness matrix The structure was assembled using finite element methods. First, the entire structure was discretized into hexahedral solid elements, each with 8 nodes and 3 translational degrees of freedom. The element mass matrix and element stiffness matrix were calculated based on the element material properties and geometric dimensions. Then, all element matrices were assembled into a global mass matrix and a global stiffness matrix according to the node numbers. The material density was taken as 7850 kg / m³, the elastic modulus as 206 GPa, and the Poisson's ratio as 0.3.

[0048] S503: Using the eccentricity index as a time-domain weighting factor, the equivalent stress in the dangerous area within the overall system response calculation window is mapped to the evolutionary characteristics, filtering out steady-state stresses and extracting the eccentric stress response dominated by sudden pulsations.

[0049] The hazardous area was determined through initial finite element analysis. The boundary conditions for the initial finite element analysis were a fully fixed constraint at the bottom of the frame, and the load conditions were steady-state loads under the rated working load. The area with equivalent stress greater than 70% of the material's allowable stress was extracted as the coordinate set of the hazardous area. The eccentric stress response retains only the stress pulsation component caused by eccentric impact, and the calculation formula is as follows: ; in, For eccentric stress response, The calculation window variable is used to calculate the overall system response, which includes multiple work cycles, and its value is... , Let be the index minimum constant, and let its value be . .

[0050] S504: Simultaneously utilize the eccentricity index to perform weighted extraction calculations on the nodal displacements of displacement-sensitive parts to isolate steady-state displacements and construct an eccentric displacement response, which, together with the eccentric stress response, constitutes an eccentric response characterizing transient eccentric load impact.

[0051] The displacement-sensitive areas were determined through initial finite element analysis, and the regions with steady-state displacements greater than 0.1 mm were extracted as the coordinate set of the displacement-sensitive areas. The eccentric displacement response retains only the displacement pulsation component caused by eccentric impact, and the calculation formula is as follows: ; in, This is the response to eccentric displacement.

[0052] S601: Fatigue damage is extracted by utilizing the stress range within the alternating force evolution cycle, and modal proximity is extracted based on the coupling interference characteristics between the applied load spectrum and the frame's natural frequency, in order to eliminate potential dynamic resonance hazards.

[0053] Fatigue damage was calculated using a modified Miner linear cumulative damage criterion, with an allowable number of cycles. The calculation formula is obtained from the SN curve of the material: ; in, and Let be the material fatigue constant, for Q235 steel, , ; For the first The stress range of the first-order fatigue load is extremely poor.

[0054] The formula for calculating fatigue damage is: ; in, For fatigue damage, For fatigue load order index, This represents the total order, with a value of 10. This represents the actual number of loops. This is the load correction factor, with a value ranging from 0.8 to 1.2.

[0055] Modal proximity reflects the degree of closeness between the applied load frequency and the rack's natural frequency; a higher value indicates a higher risk of resonance. Load harmonic order. Take 1 to 10, frame vibration mode order Take values ​​from 1 to 6. Load spectrum amplitude The Fourier transform amplitude of the nodal loads and the harmonic amplitude vector. This represents the load vector of the corresponding order. The modal proximity calculation formula is: ; in, For modal proximity, For rack number First mode vector, For the first First natural frequency, For the first External load harmonic frequency, To prevent odd small values, the value is set to... .

[0056] Load correction factor This is used to consider the effects of load amplitude, loading frequency, and stress concentration on fatigue life. When the stress difference is less than 50% of the material's yield strength, Take 1.0; when the stress range is between 50% and 80% of the yield strength, Take 1.1; when the stress range is greater than 80% of the yield strength, Take 1.2. For high-frequency loads (frequency greater than 100Hz), multiply by an additional frequency correction factor of 1.1.

[0057] S602: Maps the eccentric stress response, eccentric displacement response, fatigue damage, component mass scalar, and modal proximity to the comparative dimensions under a given initial reference variable, eliminating the numerical islanding effect across physical field parameters and combining them to form a multidimensional normalized response.

[0058] The multidimensional normalized response converts parameters with different physical dimensions into dimensionless relative values, facilitating subsequent multi-objective optimization calculations. The calculation formula is as follows: ; in, For a multidimensional normalized response column vector, Given an initial baseline variable, This is a scalar measure of component quality.

[0059] S603: Extract the discrete distribution state of the multidimensional normalized response in the contemporary particle swarm, calculate the information entropy of each response component constituting the multidimensional normalized response to quantify the structural distinguishability, and assign response weights to enhance the correction effect to parameters with significant differences.

[0060] A smaller information entropy indicates a more concentrated distribution of the response component in the particle swarm, higher discriminative power, and therefore a larger weight should be assigned. First, the multidimensional normalized response is subjected to range standardization, calculated using the following formula: ; in, For the first The first particle The range standardized variable for each item, For the corresponding unscaled original element, This is a scaling minimum constant, with a value of [value]. .

[0061] Then calculate the individual probability distribution value, using the following formula: ; in, These are the individual probability distribution values. The total population size is 40.

[0062] Next, the information entropy is calculated using the following formula: ; in, For the first Information entropy of a physical response.

[0063] Finally, the response weight is calculated using the following formula: ; in, In response weights, This represents the total number of target response items, with a value of 5.

[0064] When the probability distribution value of a single item At that time, due to Since it has no mathematical meaning, the minimum value substitution method is used to process it. Replace with The calculation formula is: ; Then use Perform information entropy calculations to ensure the stability of numerical calculations.

[0065] S604: Combines the response weights with the various response components that complete the extreme value scaling transformation, superimposes them to penalize assembly errors that exceed the physical interference limits, and outputs the fitness function.

[0066] The criterion for judging physical interference is that the minimum distance between two components is less than 0.1 mm. The assembly error penalty term uses a quadratic penalty function: when interference occurs, the penalty term is 1000 multiplied by the square of the interference volume; otherwise, it is 0. The calculation formula is: ; in, This is a penalty item for assembly errors. This represents the interference volume between components.

[0067] A smaller fitness function value indicates a better design. The calculation formula is: ; in, In order to target the Particles The fitness function scalar.

[0068] S701: Capture the extreme values ​​of fitness improvement during the iteration process, and extract the individual improvement amount of the individual historical optimization contribution and the group improvement amount of the group collaborative optimization contribution, respectively.

[0069] The individual improvement is the absolute value of the difference between the current particle's fitness and its own historical best fitness; if the difference is negative, it is set to 0. The swarm improvement is the absolute value of the difference between the current swarm's best fitness and the previous generation's swarm's best fitness; if the difference is negative, it is set to 0. The calculation formula is as follows: ; ; in, For the first Individual improvement amount in each iteration For the first The amount of group improvement in the next iteration For the first The individual's optimal historical position in the next iteration. For the first The optimal global position of the population in the next iteration.

[0070] Individual optimal historical position The update rule is: if the first In the next iteration, particles fitness Less than its historical best fitness Then update Otherwise, keep .

[0071] Group optimal global position The update rule is: iterate through the individual optimal positions of all particles, and find the position with the minimum fitness as the optimal position. It is updated once per generation.

[0072] S702: Based on the competitive comparison between individual improvement amount and group improvement amount, adaptively allocate individual coefficients to guide local development and group coefficients to guide global exploration.

[0073] When the individual improvement is large, the individual coefficient is increased to enhance the local search; when the swarm improvement is large, the swarm coefficient is increased to enhance the global search. The calculation formula is as follows: ; ; in, For the first Individual coefficients in the next iteration. For the population coefficient, To prevent the denominator from being zero, the learning rate is set to a value of [value to be filled in]. .

[0074] S703: By evaluating the discrete convergence of the contemporary particle swarm fitness distribution, the velocity coefficient used to balance the parameter search span and the position locking tendency is dynamically calculated.

[0075] When the particle swarm fitness distribution has a large dispersion, the velocity coefficient is increased to expand the search range; when the dispersion is small, the velocity coefficient is decreased to accelerate convergence. The calculation formula is as follows: ; ; ; in, The mean fitness of the population. The standard deviation of population fitness For speed coefficient, Let be the constant with the minimum dispersion, and take the value of . .

[0076] S704: Merge the velocity coefficient, individual coefficient and group coefficient to correct the current particle velocity to obtain the updated velocity, and superimpose the updated velocity at the current particle position to execute the design domain constraint to prevent manufacturing from going out of bounds, and decode to generate the updated unified design variable.

[0077] The particle velocity update formula is: ; in, For faster update speed, The current particle velocity, with an initial velocity range of 10% of the corresponding design variable range. The current particle position. For the individual's optimal historical position, To find the optimal global position for the group. , is a random number that is uniformly distributed in the interval [0,1].

[0078] The particle position update formula is: ; in, The position of the transition particle after superposition.

[0079] Design domain constraint projection function Projecting particle positions outside the allowable range onto the boundary, the upper and lower limits of each design variable are: , , , , , , , , , , , The projection rule is to take the lower limit value when the particle position is less than the lower limit, and the upper limit value when it is greater than the upper limit. The updated unified design variable calculation formula is as follows: ; in, For the updated unified design variables.

[0080] The initial positions of the particles are generated randomly and uniformly within the design domain, and the initial value of each design variable is calculated using the following formula: ; in, For the first The first particle Initial values ​​for each design variable, and The first The lower and upper limits of a design variable. is a random number that is uniformly distributed in the interval [0,1].

[0081] S801: Scan the global iteration space and lock the sample space variables that trigger the fitness function to reach the lower limit penalty threshold as the final parameters.

[0082] Maximum number of iterations in Particle Swarm Optimization The value is set to 150, and the convergence condition is that the rate of change of the optimal fitness of the population is less than 150 for 20 consecutive generations. The final parameter is the solution with the smallest fitness function value among the optimal solutions of each generation of the population, calculated using the following formula: ; in, For the final parameters, For the end generation The optimal position of the group at each round.

[0083] After each iteration, the system stores the optimal position of the current generation. and their corresponding fitness values This forms the set of optimal solutions for each generation. During storage, duplicate solutions are automatically removed, retaining only independent solutions with different fitness values, thus reducing the computational load of subsequent scans.

[0084] S802: By comparing the difference in response indices between the given initial baseline variables and the final parameters, the response improvement rate, which characterizes the performance ratio of the multidimensional physical field, is obtained analytically.

[0085] The response improvement rate is the relative improvement of each response metric. A higher value indicates a more significant optimization effect. The calculation formula is as follows: ; in, In response to the improvement rate, , These are the corresponding individual response baseline value and optimal final value, respectively.

[0086] S803: Extract the steady-state time mean of the eccentricity index corresponding to the given initial benchmark variable and the final parameter, and analyze the eccentricity attenuation rate, which characterizes the ability of the double roll gap to eliminate asymmetric hysteresis fracture impact.

[0087] The eccentric attenuation rate reflects the degree of reduction in eccentric load impact after optimization. The larger the value, the better the eccentric load impact elimination effect. The calculation formula is: ; ; in, The steady-state time mean of the eccentricity index. This represents the eccentric attenuation rate.

[0088] The steady-state time mean of the eccentricity index is adopted The integral window is determined based on the periodic characteristics of the pulverization process of rhizome-type Chinese medicinal materials. Ten working cycles are sufficient to cover the complete cycle of material retention, bridging, and breakage, accurately reflecting the steady-state statistical characteristics of the eccentricity index. If the fiber length of the material varies greatly, the integral window can be extended to 20 working cycles.

[0089] S804: Encapsulates the final parameters along with their physical associated nodal loads, equivalent stresses, nodal displacements, response improvement rates, and eccentricity attenuation rates to generate an optimized result set for distribution to manufacturing and assembly.

[0090] Discrete element simulations use fibrous particles with aspect ratios ranging from 5 to 10, and time step values ​​are... The contact model adopted is an adhesive contact model. Finite element analysis uses hexahedral solid elements with mesh sizes ranging from 5mm to 10mm, and the boundary condition is a fully fixed constraint at the bottom of the frame. Data encapsulation uses a structured data format, containing all geometric parameters and mechanical performance verification data required for manufacturing. The calculation formulas are as follows: ; in, To optimize the encapsulation format of the physical quantity data mapping set in the result set.

[0091] The optimization result set adopts a hierarchical structured data format, which includes the following four levels: 1. Geometric parameter layer: Stores final parameters The 12 design variable values ​​and their corresponding tolerance requirements; 2. Payload Data Layer: Stores node payloads Time series data, with a sampling interval of ; 3. Response Data Layer: Storing Equivalent Force and nodal displacement The maximum value and its corresponding coordinates; 4. Performance Metrics Layer: Storage 5 Response Improvement Rates and eccentric decay rate The value.

[0092] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment, applied to equipment containing an upper three-roller pulverizing module, a side-mounted transmission module, and a lower-mounted sieving module, characterized in that... Configured for execution: Extract the given initial baseline variables, write the key components as unified design variables, and establish a three-dimensional component stress model to define the physical boundaries of frame stiffness and roll gap space; Mechanical test data of rhizomes and other Chinese medicinal materials were converted into discrete element contact parameters of fiber bundles to construct a load model characterizing the anisotropic shear resistance of the materials. In the geometry corresponding to the unified design variables, contact torque and energy data are extracted based on the load model to construct an eccentricity index and quantify the load offset amplitude triggered by material retention. Multi-source mechanical loads are mapped to a finite element mesh model, and the response is weighted in the time domain using the eccentricity index to obtain an eccentric response that filters out steady-state load characteristics. Based on the eccentric response, a multidimensional normalized response is constructed. The information entropy of the contemporary particle swarm is used to generate response weights to correct dimensional bias. A fitness function is constructed in combination with the constraint conditions. The unified design variable is set as the particle position, and the fitness function is used to drive the particle position update execution parameter optimization, which converges to the target force structure that eliminates resonance interference. Extract the minimum fitness solution from the optimal solutions of each generation of population as the final parameter, and generate an optimization result set containing the final parameter.

2. The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to claim 1, characterized in that, Key components are written as unified design variables, and a three-dimensional component stress model is established, including: The unified design variables are formed by combining the active roller shaft diameter, bearing span, bearing housing thickness, bearing housing height, top beam thickness, top beam height, reinforcing rib thickness, reinforcing rib height, reinforcing rib inclination angle, material distribution ridge offset, material distribution ridge inclination angle, and roller surface pitch ratio. Based on the center distance of the three rollers in the upper three-roller crushing module, and combined with the spatial vector deviation introduced by the material distribution ridge offset and the material distribution ridge inclination angle, position compensation is performed to construct the left roll-in distance and right roll-in distance, which represent the volume of the left and right material distribution channels, respectively. Spatial operations are performed on the projected areas of the left and right distribution channels along the feeding direction to extract the left and right surface integrals that characterize the flow cross-section features; Based on the proportion of the left surface area in the sum of the left and right surface areas, a left feed coefficient is constructed to quantify the asymmetric characteristics of particle diversion.

3. The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to claim 2, characterized in that, Construct a load model characterizing the anisotropic shear properties of the material, including: The relationship between cross-sectional resistance and deformation scale during mechanical deformation is extracted to generate elastic modulus, shear modulus, fracture stress and shear strength, which characterize the deformation resistance of material fiber bundles, and the spatial dimensions of discrete contact bridges are extracted. By integrating the elastic modulus, the shear modulus, and the spatial dimensions of the discrete contact bridge, normal stiffness and tangential stiffness are constructed to constrain the relative displacement of particles. The fracture stress, shear strength and discrete contact bridge spatial dimensions are combined to construct the positive fracture force and tangential fracture force, which are used to calibrate the critical mechanical threshold for fiber bundle fracture. Based on the normal stiffness, the tangential stiffness, and the material friction coefficient, the relative normal compression and tangential slip between discrete particles are converted into torque and quantified by frictional retardation to generate an instantaneous contact force that combines elasticity and frictional dissipation, thus forming the load model.

4. The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to claim 3, characterized in that, In the geometry corresponding to the unified design variables, contact moment and energy data are extracted based on the load model to construct an eccentricity index, including: Extract the left contact torque and right contact torque within the left and right roll gaps, and construct a deviation ratio based on the relative amplitude difference between the two to eliminate the interference of the foundation load sharing, thus characterizing the transient off-center load of the roll system. Extract the elastic energy stored due to particle deformation and the frictional energy caused by sliding work, and isolate them from the system kinetic energy introduced by the material mass and roller inertia, thus separating the hysteresis dissipation source inside the material. The accumulated amount of elastic energy and frictional energy along the working cycle of the active roller is collected in the time domain to obtain the hysteresis release energy, and is proportionally mapped with the time domain integral of the system kinetic energy in the same period to construct a hysteresis multiplier characterizing the sudden breakage and release effect of the fiber bundle after short-term energy storage. The eccentricity index is generated by the energy-weighted amplification of the deviation ratio along the execution time dimension based on the lag multiplier.

5. The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to claim 4, characterized in that, Multi-source mechanical loads are mapped to a finite element mesh model, and the response is time-domain weighted using the eccentricity index to obtain the eccentric response after filtering out steady-state load characteristics, including: The multi-source mechanical loads are aligned in spatial coordinate system using a load mapping matrix, and the contact force between the left and right roller gaps, the gear meshing force, the belt drive shaft pressure force, and the screen body inertial force are reconstructed into nodal loads applied to the grid nodes, thus eliminating excitation spatial misalignment. Based on the system's physical topology, a dynamic equation is constructed that includes component mass scalars, damping matrices, and stiffness matrices. Substituting these equations into the nodal loads, the time-domain nodal displacements and corresponding equivalent stresses of the structure are obtained. Using the eccentricity index as a time-domain weighting factor, the equivalent stress in the dangerous area within the overall system response calculation window is mapped to the evolutionary characteristics, filtering out steady-state stresses and extracting the eccentric stress response dominated by sudden pulsations. Simultaneously, the eccentricity index is used to perform a weighted extraction operation on the nodal displacement of the displacement-sensitive part to isolate the steady-state displacement and construct the eccentric displacement response, which together with the eccentric stress response constitutes the eccentric response characterizing the transient eccentric load impact.

6. The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to claim 5, characterized in that, The information entropy of contemporary particle swarm optimization is used to generate response weights to correct dimensional bias. A fitness function is constructed by combining this with constraints, including: Fatigue damage is extracted by utilizing the stress range within the alternating stress evolution cycle, and modal proximity is extracted based on the coupling interference characteristics between the applied load spectrum and the frame's natural frequency, in order to eliminate potential dynamic resonance hazards. The eccentric stress response, the eccentric displacement response, the fatigue damage, the component mass scalar, and the modal proximity are mapped to the comparative dimensions under the given initial reference variable, eliminating the numerical islanding effect across physical field parameters, and are combined to form the multidimensional normalized response; Extract the discrete distribution state of the multidimensional normalized response in the contemporary particle swarm, calculate the information entropy of each response component constituting the multidimensional normalized response to quantify the structural distinguishability, and assign the response weights that enhance the correction effect to parameters with significant differences. The fitness function is output by combining the response weights with the response components that have undergone extreme value scaling transformation, superimposing assembly errors used to penalize physical interference exceeding the limits.

7. The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to claim 6, characterized in that, The fitness function is used to drive particle position updates to perform parameter optimization, converging to the target force structure that eliminates resonance interference, including: Capture the extreme values ​​of fitness improvement during the iteration process, and extract the individual improvement amount of the individual historical optimization contribution and the group improvement amount of the group collaborative optimization contribution, respectively. Based on the competitive comparison between the individual improvement amount and the group improvement amount, the individual coefficients used to guide local development and the group coefficients used to guide global exploration are adaptively allocated. By evaluating the discrete convergence of the contemporary particle swarm fitness distribution, the velocity coefficient used to balance the parameter search span and the position locking tendency is dynamically calculated. The current particle velocity is corrected by fusing the velocity coefficient, the individual coefficient, and the group coefficient to obtain an updated velocity. The updated velocity is then superimposed on the current particle position to execute design domain constraints that prevent manufacturing from going out of bounds. The updated unified design variable is then decoded and generated.

8. The structural stress optimization analysis system based on the mechanical design of traditional Chinese medicine pulverizing equipment according to claim 7, characterized in that, Extract the fitness-minimum solution from the previous population optimal solutions as the final parameter, and generate an optimization result set containing the final parameter, including: Scan the global iteration space and lock the sample space variables that trigger the fitness function to reach the lower limit penalty threshold as the final parameters; By comparing the difference in response indices corresponding to the given initial benchmark variable and the final parameter, the response improvement rate, which characterizes the performance ratio of the multidimensional physical field, is analytically obtained. Extract the steady-state time mean of the eccentric index corresponding to the given initial benchmark variable and the final parameter, and analyze the eccentric attenuation rate that characterizes the ability to eliminate asymmetric hysteresis fracture impact in the twin roll gap. The final parameters, along with their physically associated nodal loads, equivalent stresses, nodal displacements, response improvement rates, and eccentricity attenuation rates, are encapsulated to generate the optimized result set for distribution to manufacturing and assembly.

Images