Unknown boundary pipeline model optimization method and device, equipment and storage medium

Abstract

The invention provides an unknown boundary pipeline model optimization method and device, equipment and a storage medium, and relates to the technical field of pipeline structure model optimization, and the method comprises the steps: constructing an elastic boundary unit to simulate the boundary rigidity of a local pipeline model of an unknown boundary; one spring unit simulates translational stiffness or torsional stiffness in one direction; determining the spring stiffness of the plurality of first spring units as an initial parameter, and generating a plurality of first parameter combinations according to the value range of the initial parameter; and calculating dynamic weighted dominating scores among the first parameter combinations, and performing iterative optimization on the first parameter combinations based on the dynamic weighted dominating scores, so as to determine a target parameter combination of the pipeline model under the condition that the iterative optimization is completed. According to the method, the advantages and disadvantages of the to-be-optimized parameter combination can be evaluated by calculating the dynamic weighted dominating score of each parameter combination, so that the parameter combination is iteratively optimized, and the accuracy of the pipeline model is improved.

Description

Technical Field

[0001] The present application relates to the technical field of pipeline structure model optimization, and in particular to an unknown boundary pipeline model optimization method, device, equipment and storage medium. Background Art

[0002] In recent years, the petrochemical industry has become an important pillar of the national economy. As a key facility for the transportation of media and raw materials, the importance of petrochemical pipelines is self-evident. However, petrochemical pipelines are often subject to the combined effects of multiple internal and external excitation factors, such as high-temperature and high-speed gas pulsation, airflow impacting the pipe wall, and changes in the natural frequency of the gas column caused by the switching of the reactor valve. External excitation may also be transmitted through the temperature field, magnetic field, external flow field or pipeline support structure to induce local vibration of the pipeline. Long-term high-frequency local vibration will not only cause safety hazards, affect the stability and reliability of the system, but may even cause significant economic losses.

[0003] To ensure the safety and reliability of pipelines, accurate and reliable finite element models are key tools for pipeline design, vibration safety assessment, and the formulation of vibration reduction measures. However, the petrochemical pipeline system is complex, and the local vibration pipeline model often faces problems such as incomplete global structural information and difficulty in obtaining stiffness and mass matrices during the optimization process, which makes many traditional model optimization methods that rely on global structural information inapplicable. Therefore, there is currently a problem of low accuracy of pipeline models. Summary of the invention

[0004] The present application provides an unknown boundary pipeline model optimization method, device, equipment and storage medium, which can evaluate the advantages and disadvantages of the parameter combination to be optimized by calculating the power-weighted dominance score of each parameter combination, thereby iteratively optimizing the parameter combination to improve the accuracy of the pipeline model.

[0005] In a first aspect, an embodiment of the present application provides a method for optimizing an unknown boundary pipeline model, which may include: Constructing an elastic boundary unit to simulate the boundary stiffness of a local pipeline model of an unknown boundary; wherein the elastic boundary unit includes spring units in multiple directions; one of the spring units simulates the translational stiffness or torsional stiffness in one direction; Determine the spring stiffness of a plurality of first spring units as an initial parameter, and generate a plurality of first parameter combinations according to the value range of the initial parameter; wherein the first spring unit is any of the spring units; The power weighted dominance scores between the first parameter combinations are calculated, and the first parameter combinations are iteratively optimized based on the power weighted dominance scores, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed.

[0006] In the above implementation, an elastic boundary unit is proposed to simulate the unknown boundary of the pipeline structure. Therefore, it can be applicable to local pipeline structures with unclear global structural information and unclear boundary conditions. It can optimize the local vibration pipeline model under the conditions of uncertain boundary conditions, complex objective functions to be optimized, and incomplete dynamic test data. The advantages and disadvantages of the parameter combination to be optimized are evaluated by calculating the dynamic weighted dominance score of each parameter combination, so as to iteratively optimize the parameter combination to improve the accuracy of the pipeline model.

[0007] In some embodiments, the calculating of the power weighted dominance scores between the first parameter combinations and iteratively optimizing the first parameter combinations based on the power weighted dominance scores to determine the target parameter combination of the pipeline model when the iterative optimization is completed may include: Determine the frequency correlation between the local pipeline and the global pipeline represented by the first parameter combination based on the characteristic frequency residual; Determining the correlation of vibration mode components between the local pipeline and the global pipeline represented by the first parameter combination within a specified degree of freedom range based on a global-local modal assurance criterion; A dynamically weighted dominance score between the first parameter combinations is determined based on the frequency correlation and the mode shape component correlation.

[0008] In the above implementation, by calculating the characteristic frequency residual, the difference between the theoretical calculation frequency and the actual test frequency can be reflected. By using the global-local modal assurance criterion, the vibration mode component correlation between the local pipeline and the global pipeline within the specified degree of freedom can be accurately determined, thereby reflecting the consistency or difference of the vibration form of the pipeline system at different degrees of freedom, and improving the accuracy of evaluating the change of structural dynamic characteristics. By combining the frequency correlation and the vibration mode component correlation, the dynamic weighted dominance score between each first parameter combination is calculated, thereby reflecting the degree of influence of different parameter combinations on the dynamic characteristics of the pipeline system, providing a basis for the iterative optimization operation of the first parameter combination.

[0009] In some embodiments, the calculating of the power weighted dominance scores between the first parameter combinations and iteratively optimizing the first parameter combinations based on the power weighted dominance scores to determine the target parameter combination of the pipeline model when the iterative optimization is completed may also include: sorting all of the first parameter combinations according to the power-weighted dominance scores; According to the sorting result, a crossover operation and a mutation operation are performed on the plurality of target parameter combinations ranked first to generate a next-generation parameter combination; Merging the second generation parameter combination with the first parameter combination, and repeatedly performing the steps of sorting the merged parameter combination and generating the second generation parameter combination, so as to iterate the first parameter combination; When the iteration is completed, the top parameter combination is output as the target parameter combination.

[0010] In the above implementation, all first parameter combinations are sorted by the dynamic weighted dominance score, and the parameter combination with better performance can be accurately identified. This sorting method is based on the dynamic weighted dominance score that comprehensively evaluates the frequency correlation and the vibration type component correlation, so it has higher accuracy and reliability. Crossover and mutation operations are performed on multiple target parameter combinations with higher rankings to generate next-generation parameter combinations with higher potential. The crossover operation can combine the advantages of different parameter combinations, while the mutation operation can introduce new variation factors, thereby increasing the diversity of parameter combinations. In the case of incomplete dynamic test data, by making full use of known modal information, the maximum optimization of the local pipeline model is achieved, which can compensate for the impact of insufficient data to a certain extent. The iterative optimization method is also used in the embodiment of the present application, which can gradually stabilize the performance of the parameter combination, reduce the volatility and uncertainty of the parameter combination, and thus improve the reliability of pipeline model optimization.

[0011] In some embodiments, determining the dynamic weighted dominance score between the first parameter combinations according to the frequency correlation and the vibration mode component correlation may include: Determine a first less-than count based on the frequency correlation, and determine a second less-than count based on the mode shape component correlation; determining a frequency total weighted dominance score based on the first less-than count and the weighted average position; Determining a total weighted dominance score of mode shape components according to the second less-than count and the weighted average position; The dynamic weighted dominance score is determined according to the total frequency weighted dominance score and the total mode shape component weighted dominance score.

[0012] In some embodiments, the power-weighted dominance score is characterized by: Where S is the power weighted control score, S f is the total weighted dominance score of the frequency, S glmac is the total weighted dominance score of the vibration mode component, and nk is the number of all parameter combinations.

[0013] In some embodiments, determining the spring stiffness of the plurality of first spring units as initial parameters and generating a plurality of first parameter combinations according to the value range of the initial parameters may include: Determine the natural frequency and vibration mode of each of the first spring units respectively; The spring stiffness of the plurality of first spring units is determined as an initial parameter according to the natural frequency and the vibration mode.

[0014] In the above implementation, by accurately calculating the natural frequency of each first spring unit, its vibration characteristics under specific conditions can be understood, providing basic data for subsequent optimization design. Determining the vibration mode of each spring unit can more accurately characterize the deformation of the spring unit during vibration. By analyzing the natural frequency and vibration mode of the first spring unit, the spring stiffness parameters that significantly affect the dynamic characteristics can be identified. These parameters will be used as initial parameters for iterative optimization to generate the target parameter combination of the pipeline model, which can improve the accuracy of the pipeline model.

[0015] In some embodiments, the ranking all the first parameter combinations according to the power-weighted dominance scores may include: Calculating the power weighted dominance scores between any two of the first parameter combinations, and sorting the dominance relationships to obtain a first sorting result; The sorting results of all the first parameter combinations are determined according to all the first sorting results.

[0016] In the above implementation, the power weighted dominance score can comprehensively consider the less than count and weighted average position of the power correlation coefficient, so as to more comprehensively reflect the pros and cons of each parameter combination.

[0017] In a second aspect, an embodiment of the present application provides an unknown boundary pipeline model optimization device, the device comprising: A construction module is used to construct an elastic boundary unit to simulate the boundary stiffness of a local pipeline model of an unknown boundary; wherein the elastic boundary unit includes spring units in multiple directions; one of the spring units simulates the translational stiffness or torsional stiffness in one direction; A determination module, used to determine the spring stiffness of a plurality of first spring units as an initial parameter, and generate a plurality of first parameter combinations according to a value range of the initial parameter; wherein the first spring unit is any of the spring units; The calculation module is used to calculate the power weighted dominance score between each of the first parameter combinations, and iteratively optimize the first parameter combinations based on the power weighted dominance score, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed.

[0018] In a third aspect, an embodiment of the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the above description when executing the computer program.

[0019] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the muscle function assessment method as described in the above description is implemented.

[0020] Compared with the prior art, the beneficial effects of the present application are: by proposing an elastic boundary unit to simulate the unknown boundary of the pipeline structure, it can be applicable to local pipeline structures with unclear global structural information and unclear boundary conditions, and can optimize the local vibration pipeline model under the conditions of uncertain boundary conditions, complex objective functions to be optimized, and incomplete dynamic test data. By calculating the dynamic weighted dominance score of each parameter combination to evaluate the advantages and disadvantages of the parameter combination to be optimized, the parameter combination can be iteratively optimized to improve the accuracy of the pipeline model. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of the steps of the unknown boundary pipeline model optimization method provided in an embodiment of the present application.

[0022] Figure 2 Schematic diagram of sequential and reverse elastic boundary units provided in embodiments of the present application.

[0023] Figure 3 A schematic diagram of a pipeline overview provided in an embodiment of the present application.

[0024] Figure 4 A schematic diagram of the optimization results of the optimized model provided in the embodiment of the present application.

[0025] Figure 5 A schematic diagram of an unknown boundary pipeline model optimization device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The present application is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as the scope of the above subject matter of the present application being limited to the following embodiments, and all technologies implemented based on the content of the present application belong to the scope of protection of the present application.

[0027] Unless otherwise specified, in the description of the specific embodiments of the present application, the terms indicating the orientation or position relationship such as "up", "down", "left", "right", "center", "inside", "outside", "side", etc. are all expressions based on the orientation or position relationship shown in the drawings, or the orientation or position relationship when the product / equipment / device is usually used. These terms of orientation or position relationship are only for the convenience of describing the scheme of the present application or simplifying the description in the specific embodiments to facilitate the technicians to quickly understand the scheme, rather than indicating or implying that a specific device / component / element must have a specific orientation, or be constructed and operated in a specific position relationship, and therefore cannot be understood as a limitation on the present application.

[0028] In the description of the embodiments of the present application, the technical terms "first", "second", etc. only distinguish one entity or operation from another entity or operation, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0029] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0030] Example 1 During the research, the applicant found that in the optimization of dynamic models, it is usually necessary to optimize multi-order modal information, including frequency and vibration mode, which will lead to too many optimization objective functions. Traditional multi-objective optimization algorithms (such as genetic algorithms) may fall into the dilemma of dominance sorting, resulting in deviations in optimization results. In addition, existing finite element model optimization methods usually require obtaining complete modal information in advance, but in actual applications, due to noise interference and process problems, complete pipeline modal information is often difficult to obtain.

[0031] Therefore, providing a local vibration pipeline model optimization method suitable for uncertain boundary conditions, complex objective functions to be optimized, and incomplete modal test data has important engineering application value. The present invention proposes an innovative solution, which can effectively cope with the above challenges through an intelligent optimization method based on the dynamic weighted dominance score, improve the optimization accuracy and reliability of the local pipeline model, and provide strong support for the safe design and vibration control of petrochemical pipelines.

[0032] Based on this, the present application embodiment provides an unknown boundary pipeline model optimization method, please refer to Figure 1 , Figure 1 A schematic diagram of the steps of the unknown boundary pipeline model optimization method provided in an embodiment of the present application. The unknown boundary pipeline model optimization method may include: S1. Construct elastic boundary elements to simulate the boundary stiffness of the local pipeline model with unknown boundaries.

[0033] Among them, the elastic boundary unit includes spring units in multiple directions; one spring unit simulates the translational stiffness or torsional stiffness in one direction; S2. Determine the spring stiffness of a plurality of first spring units as initial parameters, and generate a plurality of first parameter combinations according to a value range of the initial parameters.

[0034] The first spring unit is any spring unit. The first parameter combination is the parameter combination to be optimized.

[0035] S3. Calculate the power weighted dominance scores between the first parameter combinations, and iteratively optimize the first parameter combinations based on the power weighted dominance scores, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed.

[0036] The unknown boundary pipeline model optimization method provided in the embodiment of the present application can be applied to situations where the boundary conditions of the pipeline model are uncertain, the objective function to be optimized is relatively complex, and the dynamic test data is incomplete.

[0037] The local pipeline model can be a finite element model, and the spring unit can be 6 Combin14 spring units. The Combin14 spring unit is a spring-damper unit in Ansys, which can be used in one-dimensional, two-dimensional or three-dimensional applications and has longitudinal or torsion functions. Three Combin14 spring units simulate the translational stiffness in the x, y, and z directions, and three Combin14 spring units simulate the torsion stiffness around the x, y, and z axes, ignoring the damping effect. Each spring unit connects two nodes, each node has one degree of freedom, and is connected to the boundary nodes of the local pipeline and the surrounding pipeline respectively.

[0038] The stiffness matrix of the 6 Combin14 spring elements is as follows: Since the structural parameters of the global pipeline are unknown and the vibration only occurs in the local pipeline, the surrounding pipelines are simplified and only one pipe unit is retained.

[0039] One node of the tube element is connected to the boundary spring element, and the other node is assumed to be consolidated. The tube element stiffness matrix can be expressed as: in, , represents the stiffness matrix of the 6-DOF pipe element with unilateral consolidation.

[0040] Please see Figure 2 , Figure 2 A schematic diagram of sequential and reverse elastic boundary units provided in an embodiment of the present application, wherein a boundary consisting of spring units in six directions and one unilateral consolidation tube unit is defined as an elastic boundary unit, which can be divided into sequential elastic boundaries and reverse elastic boundaries according to the node coupling order.

[0041] Sequential elastic boundaries and reverse elastic boundaries have different element stiffness matrices, which are: and the inverse elastic boundary element stiffness matrix : Where i represents the spring direction (i=x, y, z, rotx, roty, rotz), is the transformation matrix from the local coordinate system to the global coordinate system of the sequential boundary spring element, is the transformation matrix from the local coordinate system of the reverse boundary spring element to the global coordinate system. and The expression is: In the above implementation process, an elastic boundary unit is proposed to simulate the unknown boundary of the pipeline structure. Therefore, it can be applicable to local pipeline structures with unclear global structural information and unclear boundary conditions. It can optimize the local vibration pipeline model when the boundary conditions are uncertain, the objective function to be optimized is complex, and the dynamic test data is incomplete. The dynamic weighted dominance score of each parameter combination is calculated to evaluate the advantages and disadvantages of the parameter combination to be optimized, so as to iteratively optimize the parameter combination to improve the accuracy of the pipeline model.

[0042] Example 2 This embodiment is a specific step of determining the target parameter combination of the pipeline model in the above embodiment 1.

[0043] Calculating the power weighted dominance scores between the first parameter combinations, and iteratively optimizing the first parameter combinations based on the power weighted dominance scores, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed, may include: S31. Determine the frequency correlation between the local pipeline and the global pipeline represented by the first parameter combination based on the characteristic frequency residual.

[0044] Among them, the characteristic frequency residual is used to measure the frequency correlation between the local pipeline and the global pipeline represented by the first parameter combination, and its expression is: Among them, r f,k is the frequency correlation coefficient, which ranges from 0 to 1. The closer the value is to 0, the better the frequency correlation is.

[0045] λ glo,k =2π f glo,k , λ loc,k =2π f loc,k , f is the natural frequency of the structure (Hz), glo represents the global pipeline structure, loc represents the local pipeline structure, and k is the modal order.

[0046] S32. Determine the correlation between the vibration mode components of the local pipeline and the global pipeline represented by the first parameter combination within a specified degree of freedom range based on the global-local modal assurance criterion.

[0047] It should be noted that the traditional MAC value ignores the vibration coordinate information, while the global pipeline vibration vector φ glo,k Length greater than the local pipe mode vector φ loc,k Therefore, it is necessary to further extract the vibration mode information related to the corresponding coordinates of the local pipeline in the global pipeline vibration mode.

[0048] To this end, a global-local modal assurance criterion (GLMAC) is proposed in the embodiment of the present application to measure the correlation between the vibration mode components of the global structure and the local structure of the pipeline within a specified degree of freedom, and its expression is: in, It represents the correlation coefficient of the vibration mode component, s is the specified degree of freedom range, and represents the vibration mode vector corresponding to the degree of freedom range of the local structure node in the extracted global structure. The value of is between 0 and 1. The closer the value is to 0, the better the correlation of the vibration modes within the relevant degree of freedom.

[0049] S33. Determine the dynamic weighted dominance scores between the first parameter combinations according to the frequency correlation and the vibration mode component correlation.

[0050] Among them, S33 may specifically include: S331 , determining a first less-than count according to frequency correlation, and determining a second less-than count according to mode shape component correlation.

[0051] Less-Than Count (LTC) is a statistic used to describe the size relationship between two numeric vectors. For two numeric vectors of equal length a=[a1,a2,a3,…,an] and b=[b1,b2,b3,…,bn], LTC means that after each element in vector a is compared with the corresponding element in vector b, it is less than the number of corresponding elements in b. Its expression is: in, is an indicator function that is 1 if the condition is met and 0 otherwise.

[0052] Assuming the cutoff frequency is nk, the frequency correlation coefficients of the i-th parameter combination and the i+1-th parameter combination are and : The correlation coefficients of the vibration mode components are and : By traversing the frequency correlation coefficient according to the expression of less than count, we can get the less than count vector of frequency correlation containing position information. and : By traversing the modal component correlation coefficients according to the expression of less than count, the less than count vector of the modal component correlation coefficients containing position information can be obtained. and : in, k (k∈Z, 1≤ k ≤ n k ) is the modal order, The vector k Component representation and exist k The comparison results of the first mode position. , and Similarly, the less-than count vector can be summed to get the less-than count, expressed as: Among them, LTC f,i, LTC f,i+1 Respectively represent i and i +1 parameter combination frequency correlation coefficient first less than count, N glmac,1 and N glmac,2 Respectively represent i and i +1 parameter combination mode shape component correlation coefficient second less than count.

[0053] S332: Determine a total weighted dominance score of frequency according to the first less-than count and the weighted average position.

[0054] It should be noted that although the dynamic correlation coefficient of individual i is less than the count of individual i+1, this does not mean that the optimization effect of the local pipeline model of individual i is necessarily better than that of individual i+1. Because the less than count of individual i is mainly concentrated in high-order modes, it shows that its optimization effect on low-order modes is poor. Since structural vibration is usually dominated by low-order modes, it is impossible to scientifically sort individuals only by less than count, and a comprehensive analysis must be conducted in combination with the characteristics of modal distribution.

[0055] Weighted Average Position (WAP) is a statistic used to describe the central tendency of a set of data in a certain sort or position distribution. Its expression is: Where n represents the total number of first parameter combinations, pos j Indicates the position of the jth data, w j Represents the weight of the j-th data, and the sum of the weights is not zero.

[0056] According to the importance of the mode, weights are assigned to the vectors of the dynamic correlation coefficients of each order that are less than the count vector. In order to highlight the influence of the dominant mode, the exponential decay method can be used to reduce the weight at an exponential level: in, w 0 is the initial weight value, β is the exponential decay rate, k is the modal order, d is an arbitrarily decreasing common difference.

[0057] To ensure the stability of the algorithm, the weight vector can be normalized: in, w norm,k is the normalized modal weight value vector, assuming its display expression is: The weight calculation method defaults to taking the first-order mode as the core mode and giving it the highest weight. k To optimize the precision of the first-order mode, the weight vector needs to be adjusted w norm,k The order of the elements in k The highest weight is given to the first-order mode, and the weights of other modes are rearranged in order.

[0058] S333: Determine a total weighted dominance score of the mode shape component according to the second less-than count and the weighted average position.

[0059] Based on the second less than count and weighted average position, the power weighted dominance scores of parameter combination i and parameter combination i+1 can be calculated, and their expressions are: in, and The value range of is 0 to 1. i If the power-weighted dominance score of the individual is high, the individual is considered an excellent individual and can be considered to be able to dominate the individual i +1.

[0060] Likewise, the parameter combination i Combination of parameters i The modal component correlation dominance score of +1 is ,individual i +1 pair of individuals i The dominance score is , whose expressions are: By traversing all first parameter combinations, the relative dominance score matrix between each first parameter combination can be obtained. The relative dominance score matrices of the frequency correlation and vibration mode component correlation between individuals are: in, The subscripts of the matrix elements represent the individual numbers, for example Indicates the first parameter combination i For the first parameter combination j The dynamic weighted dominance score of the frequency correlation of The meaning of the matrix is ​​the same. The first parameter combination i The sum of the power-weighted dominance scores for all other first parameter combinations is defined as the first parameter combination i The total weighted dominance score of: in represents the total weighted dominance score of the frequency of the first combination parameter 1, Represents the total weighted dominance score of the mode shape components of the first combination parameter 1.

[0061] S334. Determine the dynamic weighted dominance score according to the total weighted dominance score of the frequency and the total weighted dominance score of the vibration mode component.

[0062] The dynamic weighted dominance score can be defined as the mean of the total frequency weighted dominance score and the total weighted dominance score of the vibration mode component, and its expression is: Among them, S is the power weighted dominance score, S f is the total weighted dominance score of frequency, S glmac is the total weighted dominance fraction of the modal components, n k The number of all parameter combinations.

[0063] In the above implementation process, by calculating the characteristic frequency residual, the difference between the theoretical calculation frequency and the actual test frequency can be reflected. By using the global-local modal assurance criterion, the vibration mode component correlation between the local pipeline and the global pipeline within the specified degree of freedom can be accurately determined, thereby reflecting the consistency or difference of the vibration form of the pipeline system in different degrees of freedom, and improving the accuracy of evaluating the changes in the dynamic characteristics of the structure. By combining the frequency correlation and the vibration mode component correlation, the dynamic weighted dominance score between each first parameter combination is calculated, thereby reflecting the degree of influence of different parameter combinations on the dynamic characteristics of the pipeline system, providing a basis for the iterative optimization operation of the first parameter combination.

[0064] Example 3 This embodiment is a specific implementation scheme for iteratively optimizing the first parameter combination based on the power weighted dominance score in the above-mentioned embodiment 1.

[0065] Calculating the power weighted dominance scores between the first parameter combinations, and iteratively optimizing the first parameter combinations based on the power weighted dominance scores, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed, may also include: S34. Sort all first parameter combinations according to the power weighted dominance scores.

[0066] All first parameter combinations are ranked according to the power-weighted dominance score, and the parameter combination of the first-ranked individual is regarded as the optimal parameter combination for model optimization in the current population.

[0067] The manner of sorting all first parameter combinations according to the power-weighted dominance scores may include: The power weighted dominance scores between any two first parameter combinations are calculated, and the dominance relationships are sorted to obtain a first sorting result; and the sorting results of all first parameter combinations are determined according to all the first sorting results.

[0068] In the above embodiment, the power weighted dominance score can comprehensively consider the less than count and weighted average position of the power correlation coefficient, so as to more comprehensively reflect the relationship between the advantages and disadvantages of each parameter combination.

[0069] S35. Perform crossover and mutation operations on multiple target parameter combinations ranked top according to the sorting results to generate next-generation parameter combinations.

[0070] In order to make the population approach the optimal combination of parameters, it is necessary to perform crossover and mutation on the current population and generate the second generation population. The crossover operation increases the spatial diversity of the population and prevents the algorithm from converging prematurely. Assume that the crossover ratio is β c , then the number of individuals that need to cross in the population of the first parameter combination of size npop is: Due to N c The parameter combinations need to be crossed in pairs, and Nc / 2 crossovers are required in one generation of population. Assuming there are two parent individuals p1 and p2, the expression is: Where np is the number of genes in the individual, that is, the number of parameters to be optimized in the local pipeline. p 1 and p 2 Perform crossover operation to obtain new individuals p 1,new and p 2,new , whose expression is: in α c is a random coefficient matrix whose elements are uniformly distributed in the interval [0, 1] and whose dimension is the same as p 1 and p 2 It is used to mix the genetic information of two parent individuals in proportion.

[0071] The mutation operation aims to introduce individual diversity, prevent the algorithm from falling into local optimality, and enhance the exploration ability of the model optimization method. Assume that the population size is npop , the variation ratio is β m , then the number of individuals that mutate is: If the current individual is p, the value range of the parameter to be optimized in p is: From this we can get the variable phase length vector σ : in, α m is the variable step length coefficient, which determines the variation range of the parameter value to be optimized. p 1 The number of parameters to be optimized in the variation is: in Indicates rounding up. μ Represents the mutation rate.

[0072] exist np Among the parameters to be optimized, randomly select n μ Locations j pos , recorded as: in is the individual after mutation, σ jpos Indicates that the corresponding j pos The value of the item, that is . Represents the disturbance vector introduced by the mutation, which obeys the normal distribution. For the position where no mutation occurs, the value of the parameter to be optimized remains unchanged, so the individual after the final mutation is expressed as: in, i =(1,2,… np ).

[0073] S36: merge the next-generation parameter combination and the first parameter combination, and repeat the steps of sorting the merged parameter combination and generating the next-generation parameter combination to iterate the first parameter combination.

[0074] After the parent population is sorted by the power-weighted dominance score, crossover and mutation are performed to generate the second-generation population. The second-generation population is sorted again, and the third-generation population is generated through crossover and mutation. This process is repeated until the preset number of iterations is reached or the modified objective function meets the threshold condition, and finally the optimal correction parameter combination is obtained.

[0075] S37. When the iteration is completed, output the top-ranked parameter combination as the target parameter combination.

[0076] In the above implementation process, all the first parameter combinations are sorted by the dynamic weighted dominance score, and the parameter combination with better performance can be accurately identified. This sorting method is based on the dynamic weighted dominance score that comprehensively evaluates the frequency correlation and the vibration type component correlation, so it has higher accuracy and reliability. Crossover and mutation operations are performed on multiple target parameter combinations with higher rankings to generate next-generation parameter combinations with higher potential. The crossover operation can combine the advantages of different parameter combinations, while the mutation operation can introduce new variation factors, thereby increasing the diversity of parameter combinations. In the case of incomplete dynamic test data, by making full use of known modal information, the maximum optimization of the local pipeline model is achieved, which can make up for the impact of insufficient data to a certain extent. The iterative optimization method is also used in the embodiment of the present application, which can gradually stabilize the performance of the parameter combination, reduce the volatility and uncertainty of the parameter combination, and thus improve the reliability of pipeline model optimization.

[0077] Example 4 This embodiment is a specific implementation method of determining initial parameters in Embodiment 1 and generating multiple first parameter combinations.

[0078] Determining the spring stiffness of the plurality of first spring units as initial parameters and generating a plurality of first parameter combinations according to the value range of the initial parameters may include: S21. Determine the natural frequency and vibration mode of each first spring unit respectively.

[0079] In the boundary condition optimization, the spring stiffness in six directions of the local pipeline elastic boundary is used as the optimization parameter to improve the accuracy of the model and reflect the mechanical properties of the actual boundary more accurately.

[0080] The natural frequency is the frequency of the spring unit's free vibration without external excitation, and is an important parameter for describing the dynamic characteristics of the spring unit. The vibration mode is the deformation mode of the spring unit corresponding to the natural frequency, describing the relative displacement of each point of the spring unit during free vibration.

[0081] In terms of structural optimization parameters, the parameters to be optimized can be selected from the initial parameters through sensitivity analysis of natural frequency and vibration mode. A larger sensitivity indicates that the optimization variable has a more significant impact on the dynamic characteristics, so variables with higher sensitivity can be preferentially selected as parameters to be optimized.

[0082] S22. Determine the spring stiffness of the plurality of first spring units as initial parameters according to the natural frequency and the vibration mode.

[0083] Natural frequency ω k For optimization variables p The relative sensitivity expression is: Among them, η r is the relative sensitivity, φ k is the vibration mode, M and K are the mass matrix and stiffness matrix of the local pipeline respectively.

[0084] Assuming a certain vibration mode φ k =[φ k,1 , φ k,2, …, φ k,d ,… φ k,nd ],in nd is the total number of modal degrees of freedom, φ k,d For the d The vibration mode value related to each degree of freedom. d For optimization variables p Relative sensitivity η r ( φ k,d , p ) can be expressed as: In the above embodiment, by accurately calculating the natural frequency of each first spring unit, its vibration characteristics under specific conditions can be understood, providing basic data for subsequent optimization design. Determining the vibration mode of each spring unit can more accurately characterize the deformation of the spring unit during vibration. By analyzing the natural frequency and vibration mode of the first spring unit, the spring stiffness parameters that significantly affect the dynamic characteristics can be identified. These parameters will be used as initial parameters for iterative optimization to generate the target parameter combination of the pipeline model, which can improve the accuracy of the pipeline model.

[0085] Example 5 This embodiment is a practical application example of the unknown boundary pipeline model optimization method in Embodiment 1.

[0086] Please see Figure 3 , Figure 3 A schematic diagram of the pipeline overview provided in the embodiment of the present application. The pipeline in this practical application example is a petrochemical dehydrogenation high-temperature pipeline. Figure 3 The figure shows the section of the petrochemical dehydrogenation high-temperature pipeline with the most significant vibration. The section is about 66.49 m long, 2.8 m in diameter, 0.03 m thick, and consists of 1 mainstream pipeline and 5 branch pipelines.

[0087] The branch pipe is connected to the reactor above and is assumed to be in a consolidated state. The main pipe section is connected through an expansion joint and hoisted by 10 constant force springs (each providing 88949 N force), and a longitudinal limit device is set on the ground. A total of 10 measuring points are arranged along the longitudinal direction of the main pipe, numbered M-01 to M-10. Due to the high temperature of the pipe wall and the outer insulation material, the acceleration sensor cannot be directly installed, so the sensor is installed on the limit pull plate of the expansion joint accessory.

[0088] Please see Figure 4 , Figure 4 A schematic diagram of the optimization results of the optimized model provided in the embodiment of the present application.

[0089] Based on the unknown boundary pipeline model optimization method provided in this application, the local vibration pipeline model is optimized. The first 20 natural frequencies of the optimized local pipeline model are shown in Table 1, and compared with the local vibration pipeline modal parameters identified based on the measured acceleration data. For the comparison results, please refer to Figure 3 . The measured results show that there are three-order dominant frequencies in the local vibration pipeline, which are 26.2 Hz, 52.4 Hz and 78.6 Hz respectively. The natural frequencies of the optimized local pipeline finite element model also exist at 25.9 Hz, 51.7 Hz and 77.8 Hz, which are consistent with the measured results, with relative errors of 1.1%, 1.3% and 1.1% respectively. In addition, the third-order modal vibration shape of the optimized finite element model is highly consistent with the vibration shape identified based on the measured acceleration data, and the MAC (Modal Assurance Criterion) value is above 97%. It shows that the vibration shape of the optimized finite element model is highly consistent with the vibration shape identified based on the measured acceleration data, which can verify the accuracy and reliability of the optimized model.

[0090] Table 1. The first 20 natural frequencies of the model after local pipeline optimization Typically, model optimization requires complete modal information, such as the first n frequencies and vibration modes. However, in this case, the measured frequency orders of 26.2 Hz, 52.4 Hz, and 78.6 Hz are unclear, and the modes may not be fully excited. Despite this, the optimized finite element model can still accurately match these three frequencies, and the modal vibration modes are consistent with the results identified based on the measured data, indicating that this application has strong adaptability and robustness in the case of incomplete modal information, and can effectively optimize the finite element model of the local vibration pipeline.

[0091] Example 6 Please see Figure 5 , Figure 5 A schematic diagram of an unknown boundary pipeline model optimization device provided in an embodiment of the present application. The unknown boundary pipeline model optimization device 50 may include: The construction module 51 is used to construct an elastic boundary unit to simulate the boundary stiffness of a local pipeline model with an unknown boundary; wherein the elastic boundary unit includes spring units in multiple directions; and one spring unit simulates the translational stiffness or torsional stiffness in one direction.

[0092] The determination module 52 is used to determine the spring stiffness of multiple first spring units as initial parameters, and generate multiple first parameter combinations according to the value range of the initial parameters; wherein the first spring unit is any spring unit.

[0093] The calculation module 53 is used to calculate the power weighted dominance scores between the first parameter combinations, and iteratively optimize the first parameter combinations based on the power weighted dominance scores, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed.

[0094] Optionally, the calculation module 53 may be specifically used for: Determine the frequency correlation between the local pipeline and the global pipeline represented by the first parameter combination based on the characteristic frequency residual; Determining the correlation of vibration mode components of the local pipeline and the global pipeline represented by the first parameter combination within a specified degree of freedom range based on the global-local modal assurance criterion; A dynamically weighted dominance score between the first parameter combinations is determined based on the frequency correlation and the mode shape component correlation.

[0095] Optionally, the calculation module 53 may be specifically used for: ranking all first parameter combinations according to the power-weighted dominance scores; According to the sorting results, crossover and mutation operations are performed on multiple target parameter combinations ranked top to generate next-generation parameter combinations; Merging the second generation parameter combination with the first parameter combination, and repeatedly performing the steps of sorting the merged parameter combination and generating the second generation parameter combination to iterate the first parameter combination; When the iteration is completed, the top parameter combination is output as the target parameter combination.

[0096] Optionally, the calculation module 53 may be specifically used for: Determining a first less-than count based on frequency correlation, and determining a second less-than count based on mode shape component correlation; determining a frequency total weighted dominance score based on the first less-than count and the weighted average position; Determining a total weighted dominance score of the mode shape component based on the second less than count and the weighted average position; The dynamic weighted dominance score is determined based on the total weighted dominance score of the frequency and the total weighted dominance score of the mode shape components.

[0097] Optionally, the determination module 52 may be specifically configured to: Determine the natural frequency and mode shape of each first spring unit respectively; The spring stiffness of the plurality of first spring units is determined as an initial parameter according to the natural frequency and the vibration mode.

[0098] Optionally, the calculation module 53 may be specifically used for: Calculating the power weighted dominance scores between any two first parameter combinations, and sorting the dominance relationships to obtain a first sorting result; The sorting results of all first parameter combinations are determined according to all first sorting results.

[0099] It should be understood that, when performing the unknown boundary pipeline model optimization, the various modules of the unknown boundary pipeline model optimization provided in the above embodiments are only illustrated by the division of the various functional modules in the above description. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0100] The functional modules in the above embodiments may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit, and the above integrated unit may be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the protection scope of the embodiments of the present application.

[0101] Based on the same application concept, an embodiment of the present application also provides a computer device, which may include a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the method described in the above description is implemented.

[0102] Based on the same application concept, an embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the method described in the above description is implemented.

[0103] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for optimizing an unknown boundary pipeline model, characterized in that: include: Constructing an elastic boundary unit to simulate the boundary stiffness of a local pipeline model of an unknown boundary; wherein the elastic boundary unit includes spring units in multiple directions; one of the spring units simulates the translational stiffness or torsional stiffness in one direction; Determine the spring stiffness of a plurality of first spring units as an initial parameter, and generate a plurality of first parameter combinations according to the value range of the initial parameter; wherein the first spring unit is any of the spring units; The power weighted dominance scores between the first parameter combinations are calculated, and the first parameter combinations are iteratively optimized based on the power weighted dominance scores, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed.

2. The method according to claim 1, characterized in that The calculating of the power weighted dominance scores between the first parameter combinations and iteratively optimizing the first parameter combinations based on the power weighted dominance scores to determine the target parameter combination of the pipeline model when the iterative optimization is completed includes: Determine the frequency correlation between the local pipeline and the global pipeline represented by the first parameter combination based on the characteristic frequency residual; Determining the correlation of vibration mode components between the local pipeline and the global pipeline represented by the first parameter combination within a specified degree of freedom range based on a global-local modal assurance criterion; A dynamically weighted dominance score between the first parameter combinations is determined based on the frequency correlation and the mode shape component correlation.

3. The method according to claim 2, characterized in that The calculating of the power weighted dominance scores between the first parameter combinations and iteratively optimizing the first parameter combinations based on the power weighted dominance scores to determine the target parameter combination of the pipeline model when the iterative optimization is completed also includes: sorting all of the first parameter combinations according to the power-weighted dominance scores; According to the sorting result, a crossover operation and a mutation operation are performed on the plurality of target parameter combinations ranked first to generate a next-generation parameter combination; Merging the second generation parameter combination with the first parameter combination, and repeatedly performing the steps of sorting the merged parameter combination and generating the second generation parameter combination, so as to iterate the first parameter combination; When the iteration is completed, the top parameter combination is output as the target parameter combination.

4. The method according to claim 2, characterized in that: The step of determining the dynamic weighted dominance scores between the first parameter combinations according to the frequency correlation and the vibration mode component correlation comprises: Determine a first less-than count based on the frequency correlation, and determine a second less-than count based on the mode shape component correlation; determining a frequency total weighted dominance score based on the first less-than count and the weighted average position; Determining a total weighted dominance score of mode shape components according to the second less-than count and the weighted average position; The dynamic weighted dominance score is determined according to the total frequency weighted dominance score and the total mode shape component weighted dominance score.

5. The method according to claim 4, characterized in that The power-weighted dominance score is characterized as: Where S is the power weighted control score, S f is the total weighted dominance score of the frequency, S glmac is the total weighted dominance fraction of the modal component, n k The number of all parameter combinations.

6. The method according to claim 1, characterized in that The step of determining the spring stiffness of the plurality of first spring units as initial parameters and generating a plurality of first parameter combinations according to the value range of the initial parameters includes: Determine the natural frequency and vibration mode of each of the first spring units respectively; The spring stiffness of the plurality of first spring units is determined as an initial parameter according to the natural frequency and the vibration mode.

7. The method according to claim 3, characterized in that The step of sorting all the first parameter combinations according to the power weighted dominance score comprises: Calculating the power weighted dominance scores between any two of the first parameter combinations, and sorting the dominance relationships to obtain a first sorting result; The sorting results of all the first parameter combinations are determined according to all the first sorting results.

8. An unknown boundary pipeline model optimization device, characterized in that: include: A construction module is used to construct an elastic boundary unit to simulate the boundary stiffness of a local pipeline model of an unknown boundary; wherein the elastic boundary unit includes spring units in multiple directions; one of the spring units simulates the translational stiffness or torsional stiffness in one direction; A determination module, used to determine the spring stiffness of a plurality of first spring units as an initial parameter, and generate a plurality of first parameter combinations according to a value range of the initial parameter; wherein the first spring unit is any of the spring units; The calculation module is used to calculate the power weighted dominance score between each of the first parameter combinations, and iteratively optimize the first parameter combinations based on the power weighted dominance score, so as to determine the target parameter combination of the pipeline model when the iterative optimization is completed.

9. A computer device, characterized in that: The computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images