Pump head structure parameter intelligent optimization design method

Abstract

The invention belongs to the technical field of intelligent optimization, and discloses a pump head structure parameter intelligent optimization design method, which comprises the following steps: collecting pump head structure parameter data, manufacturing cost data and user behavior data; preprocessing the collected pump head structure parameter data, manufacturing cost data and user behavior data to obtain a structure parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set; performing association fusion on the structure parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set; training according to the comprehensive feature data set to obtain a performance index prediction model, and predicting through the performance index prediction model to obtain n performance indexes of the pump head; according to the n performance indexes of the pump head, using an improved multi-objective optimization algorithm to generate a corresponding structure parameter optimization scheme; the pump head design can be effectively optimized, and the pump head performance is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent optimization technology, and more specifically, to an intelligent optimization design method for pump head structural parameters. Background Art

[0002] The existing intelligent optimization design method for pump head structural parameters mainly has the following problems:

[0003] Without the support of dictionary learning, feature extraction may not be able to capture the complex structure of the data, affecting the expressiveness and generalization ability of the model, resulting in poor performance; without the support of sparse coding or sparsity regularization, data samples may need to be represented by more dictionary atoms, resulting in a less compact and effective representation; not only does it increase the computational cost, but it may also introduce unnecessary noise and redundant information; it is impossible to balance the trade-off between dictionary reconstruction error and sparsity by dynamically adjusting the sparsity regularization coefficient, which may lead to unstable performance of the model on different data sets or at different time steps;

[0004] Traditional optimization methods usually only focus on a single goal, which may lead to other important performance indicators being ignored. In pump head design, single-goal optimization may only focus on efficiency, while ignoring manufacturing costs or durability, which ultimately leads to poor results in practical applications; without introducing non-dominated sorting and crowding calculation techniques, the algorithm may find it difficult to find a balance between multiple goals, resulting in the optimization results being biased towards a certain performance indicator and damaging other indicators; traditional random sampling methods may lead to uneven distribution of initial population parameters, resulting in insufficient exploration of the solution space and increasing the risk of the algorithm falling into a local optimum; without using crowding calculation technology, the solution may be concentrated in certain areas of the solution space, resulting in insufficient population diversity and increasing the risk of the algorithm converging prematurely; lack of population diversity will limit the algorithm from exploring new possible solutions, resulting in results that are not comprehensive and robust enough; without considering the mechanism of dynamically adjusting nonlinear crowding adjustment parameters, the algorithm may lack flexibility in the multi-objective optimization process; it may converge to a local optimal solution prematurely, resulting in insufficient overall optimization.

[0005] In view of this, the present invention proposes a pump head structural parameter intelligent optimization design method to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for intelligent optimization design of pump head structural parameters, comprising:

[0007] S1, collecting pump head structural parameter data, manufacturing cost data and user behavior data;

[0008] S2, preprocessing the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain a structural parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set; correlating and fusing the structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set;

[0009] S3, training a performance indicator prediction model based on the comprehensive feature data set, and predicting n performance indicators of the pump head through the performance indicator prediction model;

[0010] S4, generating a corresponding structural parameter optimization scheme using an improved multi-objective optimization algorithm according to n performance indicators of the pump head;

[0011] S5. Perform simulation tests on the pump head that implements the corresponding structural parameter optimization scheme to obtain n performance indicators of the optimized pump head; evaluate whether the n performance indicators of the optimized pump head reach the preset performance indicator thresholds, and use the structural parameter optimization scheme corresponding to the preset performance indicator threshold as the final intelligent optimization design scheme for the pump head structural parameters.

[0012] Furthermore, the pump head structural parameter data includes the length of the pressing rod, the spring stiffness, the pump chamber volume, the outlet aperture, the pump head material type, the pump head size and the nozzle size; the manufacturing cost data includes the material cost and the processing cost; and the user behavior data includes the pressing frequency and the pressing duration.

[0013] Furthermore, the method of preprocessing the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain a structural parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set includes:

[0014] Linear interpolation method is used to fill missing values ​​in the collected pump head structure parameter data, manufacturing cost data and user behavior data; density clustering algorithm is used to identify and eliminate outliers in the pump head structure parameter data, manufacturing cost data and user behavior data;

[0015] By improving the online dictionary learning algorithm, feature extraction is performed on the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets; the structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets are standardized and normalized to obtain normalized structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets.

[0016] Furthermore, the method for extracting features from the collected pump head structural parameter data, manufacturing cost data and user behavior data by improving the online dictionary learning algorithm includes:

[0017] S41. Initialize dictionary D∈R d′×k ; Where R is a real number set, that is, each element in the dictionary is a real number; d′ is the number of characteristic types of pump head structural parameter data, manufacturing cost data, and user behavior data; k is the number of dictionary atoms;

[0018] S42. Input the current time step data: Among them, a t is the input data of the current time step; It is the pump head structural parameter data; is manufacturing cost data; is the user behavior data; T is the transpose of the input data matrix at the current time step; t is the current time step;

[0019] S43, solving the current input data a by optimizing the objective function t The sparse coding coefficients under the dictionary of the previous time step, the optimization objective function is: in, To use the dictionary D t-1 β Reconstruction t The error; λ‖β‖1 is the sparsity regularization term; λ is the sparsity regularization coefficient, which controls the sparsity of sparse coding; β t is the sparse coding coefficient, which means the current input data a t Sparse representation under the dictionary of the previous time step; β is the sparse coefficient vector; argmin β is the optimization target; t-1 is the previous time step;

[0020] S44. Dynamically adjust the sparsity regularization coefficient through a sparsity regularization adjustment formula; the sparsity regularization adjustment formula is: Among them, λ′ is the sparsity regularization coefficient after dynamic adjustment; η is the control dictionary D t-1 β Reconstruction t is the adjustment factor of the error on λ′; γ is the adjustment factor of the control time step on λ′;

[0021] S45, use iterative soft threshold algorithm to optimize β t , until the preset maximum number of iterations is reached to minimize the reconstruction error; by optimizing β t , so that the input data a of the current time step t The dictionary D that can be in the previous time step t-1 Reconstructed by sparse coefficients; adjust and update the dictionary at each time step, and finally obtain the sparse coding coefficient β tThat is, it is the feature representation of the data at the current time step. All sparse coefficient vectors are combined into a feature data set to obtain the pump head structure parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set.

[0022] Furthermore, the method of associating and fusing the structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set includes:

[0023] The structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set are spliced ​​according to samples to obtain a multidimensional feature data set; the K-means clustering algorithm is applied to perform cluster analysis on the spliced ​​multidimensional feature data set, and the preset cluster number K is selected, and the data is divided into K clusters using the K-means algorithm; according to the clustering results, a cluster label is assigned to each sample, and the label is the cluster number to which the sample belongs, and the comprehensive characteristics of each cluster are represented by the cluster center, and the cluster center is the mean of the sample characteristics in the cluster; according to the clustering results and the cluster center, a comprehensive feature data set after association and fusion is finally formed.

[0024] Furthermore, the training method of the performance indicator prediction model includes:

[0025] The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample contains a historical comprehensive feature data set and n performance indicators of the corresponding pump head;

[0026] Construct a performance indicator prediction model, which includes an input layer, an encoding layer, a multimodal fusion layer, a decoding layer and an output layer; the input layer is used to input a historical comprehensive feature data set, and the output layer is used to output n performance indicators of the pump head; the output layer is provided with neurons equal to the number of performance indicator types of the pump head, and each neuron corresponds to the prediction probability of a performance indicator of the pump head; a softmax function is used as an activation function; the performance indicator prediction model is a multimodal Transformer model;

[0027] Use multi-classification cross entropy as the loss function of the model to measure the difference between the model's predicted value and the actual value; use the training set to train the performance indicator prediction model, and update the model parameters through the back propagation algorithm to minimize the loss function; use the mini-batch gradient descent method to update the model once per batch; use the validation set to evaluate the performance of the performance indicator prediction model by calculating the recall rate indicator;

[0028] The SGD optimization algorithm is selected as the optimizer, and the model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance no longer improves or the preset stopping condition is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained performance indicator prediction model is used to predict the current comprehensive feature data set to obtain n performance indicators of the pump head.

[0029] Furthermore, the n performance indicators of the pump head include flow consistency, pressure resistance, sealing, durability, corrosion resistance, pressing comfort and material utilization.

[0030] Furthermore, the method of generating a corresponding structural parameter optimization scheme using an improved multi-objective optimization algorithm according to n performance indicators of the pump head includes:

[0031] The initial population is generated by Latin hypercube sampling. The improved NSGA-II algorithm is used to perform non-dominated sorting on all individuals in the initial population and calculate the non-linear crowding degree. The parent individuals are selected from all individuals in the initial population based on the non-dominated level and crowding degree. The parent individuals are crossovered and mutated to generate the offspring individuals. The parent individuals and the offspring individuals are merged, and the next generation population is selected based on the non-dominated sorting results and non-linear crowding degree. Iteration is continued until the preset maximum number of iterations is reached, and the optimal solution set is obtained. The comprehensive scores of all solutions in the optimal solution set are calculated by weighted fusion, and the solution with the largest comprehensive score is selected as the final generated structural parameter optimization solution. The specific steps are as follows:

[0032] S81. The preset pump head structural parameter data includes m types of structural parameters, and the structural parameter space is defined as: X = {X1, X2, ..., X i ,...,X m}; where X i is the i-th structural parameter; X m is the mth structural parameter; X i The value range is [L i ,U i ]; L i and U i They are the structural parameters X i The minimum and maximum values ​​of ; i is the index of the structural parameter type;

[0033] S82. For each structural parameter X i , the value range [L i ,U i ] is divided into N intervals; the width of each interval is: Where ΔX i is the width of the interval; randomly sample a sample point in each of the N intervals and arrange them in a random order to obtain N arranged sampling points;

[0034] Combine the permuted sampling points of all structural parameters to generate the initial population P init =

[0035] {p1,p2,...,p α ,...,p N}; where p N is the Nth individual in the population; p α is the αth individual in the population, p α ={X α1 ,X α2 ,...,X αm};X α1 is the αth permutation sampling point of the 1st structural parameter;

[0036] X α2 is the αth permutation sampling point of the second structural parameter; X αm is the αth permutation sampling point of the mth structural parameter; the initial population is represented by a matrix:

[0037] S83. Classify each individual in the initial population according to the non-dominated sorting, and arrange the non-dominated sorting in descending order to obtain θ non-dominated levels: F1, F2, ..., F θ ; Among them, F1 is the optimal solution set in the initial population; F2 is the suboptimal solution set in the initial population; F θ is the θth optimal solution set in the initial population; θ is the index of the non-dominated level number;

[0038] S84. Calculate the nonlinear crowding value of each individual in the initial population. The specific formula is: ρ

[0039] Among them, ρ(p α ) is the population individual p α The nonlinear crowding value of α+1 and d α-1 are the individuals p in the current population respectively. α Its adjacent solution p α+1 and p α-1 The Euclidean distance between them; σ is the adjustment parameter for controlling nonlinear congestion; exp() is an exponential function with e as the base; the adjustment parameter for controlling nonlinear congestion is adjusted by the adjustment parameter restriction formula; the adjustment parameter restriction formula is: Among them, σ′ is the adjustment parameter for controlling the nonlinear crowding degree after restriction; Q is the population size; E is the number of non-dominated levels; C1 is a constant for controlling the influence of the population size on the adjustment parameter, C2 is an offset constant; C3 is a constant for controlling the influence of the number of non-dominated levels on the adjustment parameter;

[0040] S85, performing a selection operation according to the non-dominated sorting result and the non-linear crowding value of each individual in the initial population until N′ parent individuals are selected; arbitrarily selecting two parent individuals from the selected N′ parent individuals to perform a crossover operation to generate two new crossover individuals; collecting all the new crossover individuals generated by performing a crossover operation on all pairs of parent individuals to obtain crossover offspring individuals;

[0041] Perform mutation operation on the selected N' parent individuals, use polynomial mutation to perturb the genes of N' parent individuals to generate new mutation individuals. The specific formula is: z ′=p z +δ·(U′-L′); where p z ′ is the generated new individual; p z is the zth parent individual among N′ parent individuals; δ is the disturbance factor of the mutation operation; U′ is the upper limit of the mutation operation; L′ is the lower limit of the mutation operation; collect the mutated new individuals of all parent individuals to obtain the mutated offspring individuals;

[0042] S86, merge the crossover offspring individuals and mutation offspring individuals obtained through the crossover operation and mutation operation with the N′ parent individuals to obtain a new population P new ;

[0043] S87, repeating steps S83-S86 until the preset maximum number of iterations is reached; when the iteration stops, calculating the optimal solution set of the current population; calculating the comprehensive scores of all solutions in the optimal solution set by weighted fusion, and selecting the solution with the largest comprehensive score as the final generated structural parameter optimization solution.

[0044] Furthermore, the method of performing simulation test on the pump head implementing the corresponding structural parameter optimization scheme to obtain n performance indicators of the optimized pump head includes:

[0045] According to the structural parameter optimization plan, the structural parameter data of the pump head are adjusted to obtain the optimized structural parameter data of the pump head; a three-dimensional model of the pump head is constructed based on the optimized structural parameter data of the pump head using three-dimensional modeling software; the constructed three-dimensional model of the pump head is imported into the finite element analysis software for simulation calculation, and then n performance indicators of the optimized pump head are obtained.

[0046] Furthermore, the method for evaluating whether the n performance indicators of the optimized pump head reach a preset performance indicator threshold includes:

[0047] Compare the obtained n performance indicators of the optimized pump head with the preset performance indicator thresholds:

[0048] If the obtained n performance indicators of the optimized pump head are less than the preset performance indicator threshold, it is determined that the n performance indicators of the optimized pump head do not reach the preset performance indicator threshold;

[0049] If the obtained n performance indicators of the optimized pump head are greater than or equal to the preset performance indicator threshold, it is determined that the n performance indicators of the optimized pump head reach the preset performance indicator threshold.

[0050] Technical effects and advantages of the intelligent optimization design method for pump head structural parameters of the present invention:

[0051] The present invention automatically extracts meaningful features from data through dictionary learning without the need for artificially designed features; dictionary learning can adaptively adjust the dictionary according to the structure and characteristics of the data to capture potential patterns in the data; sparse coding allows each data sample to be represented by a small number of dictionary atoms (dictionary elements); by introducing sparsity regularization, the method makes the representation of each data sample more compact and effective, thereby improving computational efficiency; by dynamically adjusting the sparsity regularization coefficient, the method can adaptively adjust the dictionary learning process at different time steps; the dynamic adjustment of the regularization coefficient helps to balance the trade-off between dictionary reconstruction error and sparsity, thereby improving the stability and robustness of the model;

[0052] It can optimize multiple pump head performance indicators at the same time, solving the problem that traditional optimization methods only focus on a single target; through technologies such as non-dominated sorting and crowding, the algorithm can find a balance between multiple targets; generate an optimal solution set to obtain the best structural parameter optimization solution; use Latin hypercube sampling to generate the initial population, which can evenly cover the structural parameter space and avoid the problem of uneven parameter distribution that may be caused by traditional random sampling; through uniform sampling, the algorithm can explore the solution space more comprehensively, thereby improving the efficiency and accuracy of optimization; through non-dominated sorting and crowding calculation, this method can effectively screen out the optimal solution set, and avoid solutions from concentrating in the same position by calculating nonlinear crowding, thereby maintaining population diversity; by dynamically adjusting the adjustment parameters of nonlinear crowding, the algorithm can flexibly adjust parameters according to the actual situation of the population; it helps to avoid premature convergence or insufficient search in the multi-objective optimization process, and ensure better optimization results. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of a process of intelligent optimization design method for pump head structural parameters;

[0054] Figure 2 This is a schematic diagram of the structure of an intelligent optimization design system for pump head structural parameters. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Example 1

[0057] See also Figure 1 As shown, this embodiment provides a method for intelligent optimization design of pump head structural parameters, including:

[0058] S1, collecting pump head structural parameter data, manufacturing cost data and user behavior data;

[0059] S2, preprocessing the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain a structural parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set; correlating and fusing the structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set;

[0060] S3, training a performance indicator prediction model based on the comprehensive feature data set, and predicting n performance indicators of the pump head through the performance indicator prediction model;

[0061] S4, generating a corresponding structural parameter optimization scheme using an improved multi-objective optimization algorithm according to n performance indicators of the pump head;

[0062] S5. Perform simulation tests on the pump head that implements the corresponding structural parameter optimization scheme to obtain n performance indicators of the optimized pump head; evaluate whether the n performance indicators of the optimized pump head reach the preset performance indicator thresholds, and use the structural parameter optimization scheme corresponding to the preset performance indicator threshold as the final intelligent optimization design scheme for the pump head structural parameters.

[0063] The pump head structural parameter data includes the length of the pressing rod, spring stiffness, pump chamber volume, outlet aperture, pump head material type, pump head size and nozzle size; the manufacturing cost data includes material cost and processing cost; the user behavior data includes pressing frequency and pressing duration.

[0064] The method of preprocessing the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain a structural parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set includes:

[0065] Linear interpolation method is used to fill missing values ​​in the collected pump head structure parameter data, manufacturing cost data and user behavior data; density clustering algorithm is used to identify and eliminate outliers in the pump head structure parameter data, manufacturing cost data and user behavior data;

[0066] By improving the online dictionary learning algorithm, feature extraction is performed on the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets; the structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets are standardized and normalized to obtain normalized structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets.

[0067] The method of extracting features from the collected pump head structure parameter data, manufacturing cost data and user behavior data by improving the online dictionary learning algorithm includes:

[0068] S41. Initialize dictionary D∈R d′×k ; Where R is a real number set, that is, each element in the dictionary is a real number; d′ is the number of characteristic types of pump head structural parameter data, manufacturing cost data, and user behavior data; k is the number of dictionary atoms;

[0069] S42. Input the current time step data: Among them, a t is the input data of the current time step; It is the pump head structural parameter data; is manufacturing cost data; is the user behavior data; T is the transpose of the input data matrix at the current time step; t is the current time step;

[0070] S43, solving the current input data a by optimizing the objective function t The sparse coding coefficients under the dictionary of the previous time step, the optimization objective function is: in, To use the dictionary D t-1 β Reconstruction t The error; λ‖β‖1 is the sparsity regularization term; λ is the sparsity regularization coefficient, which controls the sparsity of sparse coding; β t is the sparse coding coefficient, which means the current input data a t Sparse representation under the dictionary of the previous time step; β is the sparse coefficient vector; argmin β is the optimization target; t-1 is the previous time step;

[0071] S44. Dynamically adjust the sparsity regularization coefficient through a sparsity regularization adjustment formula; the sparsity regularization adjustment formula is: Among them, λ′ is the sparsity regularization coefficient after dynamic adjustment; η is the control dictionary D t-1 β Reconstruction t is the adjustment factor of the error on λ′; γ is the adjustment factor of the control time step on λ′;

[0072] S45, use iterative soft threshold algorithm to optimize β t , until the preset maximum number of iterations is reached to minimize the reconstruction error; by optimizing β t , so that the input data a of the current time step t The dictionary D that can be in the previous time step t-1 Reconstructed by sparse coefficients; adjust and update the dictionary at each time step, and finally obtain the sparse coding coefficient β t That is, it is the feature representation of the data at the current time step. All sparse coefficient vectors are combined into a feature data set to obtain the pump head structure parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set.

[0073] The method of associating and fusing the structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set includes:

[0074] The structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set are spliced ​​according to samples to obtain a multidimensional feature data set; the K-means clustering algorithm is applied to perform cluster analysis on the spliced ​​multidimensional feature data set, and the preset cluster number K is selected, and the data is divided into K clusters using the K-means algorithm; according to the clustering results, a cluster label is assigned to each sample, and the label is the cluster number to which the sample belongs, and the comprehensive characteristics of each cluster are represented by the cluster center, and the cluster center is the mean of the sample characteristics in the cluster; according to the clustering results and the cluster center, a comprehensive feature data set after association and fusion is finally formed.

[0075] The training methods for the performance indicator prediction model include:

[0076] The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample contains a historical comprehensive feature data set and n performance indicators of the corresponding pump head;

[0077] Construct a performance indicator prediction model, which includes an input layer, an encoding layer, a multimodal fusion layer, a decoding layer, and an output layer; the input layer is used to input a historical comprehensive feature data set, and the output layer is used to output n performance indicators of the pump head; the output layer is set with neurons equal to the number of performance indicator types of the pump head, and each neuron corresponds to the prediction probability of a performance indicator of the pump head; the softmax function is used as the activation function; the performance indicator prediction model is a multimodal Transformer model;

[0078] Use multi-classification cross entropy as the loss function of the model to measure the difference between the model's predicted value and the actual value; use the training set to train the performance indicator prediction model, and update the model parameters through the back propagation algorithm to minimize the loss function; use the mini-batch gradient descent method to update the model once per batch; use the validation set to evaluate the performance of the performance indicator prediction model by calculating the recall rate indicator;

[0079] The SGD optimization algorithm is selected as the optimizer, and the model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance no longer improves or the preset stopping condition is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained performance indicator prediction model is used to predict the current comprehensive feature data set to obtain n performance indicators of the pump head.

[0080] The n performance indicators of the pump head include flow consistency, pressure resistance, sealing, durability, corrosion resistance, pressing comfort and material utilization rate; flow consistency specifically means that when using a press-type pump head, the fluid flow generated by each press with the same force remains relatively stable, that is, the amount of liquid flowing out of the pump head each time the force is pressed is roughly the same;

[0081] This flow consistency is crucial to ensure the product's use effect and user experience, especially in situations where the amount of fluid needs to be precisely controlled. For example, when dispensing lotions in the cosmetics field, flow consistency determines whether the amount of lotion produced by each press is in line with expectations. When preparing condiments in the food field, flow consistency ensures the accuracy of each condiment dosage.

[0082] Pressure resistance refers to the ability of the pump head to withstand external pressure without deformation or damage during use. Pressure resistance is directly related to the safety and reliability of the pump head and avoids damage during use or transportation. Sealing refers to the ability of the pump head to prevent liquid leakage in working and non-working states, ensuring the limit of liquid storage to avoid waste or pollution, which is especially important in the distribution of chemicals or drugs. Durability refers to the ability of the pump head to maintain stable performance during repeated, frequent and long-term use, which is used to extend the service life of the product, reduce the frequency of user replacement, and improve user experience and environmental value.

[0083] Corrosion resistance refers to the ability of the pump head to resist corrosion when it comes into contact with chemicals (such as detergents, cosmetics or drugs), and is used to improve the stability of the pump head in chemical environments; pressing comfort refers to the user's feel and ease of operation when rotating the pump head, and is used to improve user experience and reduce fatigue during long-term use; material utilization refers to the efficiency of material use in the design and manufacturing process of the pump head, reflecting whether there is material waste, and is used to optimize production costs, improve environmental protection, and reduce unnecessary material waste.

[0084] The method of using the improved multi-objective optimization algorithm to generate the corresponding structural parameter optimization scheme according to the n performance indicators of the pump head includes:

[0085] The initial population is generated by Latin hypercube sampling. The improved NSGA-II algorithm is used to perform non-dominated sorting on all individuals in the initial population and calculate the non-linear crowding degree. The parent individuals are selected from all individuals in the initial population based on the non-dominated level and crowding degree. The parent individuals are crossovered and mutated to generate the offspring individuals. The parent individuals and the offspring individuals are merged, and the next generation population is selected based on the non-dominated sorting results and non-linear crowding degree. Iteration is continued until the preset maximum number of iterations is reached, and the optimal solution set is obtained. The comprehensive scores of all solutions in the optimal solution set are calculated by weighted fusion, and the solution with the largest comprehensive score is selected as the final generated structural parameter optimization solution. The specific steps are as follows:

[0086] S81. The preset pump head structural parameter data includes m types of structural parameters, and the structural parameter space is defined as: X = {X1, X2, ..., X i ,...,X m}; where X i is the i-th structural parameter; X m is the mth structural parameter; X i The value range is [L i ,U i ]; L i and U i They are the structural parameters X i The minimum and maximum values ​​of ; i is the index of the structural parameter type;

[0087] S82. For each structural parameter X i , the value range [L i ,U i ] is divided into N intervals; the width of each interval is: Where ΔX i is the width of the interval; randomly sample a sample point in each of the N intervals and arrange them in a random order to obtain N arranged sampling points;

[0088] Combine the permuted sampling points of all structural parameters to generate the initial population P init =

[0089] {p1,p2,...,p α ,...,p N}; where p N is the Nth individual in the population; p α is the αth individual in the population, p α ={X α1 ,X α2 ,...,X αm};X α1 is the αth permutation sampling point of the 1st structural parameter;

[0090] X α2 is the αth permutation sampling point of the second structural parameter; X αm is the αth permutation sampling point of the mth structural parameter; the initial population is represented by a matrix: Among them, P init The number of rows is the number of intervals N, and the number of columns is the number of types m containing structural parameters in the pump head structural parameter data;

[0091] For example, the pump head structural parameter data contains m=2 structural parameters, the value range of structural parameter X1 is [0,10], the value range of structural parameter X2 is [20,30], and the value range is divided into 5 intervals:

[0092] The interval of X1 is: [0,2), [2,4), [4,6), [6,8), [8,10);

[0093] The interval of X2 is: [20,22), [22,24), [24,26), [26,28), [28,30);

[0094] Randomly sample a point in each interval, for example: the sampling points of X1 are {1.5, 2.8, 4.3, 7.6, 9.2}; the sampling points of X2 are {21.3, 23.5, 25.7, 27.8, 29.1}; randomly arrange the sampling points to obtain a set of arranged sampling points; for example: randomly arrange the sampling points of X1 to {4.3, 9.2, 1.5, 7.6, 2.8}; randomly arrange the sampling points of X2 to {27.8, 21.3, 29.1, 23.5, 25.7}; combine the arranged sampling point sets of all structural parameters to generate the initial population

[0095] S83. Classify each individual in the initial population according to the non-dominated sorting, and arrange the non-dominated sorting in descending order to obtain θ non-dominated levels: F1, F2, ..., Fθ ; Among them, F1 is the optimal solution set in the initial population; F2 is the suboptimal solution set in the initial population; F θ is the θth optimal solution set in the initial population; θ is the index of the non-dominated level number;

[0096] S84. Calculate the nonlinear crowding value of each individual in the initial population. The specific formula is: ρ

[0097] Among them, ρ(p α ) is the population individual p α The nonlinear crowding value of α+1 and d α-1 are the individuals p in the current population respectively. α Its adjacent solution p α+1 and p α-1 The Euclidean distance between them; σ is the adjustment parameter for controlling nonlinear congestion; exp() is an exponential function with e as the base; the adjustment parameter for controlling nonlinear congestion is adjusted by the adjustment parameter restriction formula; the adjustment parameter restriction formula is: Among them, σ′ is the adjustment parameter for controlling the nonlinear crowding degree after restriction; Q is the population size; E is the number of non-dominated levels; C1 is a constant for controlling the influence of the population size on the adjustment parameter, C2 is an offset constant; C3 is a constant for controlling the influence of the number of non-dominated levels on the adjustment parameter;

[0098] For example, if the population size is 100, the number of non-dominated levels is 5, the constant controlling the influence of the population size on the adjustment parameter is 0.5, the offset constant is 1, and the constant controlling the influence of the number of non-dominated levels on the adjustment parameter is 2, then the adjustment parameter for controlling the nonlinear crowding after restriction is approximately 1.4.

[0099] S85, performing a selection operation according to the non-dominated sorting result and the non-linear crowding value of each individual in the initial population until N′ parent individuals are selected; arbitrarily selecting two parent individuals from the selected N′ parent individuals to perform a crossover operation to generate two new crossover individuals; collecting all the new crossover individuals generated by performing a crossover operation on all pairs of parent individuals to obtain crossover offspring individuals;

[0100] Perform mutation operation on the selected N' parent individuals, use polynomial mutation to perturb the genes of N' parent individuals to generate new mutation individuals. The specific formula is: z ′=p z +δ·(U′-L′); where p z ′ is the generated new individual; p zis the zth parent individual among N′ parent individuals; δ is the disturbance factor of the mutation operation; U′ is the upper limit of the mutation operation; L′ is the lower limit of the mutation operation; collect the mutated new individuals of all parent individuals to obtain the mutated offspring individuals;

[0101] S86, merge the crossover offspring individuals and mutation offspring individuals obtained through the crossover operation and mutation operation with the N′ parent individuals to obtain a new population P new ;

[0102] S87, repeating steps S83-S86 until the preset maximum number of iterations is reached; when the iteration stops, calculating the optimal solution set of the current population; calculating the comprehensive scores of all solutions in the optimal solution set by weighted fusion, and selecting the solution with the largest comprehensive score as the final generated structural parameter optimization solution.

[0103] The method of performing simulation test on the pump head implementing the corresponding structural parameter optimization scheme to obtain n performance indicators of the optimized pump head includes:

[0104] According to the structural parameter optimization plan, the structural parameter data of the pump head are adjusted to obtain the optimized structural parameter data of the pump head; a three-dimensional model of the pump head is constructed based on the optimized structural parameter data of the pump head using three-dimensional modeling software; the constructed three-dimensional model of the pump head is imported into the finite element analysis software for simulation calculation, and then n performance indicators of the optimized pump head are obtained.

[0105] Methods for evaluating whether the n performance indicators of the optimized pump head reach the preset performance indicator thresholds include:

[0106] Compare the obtained n performance indicators of the optimized pump head with the preset performance indicator thresholds:

[0107] If the obtained n performance indicators of the optimized pump head are less than the preset performance indicator threshold, it is determined that the n performance indicators of the optimized pump head do not reach the preset performance indicator threshold; analyze which performance indicators of the n performance indicators of the optimized pump head do not reach the preset performance indicator threshold, record the difference and the deviation direction, and then obtain the substandard performance indicator data; perform parameter sensitivity analysis on the substandard performance indicator data, adjust the pump head structural parameters according to the results of the parameter sensitivity analysis, obtain a new structural parameter optimization scheme, and perform simulation test on the new structural parameter optimization scheme; repeat the above steps until the n performance indicators of the optimized pump head reach the preset performance indicator threshold and stop the iteration.

[0108] If the obtained n performance indicators of the optimized pump head are greater than or equal to the preset performance indicator threshold, it is determined that the n performance indicators of the optimized pump head reach the preset performance indicator threshold;

[0109] The preset performance index threshold is set by the staff. The performance indexes of different pump heads are collected through the intelligent optimization management terminal, and the average value of the performance indexes of multiple pump heads is taken as the preset performance index threshold.

[0110] This embodiment automatically extracts meaningful features from data through dictionary learning without the need for manually designed features; dictionary learning can adaptively adjust the dictionary according to the structure and characteristics of the data to capture potential patterns in the data; sparse coding allows each data sample to be represented by a small number of dictionary atoms (dictionary elements); by introducing sparsity regularization, the method makes the representation of each data sample more compact and effective, thereby improving computational efficiency; by dynamically adjusting the sparsity regularization coefficient, the method can adaptively adjust the dictionary learning process at different time steps; dynamic adjustment of the regularization coefficient helps to balance the trade-off between dictionary reconstruction error and sparsity, thereby improving the stability and robustness of the model;

[0111] It can optimize multiple pump head performance indicators at the same time, solving the problem that traditional optimization methods only focus on a single target; through technologies such as non-dominated sorting and crowding, the algorithm can find a balance between multiple targets; generate an optimal solution set to obtain the best structural parameter optimization solution; use Latin hypercube sampling to generate the initial population, which can evenly cover the structural parameter space and avoid the problem of uneven parameter distribution that may be caused by traditional random sampling; through uniform sampling, the algorithm can explore the solution space more comprehensively, thereby improving the efficiency and accuracy of optimization; through non-dominated sorting and crowding calculation, this method can effectively screen out the optimal solution set, and avoid solutions from concentrating in the same position by calculating nonlinear crowding, thereby maintaining population diversity; by dynamically adjusting the adjustment parameters of nonlinear crowding, the algorithm can flexibly adjust parameters according to the actual situation of the population; it helps to avoid premature convergence or insufficient search in the multi-objective optimization process, and ensure better optimization results.

[0112] Example 2

[0113] See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, and a pump head structural parameter intelligent optimization design system is provided, including:

[0114] A data collection module, used to collect pump head structural parameter data, manufacturing cost data and user behavior data;

[0115] A data processing module is used to pre-process the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain a structural parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set; and to associate and fuse the structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set;

[0116] A performance index evaluation module is used to obtain a performance index prediction model based on comprehensive feature data set training, and to predict n performance indexes of the pump head through the performance index prediction model;

[0117] Intelligent optimization module, used to generate corresponding structural parameter optimization schemes based on n performance indicators of the pump head using an improved multi-objective optimization algorithm;

[0118] The effect feedback module is used to simulate and test the pump head that implements the corresponding structural parameter optimization scheme to obtain n performance indicators of the optimized pump head; evaluate whether the n performance indicators of the optimized pump head reach the preset performance indicator threshold, and use the structural parameter optimization scheme corresponding to the preset performance indicator threshold as the final pump head structural parameter intelligent optimization design scheme.

[0119] Since the electronic device introduced in this embodiment is an electronic device used to implement a pump head structural parameter intelligent optimization design method in the embodiment of the present application, based on the pump head structural parameter intelligent optimization design method introduced in the embodiment of the present application, the technical personnel of the field can understand the specific implementation of the electronic device of the present embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present application is not described in detail here. As long as the technical personnel of the field implement the electronic device used in the pump head structural parameter intelligent optimization design method in the embodiment of the present application, it belongs to the scope of protection of this application.

[0120] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0121] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for intelligent optimization design of pump head structural parameters, characterized in that: include: S1, collecting pump head structural parameter data, manufacturing cost data and user behavior data; S2, preprocessing the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain a structural parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set; correlating and fusing the structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set; S3, training a performance indicator prediction model based on the comprehensive feature data set, and predicting n performance indicators of the pump head through the performance indicator prediction model; S4, using an improved multi-objective optimization algorithm to generate a corresponding structural parameter optimization scheme according to n performance indicators of the pump head; S5. Perform simulation tests on the pump head that implements the corresponding structural parameter optimization scheme to obtain n performance indicators of the optimized pump head; evaluate whether the n performance indicators of the optimized pump head reach the preset performance indicator thresholds, and use the structural parameter optimization scheme corresponding to the preset performance indicator threshold as the final intelligent optimization design scheme for the pump head structural parameters.

2. The intelligent optimization design method for pump head structural parameters according to claim 1 is characterized in that: The pump head structural parameter data includes the length of the pressing rod, the spring stiffness, the pump chamber volume, the outlet aperture, the pump head material type, the pump head size and the nozzle size; the manufacturing cost data includes the material cost and the processing cost; the user behavior data includes the pressing frequency and the pressing duration.

3. The intelligent optimization design method for pump head structural parameters according to claim 2 is characterized in that: The method of preprocessing the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain a structural parameter feature data set, a manufacturing cost feature data set and a user behavior feature data set includes: Linear interpolation method is used to fill missing values ​​in the collected pump head structure parameter data, manufacturing cost data and user behavior data; density clustering algorithm is used to identify and eliminate outliers in the pump head structure parameter data, manufacturing cost data and user behavior data; By improving the online dictionary learning algorithm, feature extraction is performed on the collected pump head structural parameter data, manufacturing cost data and user behavior data to obtain structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets; the structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets are standardized and normalized to obtain normalized structural parameter feature data sets, manufacturing cost feature data sets and user behavior feature data sets.

4. The intelligent optimization design method for pump head structural parameters according to claim 3 is characterized in that: The method for extracting features from the collected pump head structural parameter data, manufacturing cost data and user behavior data by improving the online dictionary learning algorithm includes: S41. Initialize dictionary D∈R d′×k ; Where R is a real number set, that is, each element in the dictionary is a real number; d′ is the number of characteristic types of pump head structural parameter data, manufacturing cost data, and user behavior data; k is the number of dictionary atoms; S42. Input the current time step data: Among them, a t is the input data of the current time step; It is the pump head structural parameter data; is manufacturing cost data; is the user behavior data; T is the transpose of the input data matrix at the current time step; t is the current time step; S43, solving the current input data a by optimizing the objective function t The sparse coding coefficients under the dictionary of the previous time step, the optimization objective function is: in, To use the dictionary D t-1 β Reconstruction t The error; λ‖β‖1 is the sparsity regularization term; λ is the sparsity regularization coefficient, which controls the sparsity of sparse coding; β t is the sparse coding coefficient, which means the current input data a t Sparse representation under the dictionary of the previous time step; β is the sparse coefficient vector; argmin β is the optimization target; t-1 is the previous time step; S44. Dynamically adjust the sparsity regularization coefficient through a sparsity regularization adjustment formula; the sparsity regularization adjustment formula is: Among them, λ′ is the sparsity regularization coefficient after dynamic adjustment; η is the control dictionary D t-1 β Reconstruction t is the adjustment factor of the error on λ′; γ is the adjustment factor of the control time step on λ′; S45, use iterative soft threshold algorithm to optimize β t , until the preset maximum number of iterations is reached to minimize the reconstruction error; by optimizing β t , so that the input data a of the current time step t The dictionary D that can be in the previous time step t-1 Reconstructed by sparse coefficients; adjust and update the dictionary at each time step, and finally obtain the sparse coding coefficient β t That is, it is the feature representation of the data at the current time step. All sparse coefficient vectors are combined into a feature data set to obtain the pump head structure parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set.

5. The intelligent optimization design method for pump head structural parameters according to claim 4 is characterized in that: The method of associating and fusing the structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set to obtain a comprehensive feature data set includes: The structural parameter feature data set, the manufacturing cost feature data set and the user behavior feature data set are spliced ​​according to samples to obtain a multidimensional feature data set; the K-means clustering algorithm is applied to perform cluster analysis on the spliced ​​multidimensional feature data set, and the preset cluster number K is selected, and the data is divided into K clusters using the K-means algorithm; according to the clustering results, a cluster label is assigned to each sample, and the label is the cluster number to which the sample belongs, and the comprehensive characteristics of each cluster are represented by the cluster center, and the cluster center is the mean of the sample characteristics in the cluster; according to the clustering results and the cluster center, a comprehensive feature data set after association and fusion is finally formed.

6. The method for intelligent optimization design of pump head structural parameters according to claim 5, characterized in that: The training method of the performance indicator prediction model includes: The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample contains a historical comprehensive feature data set and n performance indicators of the corresponding pump head; Construct a performance indicator prediction model, which includes an input layer, an encoding layer, a multimodal fusion layer, a decoding layer and an output layer; the input layer is used to input a historical comprehensive feature data set, and the output layer is used to output n performance indicators of the pump head; the output layer is provided with neurons equal to the number of performance indicator types of the pump head, and each neuron corresponds to the prediction probability of a performance indicator of the pump head; a softmax function is used as an activation function; the performance indicator prediction model is a multimodal Transformer model; Use multi-classification cross entropy as the loss function of the model to measure the difference between the model's predicted value and the actual value; use the training set to train the performance indicator prediction model, and update the model parameters through the back propagation algorithm to minimize the loss function; use the mini-batch gradient descent method to update the model once per batch; use the validation set to evaluate the performance of the performance indicator prediction model by calculating the recall rate indicator; The SGD optimization algorithm is selected as the optimizer, and the model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance no longer improves or the preset stopping condition is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained performance indicator prediction model is used to predict the current comprehensive feature data set to obtain n performance indicators of the pump head.

7. The method for intelligent optimization design of pump head structural parameters according to claim 6, characterized in that: The n performance indicators of the pump head include flow consistency, pressure resistance, sealing, durability, corrosion resistance, pressing comfort and material utilization.

8. The method for intelligent optimization design of pump head structural parameters according to claim 7, characterized in that: The method of generating a corresponding structural parameter optimization scheme using an improved multi-objective optimization algorithm according to n performance indicators of the pump head includes: The initial population is generated by Latin hypercube sampling. The improved NSGA-II algorithm is used to perform non-dominated sorting on all individuals in the initial population and calculate the non-linear crowding degree. The parent individuals are selected from all individuals in the initial population based on the non-dominated level and crowding degree. The parent individuals are crossovered and mutated to generate the offspring individuals. The parent individuals and the offspring individuals are merged, and the next generation population is selected based on the non-dominated sorting results and non-linear crowding degree. Iteration is continued until the preset maximum number of iterations is reached, and the optimal solution set is obtained. The comprehensive scores of all solutions in the optimal solution set are calculated by weighted fusion, and the solution with the largest comprehensive score is selected as the final generated structural parameter optimization solution. The specific steps are as follows: S81. The preset pump head structural parameter data includes m types of structural parameters, and the structural parameter space is defined as: X={X1,X2,...,X i ,...,X m }; where X i is the i-th structural parameter; X m is the mth structural parameter; X i The value range is [L i ,U i ]; L i and U i They are the structural parameters X i The minimum and maximum values ​​of ; i is the index of the structural parameter type; S82. For each structural parameter X i , the value range [L i ,U i ] is divided into N intervals; the width of each interval is: Where ΔX i is the width of the interval; randomly sample a sample point in each of the N intervals and arrange them in a random order to obtain N arranged sampling points; Combine the permuted sampling points of all structural parameters to generate the initial population P init = {p1,p2,...,p α ,...,p N }; where p N is the Nth individual in the population; p α is the αth individual in the population, p α ={X α1 ,X α2 ,...,X αm };X α1 is the αth permutation sampling point of the first structural parameter; X α2 is the αth permutation sampling point of the second structural parameter; X αm is the αth permutation sampling point of the mth structural parameter; the initial population is represented by a matrix: S83. Classify each individual in the initial population according to the non-dominated sorting, and arrange the non-dominated sorting in descending order to obtain θ non-dominated levels: F1, F2, ..., F θ ; Among them, F1 is the optimal solution set in the initial population; F2 is the suboptimal solution set in the initial population; F θ is the θth optimal solution set in the initial population; θ is the index of the non-dominated level number; S84. Calculate the nonlinear crowding value of each individual in the initial population. The specific formula is: Among them, ρ(p α ) is the population individual p α The nonlinear crowding value of α+1 and d α-1 are the individuals p in the current population respectively. α Its adjacent solution p α+1 and p α-1 The Euclidean distance between them; σ is the adjustment parameter for controlling nonlinear crowding; exp() is an exponential function with e as the base; The adjustment parameters for controlling nonlinear congestion are adjusted by adjusting the parameter limiting formula; the adjustment parameter limiting formula is: Among them, σ′ is the adjustment parameter for controlling the nonlinear crowding degree after restriction; Q is the population size; E is the number of non-dominated levels; C1 is a constant for controlling the influence of the population size on the adjustment parameter, C2 is an offset constant; C3 is a constant for controlling the influence of the number of non-dominated levels on the adjustment parameter; S85, performing a selection operation according to the non-dominated sorting result and the non-linear crowding value of each individual in the initial population until N′ parent individuals are selected; arbitrarily selecting two parent individuals from the selected N′ parent individuals to perform a crossover operation to generate two new crossover individuals; collecting all the new crossover individuals generated by performing a crossover operation on all pairs of parent individuals to obtain crossover offspring individuals; Perform mutation operation on the selected N' parent individuals, use polynomial mutation to perturb the genes of N' parent individuals to generate new mutation individuals. The specific formula is: z ′=p z +δ·(U′-L′); where p z ′ is the generated new individual; p z is the zth parent individual among N′ parent individuals; δ is the disturbance factor of the mutation operation; U′ is the upper limit of the mutation operation; L′ is the lower limit of the mutation operation; collect the mutated new individuals of all parent individuals to obtain the mutated offspring individuals; S86, merge the crossover offspring individuals and mutation offspring individuals obtained through the crossover operation and mutation operation with the N′ parent individuals to obtain a new population P new ; S87, repeating steps S83-S86 until the preset maximum number of iterations is reached; when the iteration stops, calculating the optimal solution set of the current population; calculating the comprehensive scores of all solutions in the optimal solution set by weighted fusion, and selecting the solution with the largest comprehensive score as the final generated structural parameter optimization solution.

9. The method for intelligent optimization design of pump head structural parameters according to claim 8, characterized in that: The method of performing simulation test on the pump head implementing the corresponding structural parameter optimization scheme to obtain n performance indicators of the optimized pump head includes: According to the structural parameter optimization plan, the structural parameter data of the pump head are adjusted to obtain the optimized structural parameter data of the pump head; a three-dimensional model of the pump head is constructed based on the optimized structural parameter data of the pump head using three-dimensional modeling software; the constructed three-dimensional model of the pump head is imported into the finite element analysis software for simulation calculation, and then n performance indicators of the optimized pump head are obtained.

10. The method for intelligent optimization design of pump head structural parameters according to claim 9, characterized in that: The method for evaluating whether the n performance indicators of the optimized pump head reach the preset performance indicator thresholds includes: Compare the obtained n performance indicators of the optimized pump head with the preset performance indicator thresholds: If the obtained n performance indicators of the optimized pump head are less than the preset performance indicator threshold, it is determined that the n performance indicators of the optimized pump head do not reach the preset performance indicator threshold; If the obtained n performance indicators of the optimized pump head are greater than or equal to the preset performance indicator threshold, it is determined that the n performance indicators of the optimized pump head reach the preset performance indicator threshold.

Images