Plunger pump oil film field calculation method and system based on deep onet

Abstract

The present application relates to the field of fluid pressure and the field of electric digital data processing, and in particular to a kind of plunger pump oil film field calculation method and system based on DeepONet, comprising: obtaining training data set;DeepONet-based oil film thickness field-oil film pressure field calculation model is constructed;The training data set is used to train the oil film thickness field-oil film pressure field calculation model;Based on the oil film thickness field-oil film pressure field calculation model after training, the pressure value at the random generation space coordinates is calculated.The present application embodiment improves the oil film field calculation resolution and calculation accuracy by constructing the oil film thickness field-pressure field DeepONet, while the calculation speed is considered.

Description

Technical Field

[0001] This invention relates to the fields of fluid pressure and electronic digital data processing, and specifically to a method and system for calculating the oil film field of a plunger pump based on DeepONet. Background Technology

[0002] When the key friction pairs (distribution pair, piston pair, and slipper pair) of an axial piston pump operate in high-pressure fluid media, their oil film characteristics have a decisive impact on the pump's energy efficiency and service life. The oil film formed between these friction pair components simultaneously performs the triple functions of support, sealing, and lubrication, and the coupling effect between the oil film thickness field and the pressure field directly determines the load-bearing capacity, leakage characteristics, and lubrication condition of the friction pair. Therefore, achieving high-precision, high-resolution oil film pressure field solutions is crucial for in-depth analysis of the friction pair oil film characteristics and for guiding the optimal design and reliability assessment of piston pumps.

[0003] Traditional methods for calculating oil film pressure fields are based on numerical models established using the Reynolds equations, employing numerical calculation methods such as the finite difference method or the finite element method. However, the oil film field in an axial piston pump involves strong coupling effects between fluid, solid, and thermal multiple physics fields, posing a dual challenge to traditional oil film field calculation methods in terms of both computational speed and accuracy. These methods require discretizing the known oil film thickness field into a grid and solving for the pressure values ​​at the grid nodes through a complex iterative process, resulting in low solution efficiency. Although existing improved methods (such as the fluid film pressure calculation method based on Garlerkin's ideas, CN109815548A) simplify the process by optimizing basis functions and reducing the order of calculation, their accuracy and resolution are still limited by the grid density. To obtain a higher accuracy pressure field, the number of grids must be increased, which in turn leads to an exponential increase in computational scale.

[0004] In recent years, machine learning technology has provided new ideas for calculating oil film pressure fields. A study (Development of a Machine Learning Model for Elastohydrodynamic Pressure Prediction in Journal Bearings) has significantly improved the calculation speed by constructing and training a machine learning model to approximate the function mapping from oil film thickness value to oil film pressure value, but still has the following limitations: (1) The resolution of the oil film field is related to the output layer neurons of the machine learning model, requiring a large number of output layer neurons to achieve high-resolution oil film field calculation; (2) Learning the mapping relationship by fitting function values ​​to function values ​​has limited generalization ability, and when the input oil film thickness value exceeds the range of the training dataset, the corresponding oil film field calculation accuracy will decrease significantly; (3) Pure data-driven black box models are difficult to satisfy the Reynolds equation constraints, which limits the improvement of oil film field calculation accuracy.

[0005] Overall, existing methods still have significant shortcomings in terms of computational efficiency, accuracy, and generalization ability, and a new solution is urgently needed to overcome these technical bottlenecks. Summary of the Invention

[0006] One of the objectives of this invention is to provide a method and system for calculating the oil film field of a plunger pump based on DeepONet, so as to solve the technical problems of low efficiency, low accuracy and poor generalization ability in the calculation of oil film pressure field in the prior art.

[0007] To achieve the above objectives, this invention provides a method for calculating the oil film field of a plunger pump based on DeepONet, comprising:

[0008] Obtain the training dataset;

[0009] Construct a calculation model for oil film thickness field and oil film pressure field based on DeepONet;

[0010] The oil film thickness field-oil film pressure field calculation model is trained using the training dataset, wherein the loss function of the oil film thickness field-oil film pressure field calculation model includes formulas (1) to (4):

[0011] (1)

[0012] (2)

[0013] (3)

[0014] (4)

[0015] in, For loss function, These are constraint terms in the Reynolds equation. For data constraints, These are boundary condition constraint terms. The total number of samples in the training dataset. The number of pressure field coordinate points in each sample. These are spatial coordinates in the polar coordinate system. For oil film thickness, The kinematic viscosity of the hydraulic oil. For the calculation model of oil film thickness field-oil film pressure field, the first Each sample is in coordinates Predicted pressure value at the location, Let be the tangential velocity of the shoe's sole relative to the swashplate. The radial flow velocity caused by changes in oil film thickness. For the first Each sample is located at coordinates Reference solution at that location, This represents the boundary pressure value. For the first Each sample is located at the boundary. The pressure solution at the location;

[0016] The pressure values ​​at randomly generated spatial coordinates are calculated based on the trained oil film thickness field-oil film pressure field calculation model.

[0017] Optionally, obtaining the training dataset includes:

[0018] Establish the polar coordinates of the oil film thickness field space;

[0019] Construct an oil film thickness field model;

[0020] Construct the Reynolds equation for the piston pump slipper pair and solve for the reference solution of the pressure field;

[0021] Discretize the oil film thickness field to obtain oil film thickness field sampling point data;

[0022] Discretize the oil film pressure field to obtain oil film pressure field sampling point data;

[0023] A training dataset is obtained based on the oil film thickness field sampling point data and the oil film pressure field sampling point data.

[0024] Optionally, establishing the polar coordinates of the oil film thickness field space includes:

[0025] The origin is the projection of the center of the skate's sole onto the swashplate. Establishment on the contact surface between the oil film and the swashplate shaft and The shaft is established in the normal direction of the contact surface between the oil film and the swashplate. axis.

[0026] Optionally, constructing an oil film thickness field model includes:

[0027] An oil film thickness model is established based on formula (5).

[0028] (5)

[0029] in, For the oil film thickness field, Based on the base oil film thickness, For the center oil film thickness, For the angle of the ski boot overturning, For the azimuth angle of the ski boot overturning, Radial coordinates, For angular coordinates, Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the bottom surface of the slipper.

[0030] Optionally, the Reynolds equations for the plunger pump's slipper pair are constructed, and the reference solution for the pressure field is solved, including:

[0031] The Reynolds equation for the plunger pump slipper pair is established based on formula (6).

[0032] (6)

[0033] in, For oil film pressure field;

[0034] Input the given oil film thickness field into the Reynolds equation to obtain the corresponding pressure field data, which serves as the reference solution for the oil film thickness field-oil film pressure field.

[0035] Optionally, discretizing the oil film thickness field to obtain oil film thickness field sampling point data includes:

[0036] Discretize the oil film thickness field according to formula (7) and obtain the corresponding... ,

[0037] (7)

[0038] in, For fixed sampling points, The number of sampling points. Radial coordinates, For angular coordinates, Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the bottom surface of the slipper.

[0039] Optionally, discretizing the oil film pressure field to obtain oil film pressure field sampling point data includes:

[0040] Discretize the oil film pressure field according to formula (8) and obtain the corresponding... ,

[0041] (8)

[0042] in, For random sampling points, It is a uniform probability distribution. It is a uniform probability distribution. Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the bottom surface of the slipper.

[0043] Optionally, constructing a DeepONet-based oil film thickness field-oil film pressure field calculation model includes:

[0044] A branch network and a backbone network are constructed. The branch network adopts a cascaded structure of an attention mechanism and a multilayer perceptron. The attention mechanism is used to capture global features, and the multilayer perceptron is used to process local features. The branch network is used to input the oil film thickness value and output the global feature vector. The backbone network is a multilayer perceptron structure, which is used to input spatial coordinates and operating conditions and calculate the position feature vector.

[0045] The pressure value is calculated based on the global feature vector and the location feature vector.

[0046] Optionally, training the oil film thickness field-oil film pressure field calculation model using the training dataset includes:

[0047] The training dataset is input into the oil film thickness field-oil film pressure field calculation model, and the predicted pressure value is output.

[0048] Calculate the loss function value;

[0049] Determine whether the loss function value is greater than the threshold;

[0050] If the loss function value is greater than the threshold, the parameters of the oil film thickness field-oil film pressure field calculation model are updated, and the pressure field is recalculated.

[0051] Training is completed when the loss function value is determined to be less than or equal to the threshold.

[0052] On the other hand, the present invention provides a plunger pump oil film field calculation system based on DeepONet, the system including a processor configured to perform any of the methods described above.

[0053] The beneficial effects of this invention are:

[0054] This invention constructs an oil film thickness field-pressure field DeepONet, avoiding the process of repeatedly solving partial differential equations. After training, the DeepONet model only requires one forward propagation to quickly calculate the pressure field. Compared with existing technologies, it avoids complex mesh generation and iterative calculations, resulting in faster pressure field calculation.

[0055] This invention directly learns the global functional mapping relationship from the oil film thickness field to the oil film pressure field by constructing DeepONet. The branch network is a cascade structure of attention mechanism and MLP. The oil film thickness field is discretized and global features are extracted. The backbone network encodes spatial coordinates and operating conditions, outputting position features. The dot product of global features and position features is summed to calculate the pressure value at any position in the domain of the pressure field. It has strong generalization ability and can maintain high solution accuracy even on data not used for training. At the same time, the loss function proposed in this invention includes Reynolds equation constraint terms, data constraint terms, and boundary condition constraint terms, which fully considers the consistency of the physical information of the oil film field and the boundary conditions, further improving the solution accuracy.

[0056] The DeepONet oil film thickness-pressure field constructed in this invention decouples the oil film thickness value from the spatial coordinates through branch networks and a backbone network. By simply changing the spatial coordinates input to the backbone network, the oil film pressure value at any location can be calculated. Therefore, the resolution of the oil film pressure field is only related to the number of spatial coordinates input to the backbone network. By increasing the number of spatial coordinates, high-resolution oil film pressure field calculations can be achieved quickly without retraining the network.

[0057] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0058] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0059] Figure 1 This is a flowchart of a plunger pump oil film field calculation method based on DeepONet according to an embodiment of the present invention;

[0060] Figure 2 A flowchart illustrating a method for obtaining a training dataset according to an embodiment of the present invention;

[0061] Figure 3 This is a schematic diagram illustrating the establishment of a spatial polar coordinate system according to an embodiment of the present invention;

[0062] Figure 4 A flowchart illustrating a method for constructing an oil film thickness field-oil film pressure field calculation model based on DeepONet according to an embodiment of the present invention;

[0063] Figure 5 A flowchart of a method for training an oil film thickness field-oil film pressure field calculation model according to an embodiment of the present invention;

[0064] Figure 6 This is a framework diagram of an oil film thickness field-oil film pressure field calculation model based on DeepONet according to an embodiment of the present invention.

[0065] Figure 7 Example 1 of an oil film field calculated by a DeepONet-based oil film thickness field-oil film pressure field calculation model according to an embodiment of the present invention;

[0066] Figure 8 Example 2 of the oil film field calculated by the DeepONet-based oil film thickness field-oil film pressure field calculation model according to an embodiment of the present invention. Detailed Implementation

[0067] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0068] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0069] like Figure 1 The diagram shows a flowchart of a plunger pump oil film field calculation method based on DeepONet according to an embodiment of the present invention. Figure 1 In this calculation method, the steps may include:

[0070] In step S10, the training dataset is obtained;

[0071] In step S11, a calculation model for the oil film thickness field and oil film pressure field based on DeepONet is constructed.

[0072] In step S12, the oil film thickness field-oil film pressure field calculation model is trained using the training dataset. The loss function of the oil film thickness field-oil film pressure field calculation model includes formulas (1) to (4):

[0073] (1)

[0074] (2)

[0075] (3)

[0076] (4)

[0077] in, For loss function, These are constraint terms in the Reynolds equation. For data constraints, These are boundary condition constraint terms. The total number of samples in the training dataset. The number of pressure field coordinate points in each sample. These are spatial coordinates in the polar coordinate system. For oil film thickness, The kinematic viscosity of the hydraulic oil. For the calculation model of oil film thickness field-oil film pressure field, the first Each sample is located at coordinates Predicted pressure value at the location, Let be the tangential velocity of the shoe's sole relative to the swashplate. The radial flow velocity caused by changes in oil film thickness. For the first Each sample is located at coordinates Reference solution at that location, This represents the boundary pressure value. For the first Each sample is located at the boundary. The pressure solution at the location;

[0078] In step S13, the pressure value at the randomly generated spatial coordinates is calculated based on the trained oil film thickness field-oil film pressure field calculation model.

[0079] In such Figure 1 In the DeepONet-based plunger pump oil film field calculation method shown, step S10 is used to obtain a training dataset. In this embodiment, the specific method for obtaining the training dataset in step S10 can be of various forms known to those skilled in the art. In one example of the present invention, step S10 may include, for example... Figure 2 The steps shown are described in this. Figure 2 In this context, step S10 may include:

[0080] In step S20, the polar coordinates of the oil film thickness field are established;

[0081] In step S21, an oil film thickness field model is constructed;

[0082] In step S22, the Reynolds equation for the piston pump slipper pair is constructed, and the reference solution for the pressure field is solved.

[0083] In step S23, the oil film thickness field is discretized and sampled to obtain oil film thickness field sampling point data;

[0084] In step S24, the oil film pressure field is discretized and sampled to obtain oil film pressure field sampling point data;

[0085] In step S25, a training dataset is obtained based on the oil film thickness field sampling point data and the oil film pressure field sampling point data.

[0086] In such Figure 2 In the method shown, step S20 is used to establish the spatial polar coordinates of the oil film thickness field. Since the oil film of the piston pump friction pair is annular, establishing a spatial rectangular coordinate system is inconvenient for subsequent calculations. Therefore, in this embodiment, a spatial polar coordinate system can be established. Specifically, as shown... Figure 3 As shown, the origin is the projection of the center of the skate's bottom surface onto the swashplate. Establishment on the contact surface between the oil film and the swashplate shaft and The shaft is established in the normal direction of the contact surface between the oil film and the swashplate. Shaft, oil film thickness field application This indicates that the oil film pressure field is used express.

[0087] The structural parameters and operating conditions of the plunger pump can be specified according to actual needs. Specifically, in this example, the structural parameters of the plunger pump can be given as follows: slipper chamber radius. Radius of the outer circle of the bottom surface of the slipper , radius of plunger distribution circle Given hydraulic oil performance parameters, including the kinematic viscosity of the hydraulic oil... Given the operating conditions of the piston pump: given spindle speed and load pressure As the solution condition, the pump casing pressure is given. As boundary conditions, in this example, it can be one of two operating conditions: a spindle speed of 1000 rpm, a load pressure of 9 MPa, and a pump housing pressure of 0.1 MPa; or a spindle speed of 5000 rpm, a load pressure of 21 MPa, and a pump housing pressure of 0.1 MPa.

[0088] Step S21 is used to construct an oil film thickness field model. In this embodiment, the specific method for constructing the oil film thickness field model in step S21 can be of various forms known to those skilled in the art. In one example of the present invention, step S21 may include:

[0089] An oil film thickness model is established based on formula (5).

[0090] (5)

[0091] in, For oil film thickness field, Based on the base oil film thickness, For the center oil film thickness, For the angle of the ski boot overturning, For the azimuth angle of the ski boot overturning, Radial coordinates, For angular coordinates, Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the shoe's bottom surface. In this example, the oil film thickness at the center point is... The range is set to {5, 6, 7, 8}. Slipper overturning angle Set to {0.001, 0.0015, 0.002}°, slipper overturning azimuth angle. Set to {240, 245, 250}°.

[0092] Step S22 is used to calculate the corresponding pressure field as a reference solution based on the Reynolds equation for the friction pair of the plunger pump. In this embodiment, the specific method for solving the pressure field reference solution in step S22 can be of various forms known to those skilled in the art. In one example of the present invention, step S22 may include:

[0093] The Reynolds equation for the plunger pump slipper pair is established based on formula (6).

[0094] (6)

[0095] in, For oil film pressure field;

[0096] Input the given oil film thickness field into the Reynolds equation to obtain the corresponding pressure field data, which serves as the reference solution for the oil film thickness field-oil film pressure field.

[0097] Specifically, in this embodiment, it can be a given specific oil film thickness field Input the Reynolds equation and use tools such as MATLAB or Pumplinx to calculate the corresponding pressure field. As a reference solution for the oil film thickness-pressure field, the number of reference solutions is determined by the total number of permutations of the following parameters: oil film thickness at the center point, slipper overturning angle, slipper overturning azimuth angle, and operating conditions. In this example, there are a total of 72 (4×3×3×2=72) oil film thickness-pressure field reference solutions. The obtained oil film thickness-pressure field reference solutions are divided into... Group training reference solution and Group verification reference solution In this example, It is 60. It is 12.

[0098] Step S23 is used to discretize and sample the oil film thickness field to obtain oil film thickness field sampling point data. In this embodiment, the specific method for obtaining oil film thickness field sampling point data in step S23 can be of various forms known to those skilled in the art. In one example of the present invention, step S23 may include:

[0099] Discretize the oil film thickness field according to formula (7) and obtain the corresponding... ,

[0100] (7)

[0101] in, For fixed sampling points, The number of sampling points. Radial coordinates, For angular coordinates, Where is the radius of the slipper oil chamber. Let be the outer radius of the bottom surface of the slipper. All oil film thickness fields. and Sharing the same sampling point locations ensures consistency in the input dimensions. In this example, It can be 256.

[0102] Step S24 is used to discretize and sample the oil film pressure field to obtain oil film pressure field sampling point data. In this embodiment, the specific method for obtaining oil film pressure field sampling point data in step S24 can be of various forms known to those skilled in the art. In one example of the present invention, step S24 may include:

[0103] Discretize the oil film pressure field according to formula (8) and obtain the corresponding... ,

[0104] (8)

[0105] in, For random sampling points, It is a uniform probability distribution. It is a uniform probability distribution. Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the bottom surface of the slipper.

[0106] Through probability distribution Randomly generated spatial coordinates For any pressure field and Spatial coordinates They are not necessarily the same. In this example, It can be 100.

[0107] Step S25 obtains the training dataset based on the oil film thickness field sampling point data and the oil film pressure field sampling point data. Specifically, in this example, the training dataset... for:

[0108] ;

[0109] Validation dataset for:

[0110] ;

[0111] The number of samples in the training dataset is... and The product of , the number of samples in the validation dataset is and The product of . In this example, the training dataset has 6000 samples and the validation dataset has 1200 samples.

[0112] Step S11 is used to construct an oil film thickness field-oil film pressure field calculation model based on DeepONet. In this embodiment, the specific method for constructing the oil film thickness field-oil film pressure field calculation model in step S11 can be of various forms known to those skilled in the art. In one example of the present invention, step S11 may include, for example... Figure 4 The steps shown are described. Figure 4 In this context, step S11 may include:

[0113] In step S30, the branch network and the backbone network are constructed.

[0114] In step S31, the global feature vector and the positional feature vector are obtained;

[0115] In step S32, the pressure value at the specified location of the oil film field is calculated.

[0116] In such Figure 4 In the method shown, step S30 is used to construct the branch network and the backbone network. Specifically, in this embodiment, the constructed oil film thickness field-pressure field DeepONet is used... This indicates that it is composed of branch networks. and backbone network The DeepONet architecture is composed of multiple layers of perceptrons, designed to approximate the global functional mapping from the oil film thickness field to the oil film pressure field. Initially, both the branch and backbone networks of DeepONet were composed of multiple layers of perceptrons, which could only capture local features and were unable to model global features such as radial gradient pressure, axial symmetry, and boundary conditions of the oil film field. Therefore, [the following is a more detailed explanation of DeepONet's architecture]. Figure 6As shown, the DeepONet branch network in this invention employs a cascaded structure of an attention mechanism and a multilayer perceptron. The attention mechanism captures global features, while the multilayer perceptron processes local features. The backbone network retains the multilayer perceptron structure to encode spatial coordinates and operational conditions. Both the branch network and the backbone network output F-dimensional feature vectors. In this example, F can be 16, representing the key matrix of the attention mechanism in the branch network. Query matrix Sum matrix All are 256×256 matrices. The multilayer perceptron with branch networks includes an input layer. Hidden layer Hidden layer and output layer The number of neurons were 256, 64, 16, and 16, respectively; the MLP backbone network contained an input layer. Hidden layer Hidden layer and output layer The number of neurons were 4, 8, 16, and 16, respectively.

[0117] Step S31 involves the branch network taking the oil film thickness value at a fixed sampling point as input and outputting a global feature vector, while the backbone network takes randomly generated spatial coordinates and operating conditions as input and calculates a position feature vector. Specifically, the input to the branch network can be the oil film thickness value at a fixed sampling point. The output is an F-dimensional global feature vector. :

[0118] ;

[0119] The input to the backbone network is randomly generated spatial coordinates. The output is an F-dimensional positional feature vector. :

[0120] .

[0121] Step S32 is used to calculate the pressure value at a specified location in the oil film field. Specifically, in this example, it can be achieved by multiplying the global feature vector output by the branch network with the location feature vector output by the backbone network to calculate the spatial coordinates. Pressure value at the location :

[0122] ;

[0123] Input oil film thickness value for fixed branch network traversal Then input the data into the backbone network sequentially to calculate the pressure value at the specified location in the oil film field. .

[0124] Step S12 is used to train the oil film thickness field-oil film pressure field calculation model. In this embodiment, the specific method for training the oil film thickness field-oil film pressure field calculation model in step S12 can be of various forms known to those skilled in the art. In one example of the present invention, step S12 may include, for example... Figure 5 The steps shown are described in this. Figure 5 In this context, step S12 may include:

[0125] In step S40, the training dataset is input into the oil film thickness field-oil film pressure field calculation model, and the predicted pressure value is output.

[0126] In step S41, the loss function value is calculated;

[0127] In step S42, it is determined whether the loss function value is greater than the threshold;

[0128] In step S43, if the loss function value is greater than the threshold, the parameters of the oil film thickness field-oil film pressure field calculation model are updated, and the pressure field is recalculated.

[0129] Training is completed when the loss function value is less than or equal to the threshold.

[0130] In such Figure 4 In the method shown, steps S40 to S43 are used to train the oil film thickness field-pressure field DeepONet. In this embodiment, the training and accuracy evaluation of the oil film thickness field-pressure field DeepONet can be performed in a Python environment. First, the parameters of the oil film thickness field-pressure field DeepONet are initialized, including the key matrix of the branch network. Query matrix Value matrix The weights of each neuron, the biases of each neuron in the branch network, the weights of each neuron in the backbone network, and the biases of each neuron in the backbone network are determined. Then, the threshold for the loss function is set to... In this embodiment, Set as Step S40 is used to obtain the predicted stress value. The training dataset is used... Input DeepONet to calculate the pressure field based on a given oil film thickness field. Step S41 is used to calculate the loss function value. Although DeepONet has the ability to learn functional mappings compared to machine learning models, it is essentially still a data-driven black box model, lacking the corresponding physical information of the oil film field and consistency of boundary conditions, which limits further improvement in solution accuracy. Therefore, this invention adds the Reynolds equation as a physical constraint to the training process of DeepONet. The pressure field calculated by DeepONet is compared with the reference solution, and the loss function value is calculated according to the following formula:

[0131] (1)

[0132] (2)

[0133] (3)

[0134] (4)

[0135] (9)

[0136] (10)

[0137] in, For loss function, These are constraint terms in the Reynolds equation. For data constraints, These are boundary condition constraint terms. The total number of samples in the training dataset. The number of pressure field coordinate points in each sample. These are spatial coordinates in the polar coordinate system. For oil film thickness, The kinematic viscosity of the hydraulic oil. For the calculation model of oil film thickness field-oil film pressure field, the first Each sample is located at coordinates Predicted pressure value at the location, Let be the tangential velocity of the shoe's sole relative to the swashplate. The radial flow velocity caused by changes in oil film thickness. For the first Each sample is located at coordinates Reference solution at that location, This represents the boundary pressure value. For the first Each sample is located at the boundary. The pressure solution at the location. Among them, and It can be obtained in Python through automatic differentiation.

[0138] Step S42 determines whether the loss function value is greater than a threshold. Step S43, if the loss function value is greater than the threshold, updates the parameters of the oil film thickness field-oil film pressure field calculation model using gradient descent and recalculates the pressure field. In this example, the gradient descent update algorithm can be Adam. If the loss function value is less than or equal to the threshold, training is complete.

[0139] Step S13 is used to calculate the pressure value at randomly generated spatial coordinates using the trained oil film thickness field-oil film pressure field calculation model.

[0140] This invention also evaluates the accuracy of DeepONet for the oil film thickness-pressure field. Specifically, this includes: using a validation dataset... Input the trained DeepONet and calculate the oil film field. Calculate the evaluation index according to formula (11). , The smaller the value, the higher the accuracy of DeepONet in calculating the oil film field.

[0141] (11)

[0142] In this example, the trained DeepONet calculates the oil film field under two operating conditions: a spindle speed of 1000 rpm and a load pressure of 9 MPa, and a spindle speed of 5000 rpm and a load pressure of 21 MPa. The calculations are then converted to a spatial rectangular coordinate system, as shown below. Figure 7 (Spindle speed 1000 rpm, load pressure 9 MPa) and Figure 8 (Spindle speed 5000 rpm, load pressure 21 MPa) As shown. DeepONet calculated the average time for a single oil film field to be 116.6 ms. Evaluation metrics This enables fast and accurate calculation of the oil film field in a plunger pump. Furthermore, after training DeepONet, by simply changing the spatial coordinates of the input backbone network based on the oil film thickness values ​​of the input branch network, the oil film pressure value at any location can be calculated. Therefore, the resolution of DeepONet's oil film field calculation is only related to the number of spatial coordinates in the input backbone network; simply increasing the number of spatial coordinates allows for fast calculation of high-precision, high-resolution oil film fields without requiring network retraining.

[0143] On the other hand, the present invention provides a plunger pump oil film field calculation system based on DeepONet, the system including a processor configured to perform any of the methods described above.

[0144] The beneficial effects of this invention are:

[0145] This invention constructs an oil film thickness field-pressure field DeepONet, avoiding the process of repeatedly solving partial differential equations. After training, the DeepONet model only requires one forward propagation to quickly calculate the pressure field. Compared with existing technologies, it avoids complex mesh generation and iterative calculations, resulting in faster pressure field calculation.

[0146] This invention directly learns the global functional mapping relationship from the oil film thickness field to the oil film pressure field by constructing DeepONet. The branch network is a cascade structure of attention mechanism and MLP. The oil film thickness field is discretized and global features are extracted. The backbone network encodes spatial coordinates and operating conditions, outputting position features. The dot product of global features and position features is summed to calculate the pressure value at any position in the domain of the pressure field. It has strong generalization ability and can maintain high solution accuracy even on data not used for training. At the same time, the loss function proposed in this invention includes Reynolds equation constraint terms, data constraint terms, and boundary condition constraint terms, which fully considers the consistency of the physical information of the oil film field and the boundary conditions, further improving the solution accuracy.

[0147] The DeepONet oil film thickness-pressure field constructed in this invention decouples the oil film thickness value from the spatial coordinates through branch networks and a backbone network. By simply changing the spatial coordinates input to the backbone network, the oil film pressure value at any location can be calculated. Therefore, the resolution of the oil film pressure field is only related to the number of spatial coordinates input to the backbone network. By increasing the number of spatial coordinates, high-resolution oil film pressure field calculations can be achieved quickly without retraining the network.

[0148] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0149] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0150] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0151] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0152] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0153] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0154] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0155] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0156] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for calculating the oil film field of a plunger pump based on DeepONet, characterized in that, The calculation method includes: Obtain the training dataset; Construct a calculation model for oil film thickness field and oil film pressure field based on DeepONet; The oil film thickness field-oil film pressure field calculation model is trained using the training dataset, wherein the loss function of the oil film thickness field-oil film pressure field calculation model includes formulas (1) to (4): ，（1） ，（2） ，（3） ，（4） in, For loss function, These are constraint terms in the Reynolds equation. For data constraints, These are boundary condition constraint terms. The total number of samples in the training dataset. The number of pressure field coordinate points in each sample. These are spatial coordinates in the polar coordinate system. For oil film thickness, The kinematic viscosity of the hydraulic oil. For the calculation model of oil film thickness field-oil film pressure field, the first Each sample is in coordinates Predicted pressure value at the location, Let be the tangential velocity of the shoe's sole relative to the swashplate. The radial flow velocity caused by the change in oil film thickness. For the first Each sample is in coordinates Reference solution at that location, This represents the boundary pressure value. For the first Each sample is located at the boundary. The pressure solution at the location; The pressure values ​​at randomly generated spatial coordinates are calculated based on the trained oil film thickness field-oil film pressure field calculation model.

2. The calculation method according to claim 1, characterized in that, Obtaining the training dataset includes: Establish the polar coordinates of the oil film thickness field space; Construct an oil film thickness field model; Construct the Reynolds equation for the piston pump slipper pair and solve for the reference solution of the pressure field; Discretize the oil film thickness field to obtain oil film thickness field sampling point data; Discretize the oil film pressure field to obtain oil film pressure field sampling point data; A training dataset is obtained based on the oil film thickness field sampling point data and the oil film pressure field sampling point data.

3. The calculation method according to claim 2, characterized in that, Establishing the polar coordinates of the oil film thickness field space includes: The origin is the projection of the center of the skate's sole onto the swashplate. Establishment on the contact surface between the oil film and the swashplate shaft and The shaft is established in the normal direction of the contact surface between the oil film and the swashplate. axis.

4. The calculation method according to claim 2, characterized in that, The construction of the oil film thickness field model includes: An oil film thickness model is established based on formula (5). ，（5） in, For oil film thickness field, Based on the base oil film thickness, For the center oil film thickness, For the angle of the ski boot overturning, For the azimuth angle of the ski boot overturning, Radial coordinates, For angular coordinates, Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the bottom surface of the slipper.

5. The calculation method according to claim 4, characterized in that, The Reynolds equations for the piston pump slipper pair are constructed, and the reference solution for the pressure field is solved, including: The Reynolds equation for the plunger pump slipper pair is established based on formula (6). ，（6） in, For oil film pressure field; Input the given oil film thickness field into the Reynolds equation to obtain the corresponding pressure field data, which serves as the reference solution for the oil film thickness field-oil film pressure field.

6. The calculation method according to claim 2, characterized in that, Discretizing the oil film thickness field to obtain oil film thickness field sampling point data includes: Discretize the oil film thickness field according to formula (7) and obtain the corresponding... , ，（7） in, For fixed sampling points, The number of sampling points. Radial coordinates, For angular coordinates, Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the bottom surface of the slipper.

7. The calculation method according to claim 2, characterized in that, Discretizing the oil film pressure field to obtain oil film pressure field sampling point data includes: Discretize the oil film pressure field according to formula (8) and obtain the corresponding... , ，（8） in, For random sampling points, It is a uniform probability distribution. It is a uniform probability distribution. Where is the radius of the slipper oil chamber. Let be the radius of the outer circle of the bottom surface of the slipper.

8. The calculation method according to claim 1, characterized in that, The construction of the oil film thickness field-oil film pressure field calculation model based on DeepONet includes: A branch network and a backbone network are constructed. The branch network adopts a cascaded structure of an attention mechanism and a multilayer perceptron. The attention mechanism is used to capture global features, and the multilayer perceptron is used to process local features. The branch network is used to input the oil film thickness value and output the global feature vector. The backbone network is a multilayer perceptron structure, which is used to input spatial coordinates and operating conditions and calculate the position feature vector. The pressure value is calculated based on the global feature vector and the location feature vector.

9. The calculation method according to claim 1, characterized in that, Training the oil film thickness field-oil film pressure field calculation model using the training dataset includes: The training dataset is input into the oil film thickness field-oil film pressure field calculation model, and the predicted pressure value is output. Calculate the loss function value; Determine whether the loss function value is greater than the threshold; If the loss function value is greater than the threshold, the parameters of the oil film thickness field-oil film pressure field calculation model are updated, and the pressure field is recalculated. Training is completed when the loss function value is determined to be less than or equal to the threshold.

10. A plunger pump oil film field calculation system based on DeepONet, characterized in that, The system includes a processor configured to perform the method as described in any one of claims 1 to 9.

Images