Optimization design method for spring-energized seal ring of single point mooring system liquid stuffing box

Abstract

The present application relates to the technical field of electric digital data processing, and especially relates to a single-point mooring system liquid slip ring spring energy storage sealing ring optimization design method, comprising the following steps: a two-dimensional axisymmetric model of the liquid slip ring spring energy storage sealing ring is established; the peak contact stress, the linear contact stress and the maximum shear stress of the liquid slip ring spring energy storage sealing ring under actual working conditions are obtained; a sealing performance evaluation system of the liquid slip ring spring energy storage sealing ring is established; key structure parameters are screened out to form a data set; a back propagation neural network model is constructed, and the weight and bias of the back propagation neural network model are trained; the trained back propagation neural network model is re-evaluated, and finally the optimal parameter combination and the finite element model based on the optimal parameter combination are obtained. The method provided by the present application can determine the optimal parameter combination that meets the sealing performance requirements, so as to improve the overall sealing effect and improve the reliability of the sealing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric digital data processing, and in particular to a single-point mooring system liquid slip ring spring energy storage sealing ring optimization design method. BACKGROUND

[0002] At present, the development of offshore oil fields in China generally adopts the production mode of combining drilling platforms, oil storage barges and single-point mooring systems. Among them, the single-point mooring system, as an important part of modern marine crude oil exploitation and transportation, is one of the key equipment in the marine oil and gas industry. In the single-point mooring system, the liquid rotary joint structure is one of the core technologies. Through the rotary joint, effective transportation of crude oil can be realized, and the ship can be automatically adjusted to the direction with the smallest force to stay according to the wind and wave conditions. The basic principle of the rotary joint relies on the relative rotation of the inner and outer slip rings, so as to realize the free movement of the ship with the wind and wave and the transportation of oil and gas. Therefore, the sealing design of the liquid rotary joint slip ring structure becomes a key point.

[0003] The liquid slip ring needs to be equipped with multiple sealing devices at the rotary joint to prevent leakage of crude oil. It often works in a high-temperature and high-pressure dynamic sealing process, and a single rubber material cannot meet the requirements of high-temperature and high-pressure dynamic sealing, so the spring energy storage sealing ring appears. The main material of the spring energy storage sealing ring is generally polytetrafluoroethylene, which has good resistance to high temperature, corrosion, wear resistance and lubricity, and can cope with harsh working environments. At present, the research on the spring energy storage sealing ring mainly focuses on materials, geometric structures and working conditions. Numerous studies have shown that the changes in the structural parameters of the spring energy storage sealing ring have a significant impact on the sealing performance. However, most of the research only focuses on the impact of single parameter changes on the sealing performance, and the evaluation criteria for sealing performance are not accurate, which cannot systematically describe the influence of the structural parameters of the spring energy storage sealing ring on the sealing performance, thereby reducing the reliability of the spring energy storage sealing ring. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a single-point mooring system liquid slip ring spring energy storage sealing ring optimization design method, which can determine the optimal parameter combination that meets the sealing performance requirements, to improve the overall sealing effect and improve the reliability of the sealing.

[0005] The present application is realized by the following technical solutions:

[0006] The single-point mooring system liquid slip ring spring energy storage sealing ring optimization design method comprises the following steps:

[0007] S1: According to the structure form, geometric shape and initial structure parameters of the single-point mooring system liquid slip ring spring energy storage sealing ring, a two-dimensional axisymmetric model of the liquid slip ring spring energy storage sealing ring is established;

[0008] S2: Based on the two-dimensional axisymmetric model of the spring-energized seal ring of the liquid sliding ring, the finite element analysis software is used to numerically simulate the mechanical behavior of the spring-energized seal ring of the liquid sliding ring under actual working conditions, and the peak contact stress, linear contact stress and maximum shear stress of the spring-energized seal ring of the liquid sliding ring under actual working conditions are obtained.

[0009] S3: Based on the peak contact stress, linear contact stress and maximum shear stress of the spring-energized seal ring of the liquid sliding ring under actual working conditions, the sealing performance evaluation criterion of the spring-energized seal ring of the liquid sliding ring is proposed, and the sealing performance evaluation system of the spring-energized seal ring of the liquid sliding ring is established.

[0010] S4: Based on the sealing performance evaluation system of the spring-energized seal ring of the liquid sliding ring, according to the influence degree of each structural parameter on the sealing performance and the influence of the structural parameter combination on the performance of the spring-energized seal ring of the liquid sliding ring, the structural parameters with sealing performance level exceeding the set standard are selected as the key structural parameters, and a data set containing the key structural parameters and the sealing performance level is formed, and the data set is divided into training set and test set.

[0011] S5: Construct a back propagation neural network model based on particle swarm optimization algorithm optimization, and train the weight and bias of the back propagation neural network model through the training set, and then evaluate the trained back propagation neural network model through the test set, and finally obtain the optimal parameter combination and the finite element model based on the optimal parameter combination that meet the sealing performance requirements.

[0012] The optimized structure of the spring-energized seal ring of the liquid sliding ring includes an outer U-shaped sealing jacket and an internal V-shaped energy storage spring, and the structural parameters include an inner inclination angle, an outer inclination angle, a lip thickness, a spring thickness, a compressed ring thickness, a lip length and a root arc radius. The mechanical behavior of the spring-energized seal ring of the liquid sliding ring under actual working conditions is numerically simulated by using finite element analysis software, including material property assignment, mesh division, contact definition, boundary condition and load application.

[0013] In the material property assignment, the U-shaped sealing jacket is made of polytetrafluoroethylene material, the elastic modulus is 800 MPa, and the Poisson's ratio is 0.46. The V-shaped energy storage spring is made of stainless steel material, the elastic modulus is 210 GPa, and the Poisson's ratio is 0.3.

[0014] In the mesh division, the mesh of each structural component of the spring-energized seal ring of the liquid sliding ring is CAX4R unit, the unit shape is quadrilateral, and the main contact parts of the spring-energized seal ring and the liquid sliding ring are locally refined.

[0015] In the contact definition, the contact algorithm adopts the penalty function method, and the friction coefficient is 0.02.

[0016] When optimized, with boundary conditions and loads applied, binding constraints are applied between the U-shaped sealing jacket and the V-shaped energy storage spring, full constraints are applied to the lower wall of the slip ring, and constraints are applied to the upper arm of the slip ring. Axial direction and In the axial direction, firstly in A displacement load of 0.2 mm was applied in the axial direction to simulate the assembly process, and then a medium pressure of 20 MPa was applied to simulate the medium working condition of the spring energy storage seal.

[0017] Furthermore, the evaluation criteria for the sealing performance of the liquid slip ring spring energy storage seal in step S3 are as follows:

[0018] Rule 1: The peak contact pressure between the upper lip of the spring energy storage seal ring and the contact surface is greater than the medium pressure;

[0019] Criterion 2: The maximum shear stress on the U-shaped sealing jacket is less than the ultimate shear strength of the U-shaped sealing jacket material;

[0020] Criterion 3: The line contact stress in the upper lip contact area of ​​the spring energy storage seal ring is the minimum stress required to satisfy the sealing requirement.

[0021] Furthermore, in step S4, key structural parameters are selected using the following method: a five-factor, ten-level orthogonal experiment is designed to simulate the mechanical response of the liquid slip ring spring energy storage seal ring with different combinations of structural parameters during actual operation. The influence of the combination of structural parameters on the performance of the liquid slip ring spring energy storage seal ring is analyzed. The finite element stress results of each combination of structural parameters are evaluated according to the performance evaluation system of the liquid slip ring spring energy storage seal ring, and structural parameters with high sealing performance are selected as key structural parameters.

[0022] Furthermore, the method for constructing the backpropagation neural network model optimized by the particle swarm optimization algorithm in step S5 is as follows:

[0023] S511: The position of each particle in the particle swarm optimization algorithm is associated with the weight matrix and bias vector parameter combination in the backpropagation neural network model, and the particle velocity is associated with the position update magnitude in the backpropagation neural network model. Iterative updates are performed according to equation (1):

[0024] (1);

[0025] in: express Time of the first The speed of each particle express Time of the first The speed of each particle Indicates the first learning factor. Represents the first random number. denotes the first particle history optimal position, denotes the position of the first particle at the time point, denotes the position of the first particle at the time point, denotes the position of the first particle at the time point, denotes the second learning factor, denotes the second random number, denotes the global optimal position;

[0026] S512: In the process of iterative updating, the particle searches the best combination of weights and biases in a dynamic manner until the mean square error of the back propagation neural network model is minimized, and the construction of the back propagation neural network model is completed.

[0027] Further, the optimal parameter combination satisfying the sealing performance requirement is obtained in step S5 according to the following method, and a finite element model based on the optimal parameter combination is established:

[0028] S521: The key structure parameters in the training set are normalized according to formula (2) to map the data to the interval [0, 1] to obtain the normalized structure parameter output variable:

[0029] (2);

[0030] wherein: denotes the normalized structure parameter output variable, denotes the structure parameter input variable before normalization, denotes the minimum value of the structure parameter input variable before normalization, denotes the maximum value of the structure parameter input variable before normalization;

[0031] S522: The normalized structure parameter output variable is taken as the input variable of the back propagation neural network model, and the output value of the back propagation neural network model is obtained according to formula (3):

[0032] (3);

[0033] wherein: denotes the output value of the back propagation neural network model, denotes the weight matrix of the back propagation neural network model from the input layer to the hidden layer, denotes the weight matrix of the back propagation neural network model from the hidden layer to the output layer, denotes the bias vector of the hidden layer of the back propagation neural network model, ​​​​​a bias vector representing an output layer of the back propagation neural network model, is an activation function,

[0034] S523: Calculate the root mean square error, mean absolute error, bias error and determination coefficient of the output value of the back propagation neural network model according to formula (4) using the test set:

[0035] (4);

[0036] wherein: the root mean square error of the output value of the back propagation neural network model, the total number of samples in the test set, the actual output value of the i-th sample, the predicted output value of the i-th sample, the mean absolute error of the output value of the back propagation neural network model, the bias error of the output value of the back propagation neural network model, the determination coefficient of the back propagation neural network model, the mean value of the actual output value of all samples; S524: Compare the calculated root mean square error, mean absolute error, bias error and determination coefficient of the output value of the back propagation neural network model with the corresponding preset range respectively, if the root mean square error, mean absolute error, bias error and determination coefficient of the output value of the back propagation neural network model are all within the corresponding preset range, the optimization is completed, thereby obtaining the optimal parameter combination that meets the sealing performance requirement, otherwise repeat steps S521 to S523 until the root mean square error, mean absolute error, bias error and determination coefficient of the output value of the back propagation neural network model are all within the corresponding preset range, obtain the optimal parameter group that meets the sealing performance requirement, and form a finite element model based on the optimal parameter combination.

[0037] The beneficial effects of the application are as follows:

[0038] The single-point mooring system liquid slip ring spring energy storage sealing ring optimization design method provided by the application has the following advantages:

[0039] 1. By selecting multiple key parameters for joint optimization, the comprehensive influence of parameter combination on sealing performance is fully considered. Compared with the traditional single variable analysis method, it is more rigorous and systematic, and can fully reflect the coupling effect and nonlinear relationship between parameters.

[0040] 1. By selecting multiple key parameters for joint optimization, the comprehensive influence of parameter combination on sealing performance is fully considered. Compared with the traditional single variable analysis method, it is more rigorous and systematic, and can fully reflect the coupling effect and nonlinear relationship between parameters.

[0041] ​2. By introducing peak contact stress, line contact stress and maximum shear stress indicators into the liquid sliding ring spring energy storage seal ring sealing performance evaluation system, a quantitative liquid sliding ring spring energy storage seal ring sealing performance evaluation criterion is constructed, which greatly improves the accuracy and scientificity of the sealing performance evaluation compared with the traditional qualitative evaluation criterion.

[0042] 3. By designing and applying a back propagation neural network model based on a particle swarm optimization algorithm, the fitting degree and optimization efficiency of the key parameter combination of the liquid sliding ring spring energy storage seal ring are greatly improved, which shows excellent performance in predicting and optimizing the performance of the seal ring, so that the obtained parameter combination is more in line with the actual working condition requirements.

[0043] 4. The particle swarm optimization algorithm and the back propagation neural network model are applied to the parameter optimization in the field of liquid sliding ring spring energy storage seal ring, and the data set is quickly constructed through a self-defined script, so that the data processing speed is significantly improved, and it has great potential in dealing with complex engineering problems. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 is the flowchart of the present application.

[0045] Figure 2 is the structure and structure parameter diagram of the liquid sliding ring spring energy storage seal ring of the present application.

[0046] Figure 3 is the peak contact stress and contact node curve diagram of the upper lip contact area of the spring energy storage seal ring. DETAILED DESCRIPTION

[0047] The single point mooring system liquid sliding ring spring energy storage seal ring optimization design method has a flowchart as shown in Figure 1 , and specifically includes the following steps:

[0048] S1: According to the structure form, geometric shape and initial structure parameters of the single point mooring system liquid sliding ring spring energy storage seal ring, a two-dimensional axisymmetric model of the liquid sliding ring spring energy storage seal ring is established;

[0049] Specifically, the structure of the liquid sliding ring spring energy storage seal ring includes a U-shaped sealing jacket and a V-shaped energy storage spring inside, and the structure parameters include the inner inclination angle, the outer inclination angle, the lip thickness, the spring thickness, the compressed ring thickness, the lip length and the root arc radius. The structure and structure parameter diagram of the liquid sliding ring spring energy storage seal ring is shown in Figure 2 , and the initial assignment of the structure parameters is shown in Table 1:

[0050] Table 1

[0051]

[0052] S2: based on the two-dimensional axisymmetric model of the spring energy sealing ring of the liquid sliding ring, the finite element analysis software is used to simulate the mechanical behavior of the spring energy sealing ring of the liquid sliding ring under actual working conditions, and the peak contact stress, linear contact stress and maximum shear stress of the spring energy sealing ring of the liquid sliding ring under actual working conditions are obtained;

[0053] Specifically, the numerical simulation of the mechanical behavior of the spring energy sealing ring of the liquid sliding ring under actual working conditions by the finite element analysis software includes material property assignment, mesh division, contact definition, boundary condition and load application.

[0054] Optimally, when the material performance is assigned, the U-shaped sealing jacket is made of polytetrafluoroethylene material, the elastic modulus is 800 MPa, and the Poisson's ratio is 0.46; the V-shaped energy spring is made of stainless steel material, the elastic modulus is 210 GPa, and the Poisson's ratio is 0.3.

[0055] The U-shaped sealing jacket is made of polytetrafluoroethylene with low friction coefficient and self-lubricating property, the V-shaped energy spring is made of stainless steel, and the elastic modulus of the inner and outer rings of the liquid sliding ring is much larger than that of the spring energy sealing ring material, which can be equivalent to a rigid body for analysis.

[0056] Optimally, when the mesh is divided, the mesh of each structural component of the spring energy sealing ring of the liquid sliding ring adopts CAX4R unit, the unit shape is quadrilateral, and the main contact parts of the spring energy sealing ring and the liquid sliding ring are locally refined.

[0057] Optimally, in the contact definition, the contact algorithm can adopt the penalty function method, and the friction factor is 0.02. The contact surface can be set as face-to-face contact, the upper wall of the sliding ring, the lower wall of the sliding ring and the contact surface of the sliding ring and the root of the sealing ring are main surfaces, the upper lip of the sealing ring, the lower lip of the sealing ring and the contact surface of the root of the sealing ring are secondary surfaces, the normal contact attribute is selected as hard contact, and the penetration between the surfaces is prohibited.

[0058] Optimally, in the boundary condition and load application, the binding constraint is applied between the U-shaped sealing jacket and the V-shaped energy spring, the lower wall of the sliding ring is applied with full constraint, the upper arm of the sliding ring is applied with constraint in the axial direction and in the axial direction, first, 0.2mm displacement load is applied in the axial direction to simulate the assembly process, and then medium pressure of 20Mpa is applied to simulate the medium working condition of the spring energy sealing ring.

[0059] The peak contact stress and contact node curve diagram of the upper lip contact area of the spring energy sealing ring are as Figure 3 shown.

[0060] S3: The peak contact stress, linear contact stress and maximum shear stress of the spring-energized sealing ring under actual working conditions are comprehensively considered, the sealing performance evaluation criterion of the spring-energized sealing ring is proposed, and the sealing performance evaluation system of the spring-energized sealing ring is established;

[0061] The static sealing characteristics of the spring-energized sealing ring depend on the resilience of the U-shaped sealing jacket and the resilience force generated by the V-shaped energy storage spring in the pre-compression.

[0062] In engineering, the peak contact pressure greater than the medium pressure is usually regarded as the criterion for evaluating the sealing performance. However, the single parameter variable analysis of the structural parameters of the spring-energized sealing ring shows that the increase of some sizes, such as the inner inclination angle, increases the peak contact pressure, and at the same time, a large shear stress is also increased, which may cause local deformation or even failure of the material, especially in the contact area with obvious stress concentration. Therefore, while considering the influence of size change on the peak contact pressure, the shear stress effect caused thereby must be fully considered to avoid adverse effects on the material performance and structural integrity. At the same time, the liquid sliding ring belongs to a rotating mechanism, and the friction generated by the contact will cause the spring-energized sealing ring to wear, reduce the service life and affect the sealing effect. Therefore, in actual work, the spring-energized sealing ring not only requires sealing performance, but also needs to maintain a low friction force. For the above reasons, the present application proposes a new sealing performance evaluation criterion combining the peak contact pressure, the maximum shear stress and the linear contact pressure:

[0063] Criterion one: the peak contact pressure between the upper lip of the spring-energized sealing ring and the contact surface is greater than the medium pressure;

[0064] Criterion two: the maximum shear stress on the U-shaped sealing jacket is less than the ultimate shear strength of the U-shaped sealing jacket material;

[0065] Criterion three: the linear contact stress of the contact area of the upper lip of the spring-energized sealing ring is the minimum stress under the premise of meeting the sealing.

[0066] Based on the above criteria, the linear contact pressure of the contact area is defined as follows:

[0067]

[0068] Wherein: represents the linear contact pressure, represents the coordinate value of the contact point in the Y-axis direction, represents the contact pressure distribution of the contact area, represents the coordinate value of the final contact point in the Y-axis direction, represents the coordinate value of the initial contact point in the Y-axis direction.

[0069] The contact area friction force is given by:

[0070] ;

[0071] wherein: represents the contact area friction force, represents the friction factor.

[0072] On this basis, the peak contact pressure, the maximum shear stress, and the stress level of the linear contact pressure are used to establish a sealing performance evaluation system for the spring-energized seal ring of the liquid lubricated ring.

[0073] The stress result scoring criteria are shown in Table 2:

[0074] Table 2

[0075]

[0076] The performance evaluation score of the spring-energized seal ring of the liquid lubricated ring is defined as:

[0077] wherein: is the performance evaluation score of the spring-energized seal ring of the liquid lubricated ring, is the maximum contact stress evaluation score of the seal ring lip under a certain parameter combination, is the maximum shear stress evaluation score of the clamping sleeve, is the linear contact stress evaluation score.

[0078] The performance level of the seal ring can be evaluated by the performance evaluation score of the spring-energized seal ring of the liquid lubricated ring The evaluation criteria are shown in Table 3:

[0079] Table 3

[0080]

[0081] Taking the initial structural parameters of the inner inclination angle A = 10°, the outer inclination angle B = 45°, the lip thickness C = 1.5 mm, the spring thickness D = 0.2 mm, the lip length F = 6 mm, and the root arc radius R = 1.5 mm as an example, the stress results obtained by the finite element analysis are the maximum contact stress of 35.848 Mpa, the maximum shear stress of 8.3292 Mpa, and the linear contact pressure of 17.977 Mpa, so that is 4, is 4, is 3, , and according to Table 3, the sealing performance level N of the seal ring under this parameter combination is level 4.

[0082] S4: Based on the liquid ring spring energy storage sealing ring sealing performance evaluation system, according to the influence degree of each structure parameter on the sealing performance and the influence of the structure parameter combination on the performance of the liquid ring spring energy storage sealing ring, the structure parameters with sealing performance level exceeding the set standard are selected as the key structure parameters, a data set containing the key structure parameters and the sealing performance level is formed, and the data set is divided into a training set and a test set;

[0083] In order to study the influence of the geometric size of the liquid ring spring energy storage sealing ring on the performance of the sealing ring, through single variable test, several size physical parameters which have greater influence on the sealing performance of the sealing ring are selected, the influence of parameter combination on the performance of the sealing ring is considered, and the key structure parameters are selected by the following method: design a five-factor ten-level orthogonal experiment, simulate the mechanical response of the liquid ring spring energy storage sealing ring with different structure parameter combinations in the actual operation process, analyze the influence law of the structure parameter combination of the liquid ring spring energy storage sealing ring on the performance of the liquid ring spring energy storage sealing ring, evaluate the finite element stress results of each structure parameter combination according to the sealing performance evaluation system of the liquid ring spring energy storage sealing ring, and select the structure parameters with sealing performance level exceeding the set standard as the key structure parameters.

[0084] The factors of the five-factor ten-level orthogonal experiment and the sealing performance level are shown in Table 4:

[0085] Table 4

[0086]

[0087] According to the finite element stress results of each parameter combination, the data set is constructed with the parameter combinations A, B, C, D and R as input variables and the sealing performance level N as output variable, which can be used for subsequent back propagation neural network model training and testing.

[0088] S5: Construct a back propagation neural network model based on particle swarm optimization algorithm optimization, train the weights and biases of the back propagation neural network model through the training set, and then evaluate the trained back propagation neural network model through the test set, finally obtain the optimal parameter combination meeting the sealing performance requirement and the finite element model based on the optimal parameter combination.

[0089] The back propagation neural network model is a kind of multi-layer feedforward neural network model for supervised learning, which adjusts the weights and biases of the network by minimizing the error function, so that the predicted value is as close to the true value as possible. The back propagation neural network model predicts the output value through the nonlinear mapping relationship between the input parameters and the output results obtained by training.

[0090] First, the dataset can be divided into two parts: 80% as the training set and 20% as the test set. The training set is used to train the weights and biases of the backpropagation neural network model, while the test set is used to evaluate the predictive performance of the trained backpropagation neural network model on unseen data, ensuring that the backpropagation neural network model can not only fit the training data well, but also make accurate predictions on new data.

[0091] Traditional backpropagation neural network models suffer from problems such as getting trapped in local optima during gradient descent and slow training speed. To address this, this invention introduces a particle swarm optimization algorithm to globally optimize the initial weights and biases of the backpropagation neural network model, thereby accelerating convergence and improving network performance.

[0092] Particle Swarm Optimization (PSO) is a global optimization algorithm based on swarm intelligence. It simulates the foraging behavior of birds, gradually finding the optimal solution to a problem through the updating and iteration of the population, i.e., particles. Each particle represents a candidate solution, updating its position by following both the individual extremum (local optimum) and the swarm extremum (global optimum). In this method, each particle in the swarm represents a combination of all values ​​and biases in the backpropagation neural network model, with a population size of 8. After initialization, the PSO algorithm is used for iterative updates, aiming to minimize the mean squared error of the backpropagation neural network model.

[0093] In particle swarm optimization, the position of each particle represents a set of parameters of the backpropagation neural network model, including the weight matrix and bias vector. Assume the backpropagation neural network model has... Each weight and The position of each biased particle is a length of... The vector represents the update magnitude of the position corresponding to the velocity of each particle, which is the update step size of the weights and biases of the backpropagation neural network model.

[0094] Specifically, the method for constructing a backpropagation neural network model optimized by the particle swarm optimization algorithm is as follows:

[0095] S511: The position of each particle in the particle swarm optimization algorithm is associated with the weight matrix and bias vector parameter combination in the backpropagation neural network model, and the particle velocity is associated with the position update magnitude in the backpropagation neural network model. Iterative updates are performed according to equation (1):

[0096] (1);

[0097] in: express Time of the first The speed of each particle express Time of the first The speed of each particle Indicates the first learning factor. Represents the first random number. Indicates the first The best position in the history of each particle express Time of the first The position of each particle. express Time of the first The position of each particle. Indicates the second learning factor. Represents the second random number. Indicates the globally optimal position;

[0098] S512: During the iterative update process, the particles dynamically search for the optimal combination of weights and biases until the mean square error of the backpropagation neural network model is minimized, thus completing the construction of the backpropagation neural network model.

[0099] After each position update, the particle's new position corresponds to the new weights and biases of the backpropagation neural network model;

[0100] Through this optimization process, the particles dynamically search for the optimal combination of weights and biases to minimize the error function of the backpropagation neural network model, thereby enabling the efficient search for the globally optimal weight and bias initialization of the backpropagation neural network model.

[0101] Furthermore, in step S5, the optimal parameter combination that satisfies the sealing performance requirements is obtained according to the following method, and a finite element model based on the optimal parameter combination is established:

[0102] S521: Normalize the key structural parameters in the training set according to equation (2), and map the data to... The interval is used to obtain the normalized structure parameter output variables:

[0103] (2);

[0104] in: The output variable represents the normalized structure parameters. This represents the structural parameter input variables before normalization. This represents the minimum value of the input variables representing the structural parameters before normalization. This represents the maximum value of the input variables for the structural parameters before normalization;

[0105] S522: Using the normalized structural parameter output variables as the input variables of the backpropagation neural network model, the output value of the backpropagation neural network model is obtained according to equation (3):

[0106] (3);

[0107] in: This represents the output value of the backpropagation neural network model. This represents the weight matrix from the input layer to the hidden layer of the backpropagation neural network model. This represents the weight matrix from the hidden layer to the output layer of the backpropagation neural network model. This represents the bias vector of the hidden layer in the backpropagation neural network model. This represents the bias vector of the output layer of the backpropagation neural network model. Represented as an activation function,

[0108] S523: Using the test set, calculate the root mean square error, mean absolute error, bias error, and coefficient of determination of the output value of the backpropagation neural network model according to equation (4):

[0109] (4);

[0110] in: This represents the root mean square error of the output value of the backpropagation neural network model. This represents the total number of samples in the test set. Indicates the first The actual output value of each sample Indicates the first The predicted output value for each sample. This represents the mean absolute error of the output values ​​of the backpropagation neural network model. This represents the deviation error of the output value of the backpropagation neural network model. The coefficients of determination represent the backpropagation neural network model. This represents the average of the actual output values ​​for all samples.

[0111] S524: Compare the calculated root mean square error, mean absolute error, deviation error, and coefficient of determination of the backpropagation neural network model with the corresponding preset ranges. If the root mean square error, mean absolute error, deviation error, and coefficient of determination of the backpropagation neural network model are all within the corresponding preset ranges, the optimization is complete, thus obtaining the optimal parameter combination that meets the sealing performance requirements. Otherwise, repeat steps S521 to S523 until the root mean square error, mean absolute error, deviation error, and coefficient of determination of the backpropagation neural network model are all within the corresponding preset ranges, thus obtaining the optimal parameter set that meets the sealing performance requirements and forming a finite element model based on the optimal parameter combination.

[0112] The present application can realize the construction and training of the back propagation neural network model optimized based on the particle swarm optimization algorithm by using the MATLAB program. The high-precision fitting model with a determination coefficient of 0.8928 is finally obtained through 1000 times of iterative training of the model

[0113] Table 5

[0114]

[0115] Based on the obtained optimal parameter combination, a corresponding finite element model can be established and solved. The calculation results show that the peak contact stress is 55.473236 MPa, the maximum shear stress is 7.864998 MPa, and the linear contact stress is 8.561987 MPa, and the corresponding sealing performance grade N is Level 6, which is consistent with the prediction result of the back propagation neural network model. This result verifies the rigor and reliability of the method, and proves the effectiveness of the back propagation neural network model based on the particle swarm optimization algorithm in the prediction of sealing performance and parameter optimization.

[0116] In summary, the present application provides a single-point mooring system liquid slip ring spring energy storage seal ring optimization design method, which can determine the optimal parameter combination that meets the sealing performance requirements, thereby improving the overall sealing effect of the liquid slip ring spring energy storage seal ring and improving the reliability of the seal.

[0117] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. An optimized design method for a liquid slip ring spring energy storage seal in a single-point mooring system, characterized in that: Includes the following steps: S1: Establish a two-dimensional axisymmetric model of the liquid slip ring spring energy storage seal ring based on its structural form, geometry, and initial structural parameters in a single-point mooring system. S2: Based on the two-dimensional axisymmetric model of the liquid slip ring spring energy storage seal, the mechanical behavior of the liquid slip ring spring energy storage seal under actual working conditions is numerically simulated using finite element analysis software, and the peak contact stress, line contact stress and maximum shear stress of the liquid slip ring spring energy storage seal under actual working conditions are obtained. S3: Based on the peak contact stress, line contact stress and maximum shear stress experienced by the liquid slip ring spring energy storage seal under actual working conditions, we propose the sealing performance evaluation criteria for the liquid slip ring spring energy storage seal and establish the sealing performance evaluation system for the liquid slip ring spring energy storage seal. S4: Based on the sealing performance evaluation system of liquid slip ring spring energy storage seal, according to the influence of each structural parameter on the sealing performance and the influence of the combination of structural parameters on the performance of liquid slip ring spring energy storage seal, structural parameters whose sealing performance level exceeds the set standard are selected as key structural parameters, forming a dataset containing key structural parameters and sealing performance level, and dividing the dataset into training set and test set; S5: Construct a backpropagation neural network model optimized by particle swarm optimization algorithm, train the weights and biases of the backpropagation neural network model through the training set, evaluate the trained backpropagation neural network model through the test set, and finally obtain the optimal parameter combination that meets the sealing performance requirements and the finite element model based on the optimal parameter combination.

2. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 1, characterized in that: The structure of the liquid slip ring spring energy storage seal ring includes an outer U-shaped sealing jacket and an inner V-shaped energy storage spring. The structural parameters include inclination angle, outclination angle, lip thickness, spring thickness, pressed ring thickness, lip length, and root arc radius. The mechanical behavior of the liquid slip ring spring energy storage seal under actual working conditions was numerically simulated using finite element analysis software, including material property assignment, mesh generation, contact definition, boundary conditions, and load application.

3. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 2, characterized in that: When assigning material properties, the U-shaped sealing jacket is made of polytetrafluoroethylene with an elastic modulus of 800 MPa and a Poisson's ratio of 0.

46. The V-shaped energy storage spring is made of stainless steel with an elastic modulus of 210 GPa and a Poisson's ratio of 0.

3.

4. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 2, characterized in that: During mesh generation, the mesh of each structural component of the liquid slip ring spring energy storage seal ring adopts CAX4R elements, and the element shape is quadrilateral. The main contact areas between the spring energy storage seal ring and the liquid slip ring are locally refined.

5. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 2, characterized in that: When defining contact, the contact algorithm uses the penalty function method, and the friction coefficient is 0.

02.

6. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 2, characterized in that: When boundary conditions and loads are applied, binding constraints are applied between the U-shaped sealing jacket and the V-shaped energy storage spring, full constraints are applied to the lower wall of the slip ring, and constraints are applied to the upper arm of the slip ring. Axial direction and In the axial direction, firstly in A displacement load of 0.2 mm was applied in the axial direction to simulate the assembly process, and then a medium pressure of 20 MPa was applied to simulate the medium working condition of the spring energy storage seal.

7. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 1, characterized in that: The sealing performance evaluation criteria for the liquid slip ring spring energy storage seal in step S3 are as follows: Rule 1: The peak contact pressure between the upper lip of the spring energy storage seal ring and the contact surface is greater than the medium pressure; Criterion 2: The maximum shear stress on the U-shaped sealing jacket is less than the ultimate shear strength of the U-shaped sealing jacket material; Criterion 3: The line contact stress in the upper lip contact area of ​​the spring energy storage seal ring is the minimum stress required to satisfy the sealing requirement.

8. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 1, characterized in that: In step S4, key structural parameters are selected using the following method: a five-factor, ten-level orthogonal experiment is designed to simulate the mechanical response of the liquid slip ring spring energy storage seal ring with different combinations of structural parameters during actual operation. The influence of the combination of structural parameters on the performance of the liquid slip ring spring energy storage seal ring is analyzed. The finite element stress results of each combination of structural parameters are evaluated according to the performance evaluation system of the liquid slip ring spring energy storage seal ring, and structural parameters with high sealing performance are selected as key structural parameters.

9. The optimized design method for the liquid slip ring spring energy storage seal ring of a single-point mooring system according to claim 1, characterized in that: The method for constructing the backpropagation neural network model optimized by the particle swarm optimization algorithm in step S5 is as follows: S511: The position of each particle in the particle swarm optimization algorithm is associated with the weight matrix and bias vector parameter combination in the backpropagation neural network model, and the particle velocity is associated with the position update magnitude in the backpropagation neural network model. Iterative updates are performed according to equation (1): （1）； in: express Time of the first The speed of each particle express Time of the first The speed of each particle Indicates the first learning factor. Represents the first random number. Indicates the first The best position in the history of each particle express Time of the first The position of each particle. express Time of the first The position of each particle. Indicates the second learning factor. Represents the second random number. Indicates the globally optimal position; S512: During the iterative update process, the particles dynamically search for the optimal combination of weights and biases until the mean square error of the backpropagation neural network model is minimized, thus completing the construction of the backpropagation neural network model.

10. The optimized design method for the liquid slip ring spring energy storage seal ring of the single-point mooring system according to claim 1, characterized in that: In step S5, the optimal parameter combination that meets the sealing performance requirements is obtained as follows, and a finite element model based on the optimal parameter combination is established: S521: Normalize the key structural parameters in the training set according to equation (2), and map the data to... The interval is used to obtain the normalized structure parameter output variables: （2）； in: The output variable represents the normalized structure parameters. This represents the structural parameter input variables before normalization. This represents the minimum value of the input variables representing the structural parameters before normalization. This represents the maximum value of the input variables for the structural parameters before normalization; S522: Using the normalized structural parameter output variables as the input variables of the backpropagation neural network model, the output value of the backpropagation neural network model is obtained according to equation (3): （3）； in: This represents the output value of the backpropagation neural network model. This represents the weight matrix from the input layer to the hidden layer of the backpropagation neural network model. This represents the weight matrix from the hidden layer to the output layer of the backpropagation neural network model. This represents the bias vector of the hidden layer in the backpropagation neural network model. This represents the bias vector of the output layer of the backpropagation neural network model. Represented as an activation function, S523: Using the test set, calculate the root mean square error, mean absolute error, bias error, and coefficient of determination of the output value of the backpropagation neural network model according to equation (4): （4）； in: This represents the root mean square error of the output value of the backpropagation neural network model. This represents the total number of samples in the test set. Indicates the first The actual output value of each sample Indicates the first The predicted output value for each sample. This represents the mean absolute error of the output values ​​of the backpropagation neural network model. This represents the deviation error of the output value of the backpropagation neural network model. The coefficients of determination represent the backpropagation neural network model. This represents the average of the actual output values ​​for all samples. S524: Compare the calculated root mean square error, mean absolute error, deviation error, and coefficient of determination of the backpropagation neural network model with the corresponding preset ranges. If the root mean square error, mean absolute error, deviation error, and coefficient of determination of the backpropagation neural network model are all within the corresponding preset ranges, the optimization is complete, thus obtaining the optimal parameter combination that meets the sealing performance requirements. Otherwise, repeat steps S521 to S523 until the root mean square error, mean absolute error, deviation error, and coefficient of determination of the backpropagation neural network model are all within the corresponding preset ranges, thus obtaining the optimal parameter set that meets the sealing performance requirements and forming a finite element model based on the optimal parameter combination.

Images