Method, system and equipment for predicting sloshing wave height of fluid under random load and medium

By comprehensively considering the global optimal value and the optimal value of all particles in the particle swarm algorithm, the particle swarm algorithm is improved and the RBF neural network is optimized, which solves the computational complexity and inaccuracy of predicting hydrodynamic characteristics of fluids in the tank under random loads. The problem of sexual problems achieves efficient and accurate prediction of swing waves.

CN119005045APending Publication Date: 2024-11-22CHENGDU UNIV OF INFORMATION TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411023601.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The prior art has high computational complexity when predicting the hydrodynamic characteristics of fluids in the tank under random loads, resulting in a long calculation time and inaccurate prediction results, making it difficult to deal with nonlinear and complex physical phenomena.

Method used

By adding the optimal value of each particle in the global and the optimal value of all particles in the global update formula in the speed update formula of the particle swarm algorithm, the particle swarm algorithm is improved, and the parameters of the RBF neural network are used as the improved particle swarm algorithm. Particles are optimized to construct data sets to predict the sloshing wave height of fluid in the tank under random loads.

Benefits of technology

The overfit linearity of RBF neural network is reduced, its generalization is improved, the global optimal value can be found accurately and quickly, and the prediction efficiency and accuracy are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119005045A_ABST
    Figure CN119005045A_ABST
Patent Text Reader

Abstract

The invention provides a method, a system, equipment and a medium for predicting sloshing wave height of fluid under random load, and belongs to the field of machine learning, and the method comprises the following steps: adding the optimal value of each particle in the global situation, the optimal values of all particles in the global situation and random numbers in a speed updating formula of a particle swarm algorithm; improving a speed updating formula of the particle swarm algorithm to obtain an improved particle swarm algorithm; taking the parameters of the RBF neural network as particles of the improved particle swarm algorithm, and optimizing the parameters of the RBF neural network by using the improved particle swarm algorithm to obtain the improved RBF neural network; and acquiring multiple groups of random load data and multiple groups of storage tank parameter data, constructing a data set by using the multiple groups of load data and the multiple groups of storage tank parameter data, and inputting the data in the data set into the improved RBF neural network to obtain the sloshing wave height of the fluid in the storage tank under the random load. According to the invention, the efficiency and accuracy of sloshing wave height prediction can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of machine learning, and in particular relates to a method, system, equipment and medium for predicting sloshing wave height of a fluid under random load. Background Art

[0002] Earthquakes and waves are typical random loads. Earthquakes occur frequently and with great intensity in my country. The latter has a significant impact on marine transportation and industrial activities. They can cause irregular movement of storage tanks, thereby causing violent nonlinear sloshing of the fluid in the tanks. This oscillation is accompanied by strong impact loads and overturning moments, which affect the stability of the tanks. Once a tank is damaged, it will not only cause economic losses, but may also cause serious secondary disasters such as environmental pollution and fire accidents. Therefore, a deep understanding of the hydrodynamic characteristics of the fluid in the tank under random loads is crucial for the safety analysis and structural optimization design of the tank.

[0003] The existing technology mainly uses numerical simulation and structural analysis to predict the hydrodynamic characteristics of fluids in storage tanks under random loads. In terms of numerical simulation, the mainstream is computational fluid dynamics (CFD) simulation, which can predict the flow field and stress distribution of fluids under random loads by solving the Navier-Stokes equations. For two-phase or multi-phase fluids containing gas and liquid, multiphase flow models can also be used for simulation, considering the behavior of fluids with different densities and viscosities under random loads; in terms of structural analysis, finite element analysis combined with random load spectra and fluid dynamics models can simulate the vibration response of storage tanks under random conditions and the stress conditions of the fluid, taking into account the interaction between the structure and the fluid.

[0004] However, the above methods have several major problems in predicting the hydrodynamic characteristics of the fluid in the tank under random loads. The first is the high computational complexity, which requires solving a large number of partial differential equations, especially for complex fluid dynamics phenomena, resulting in a long calculation time; the second is the difficulty in handling nonlinear and complex physical phenomena, resulting in inaccurate prediction results of the fluid hydrodynamic characteristics (sloshing wave height) under random loads. Summary of the invention

[0005] In order to overcome the inaccuracy of the prediction results of fluid sloshing wave height under random loads, the present invention provides a method for predicting the sloshing wave height of fluid under random loads, comprising the following steps:

[0006] The optimal value of each particle in the world, the optimal value of all particles in the world and random numbers are added to the speed update formula of the particle swarm algorithm, and the speed update formula of the particle swarm algorithm is improved to obtain the improved particle swarm algorithm.

[0007] The parameters of the RBF neural network are used as particles of the improved particle swarm algorithm, and the improved particle swarm algorithm is used to optimize the parameters of the RBF neural network to obtain the improved RBF neural network;

[0008] Multiple sets of random load data and multiple sets of parameter data of the fluid in the tank are obtained, and a data set is constructed using the multiple sets of load data and multiple sets of parameter data of the fluid in the tank. The data in the data set are input into the improved RBF neural network to obtain the sloshing wave height of the fluid in the tank under random load.

[0009] Preferably, the improved particle swarm algorithm is used to optimize the parameters of the RBF neural network, comprising the following steps:

[0010] Initialize the improved particle swarm algorithm and determine the parameters of the improved particle swarm algorithm;

[0011] Initialize the RBF neural network, determine the number of hidden layer nodes of the BF neural network and the type of radial basis function;

[0012] The parameters of RBF neural network are used as the initial population particles of the improved particle swarm algorithm;

[0013] Calculate the fitness value of the initial population particles;

[0014] Use the improved particle swarm algorithm to update the position and velocity of the initial population particles to obtain the updated fitness value;

[0015] Repeat the steps of calculating the fitness value of the initial population particles and updating the positions and velocities of the initial population particles until the predetermined number of iterations is reached, then stop the condition and obtain the best fitness value, that is, the best parameters of the RBF neural network.

[0016] Preferably, the random load data includes peak acceleration PGA, peak velocity PGV, PGA / PGV and duration of the random load.

[0017] Preferably, the parameter data of the fluid in the storage tank is different filling level data of the storage tank.

[0018] Preferably, before inputting the data in the data set into the improved RBF neural network, the method further includes preprocessing the data set, specifically processing the missing values, abnormal values ​​and duplicate values ​​in the data set.

[0019] Preferably, the improved speed update formula is:

[0020]

[0021] In the formula, v iis the speed of the particle, w is the inertia weight that affects the updated particle speed, g i is the optimal value of each particle in the global gb is the optimal value of all particles in the world, r1 and r2 represent random numbers between 0 and 1, r0 is a newly added random number, c1 represents the individual acceleration constant, c2 represents the social acceleration constant, p i is the optimal position of each particle, s i represents the current position of the particle, p gd represents the best position of all particles in the world.

[0022] Preferably, the parameters of the improved particle swarm algorithm include: number of particles, number of iterations, inertia weight and acceleration coefficient.

[0023] The present invention also provides a sloshing wave height prediction system for fluid under random load, comprising:

[0024] The algorithm optimization module is used to add the optimal value of each particle in the world, the optimal value of all particles in the world and the random number to the speed update formula of the particle swarm algorithm, improve the speed update formula of the particle swarm algorithm, and obtain the improved particle swarm algorithm;

[0025] The RBF neural network optimization module is used to use the parameters of the RBF neural network as particles of the improved particle swarm algorithm, and use the improved particle swarm algorithm to optimize the parameters of the RBF neural network to obtain the improved RBF neural network;

[0026] The data processing module is used to obtain multiple sets of random load data and multiple sets of parameter data of the fluid in the tank, and use the multiple sets of load data and multiple sets of parameter data of the fluid in the tank to construct a data set, input the data in the data set into the improved RBF neural network, and obtain the sloshing wave height of the fluid in the tank under random load.

[0027] The present invention also provides a computer device, comprising a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the method for predicting the sloshing wave height of the fluid under random load.

[0028] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the method for predicting the sloshing wave height of the fluid under random load.

[0029] The method, system, device and medium for predicting sloshing wave height of fluid under random load provided by the present invention have the following beneficial effects:

[0030] The present invention improves the speed update formula of the particle swarm algorithm by comprehensively considering the influence of the optimal value of each particle in the global and the optimal value of all particles in the global in the particle swarm algorithm; the parameters of the RBF neural network can be optimized by the improved particle swarm algorithm, so as to obtain the improved RBF neural network; the sloshing wave height of the fluid under the random load can be calculated by inputting random load data and parameter data of multiple groups of fluids in the storage tank into the improved RBF neural network; the present invention can reduce the overfitting linearity of the RBF neural network and improve its generalization through the improved particle swarm algorithm, so as to accurately and quickly find the global optimal value, that is, the sloshing wave height, so as to improve the efficiency and accuracy of prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 is a flow chart of the sloshing wave height prediction method of the present invention;

[0033] Figure 2 The present invention is a flow chart for optimizing the parameters of the RBF neural network;

[0034] Figure 3 A two-dimensional rectangular tank model for wave height prediction according to an embodiment of the present invention;

[0035] Figure 4 The wave height prediction training set and test set effect diagram according to the implementation case of the present invention; wherein, Figure 4 (a) is the prediction result of the improved RBF neural network on the training set; Figure 4 (b) is the prediction result of the improved RBF neural network on the test set. DETAILED DESCRIPTION

[0036] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0037] In the description of the present invention, it is to be understood that the terms “center”, “longitudinal”, “lateral”, “length”, “width”, “thickness”, “up”, “down”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside”, “outside”, “axial”, “radial”, “circumferential”, etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the technical solutions of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0038] In addition, the terms "first", "second", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance. In the description of the present invention, it should be noted that, unless otherwise clearly specified or limited, the terms "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. In the description of the present invention, unless otherwise specified, "plurality" means two or more, which will not be described in detail here.

[0039] Example 1

[0040] The present invention provides a method for predicting the sloshing wave height of a fluid under random loads, specifically: Figure 1 As shown, the following steps are included:

[0041] Step 1: Improve the speed update formula of the particle swarm algorithm by adding the optimal value of each particle in the world, the optimal value of all particles in the world, and the random number to the speed update formula to obtain the improved particle swarm algorithm. The parameters of the improved particle swarm algorithm are shown in Table 1.

[0042] Step 2: Use the parameters of the RBF neural network as particles of the improved particle swarm algorithm, and use the improved particle swarm algorithm to optimize the parameters of the RBF neural network to obtain the improved RBF neural network, as shown in the following figure. Figure 2 shown.

[0043] Table 1 Parameters of particle swarm algorithm

[0044] parameter value Local search capability c1 1.5 Global search capability c2 1.7 Maximum number of evolutions 50 Maximum population size 5 Position / speed minimum 0.1 Position / speed maximum value 1000 Elastic modulus 1

[0045] Step 2.1: Initialize PSO parameters: Determine the parameters of the PSO algorithm (i.e., particle swarm algorithm), including the number of particles, number of iterations, inertia weight, acceleration factor, etc. When determining the number of particles, it is necessary to consider the complexity of the problem and the limitations of computing resources. Some experiments are usually performed to determine the optimal number. The number of iterations is adjusted to balance the computational cost of the algorithm and the convergence speed of the results. The inertia weight and acceleration factor are set to balance the trade-off between local search and global search.

[0046] Step 2.2: Initialize the RBF neural network: Determine the structure of the RBF neural network, including the number of hidden layer nodes and the type of radial basis function (such as Gaussian function). Initialize the network parameters, including weights and biases.

[0047] Step 2.3: PSO initialization: Randomly initialize a group of particles, each particle represents a possible solution of the RBF neural network. Each particle has a position and velocity, and the position represents the parameter value of the RBF neural network.

[0048] Step 2.4: Evaluate fitness: Use appropriate loss functions or evaluation indicators to measure the prediction effect of the model. Common ones include mean square error (MSE), mean absolute error (MAE), etc. Divide the data set into a training set and a validation set, and use the validation set to evaluate the generalization ability of the model.

[0049] Step 2.5: Update particles. The improved speed update formula is:

[0050]

[0051] In the formula, v i is the speed of the particle, w is the inertia weight that affects the updated particle speed, g i is the optimal value of each particle in the global gb is the optimal value of all particles in the world, r1 and r2 represent random numbers between 0 and 1, r0 is a newly added random number, c1 represents the individual acceleration constant, c2 represents the social acceleration constant, p i is the optimal position of each particle, s i represents the current position of the particle, p gd represents the best position of all particles in the world.

[0052] The position and velocity of each particle are updated by using the improved velocity update formula so that the particle moves in a direction with better fitness. The update formula includes consideration of the current position, velocity, and the influence of the global optimal solution and the individual optimal solution.

[0053] Step 2.6: Test the model: Use the optimal RBF neural network parameters to predict the test set and evaluate the performance of the model. Evaluation indicators such as mean square error and mean absolute error are usually used to evaluate the accuracy and generalization ability of the model.

[0054] Step 2.7: Result analysis and optimization: Analyze the prediction effect of the model. If the effect is not satisfactory, try to adjust the parameters of the PSO and RBF neural network, retrain and test until satisfactory results are achieved.

[0055] Step 3: Obtain multiple sets of random load data and multiple sets of parameter data of the fluid in the tank, and use the multiple sets of load data and multiple sets of parameter data of the fluid in the tank to construct a data set, input the data in the data set into the improved RBF neural network, and obtain the sloshing wave height of the fluid in the tank under random load.

[0056] Obtain multiple sets of random load data and multiple sets of tank parameter data. The random load data include the peak acceleration PGA, peak velocity PGV, PGA / PGV and duration of the random load; the parameter data of the fluid in the tank are different filling level data of the tank. Then use multiple sets of random load data and multiple sets of tank parameter data to construct a data set. In order to ensure the quality and integrity of the data set, the data set needs to be preprocessed, specifically processing missing values, outliers and duplicate values. Data visualization and descriptive statistical analysis can be used to gain an in-depth understanding of the characteristics and distribution of the data, as well as possible correlations.

[0057] Then the data in the data set are input into the improved RBF neural network to obtain the sloshing wave height of the fluid in the tank under random load.

[0058] The present invention also provides a sloshing wave height prediction system for fluid under random load, comprising an algorithm optimization module, an RBF neural network optimization module and a data processing module. The algorithm optimization module is used to improve the velocity update formula of the particle swarm algorithm, and the optimal value of each particle in the global, the optimal value of all particles in the global and a random number are added to the velocity update formula to obtain an improved particle swarm algorithm; the RBF neural network optimization module is used to use the parameters of the RBF neural network as particles of the improved particle swarm algorithm, and use the improved particle swarm algorithm to optimize the parameters of the RBF neural network to obtain an improved RBF neural network; the data processing module is used to obtain multiple sets of random load data and multiple sets of tank parameter data, and use the multiple sets of random load data and multiple sets of tank parameter data to construct a data set, and input the data in the data set into the improved RBF neural network to obtain the hydrodynamic characteristics of the fluid in the tank under random load.

[0059] The present invention also provides a computer device, comprising a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the method for predicting the sloshing wave height of the fluid under random load.

[0060] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the method for predicting the sloshing wave height of the fluid under random load.

[0061] Example 2

[0062] Next, the wave height prediction method under earthquake load is further explained through a specific example.

[0063] like Figure 3 As shown, the present invention adopts a two-dimensional rectangular tank model with a length of L = 20.0m and a height of H = 20.0m. Seismic excitation is applied in the x direction. In engineering, water tanks are usually divided into "shallow" tanks and "high" tanks, corresponding to h / L≤0.5 and h / L>0.5, respectively. The depth of the storage tank is also divided into shallow (h / L≤0.05~0.1), medium (0.05~0.1≤h / L≤0.2~0.25), limited (tanh(πh / L)<1) and deep (tanh(πh / L)≈1). Depending on the experimental requirements and the selected equipment, the division range has a certain fluctuation, among which h / L=0.34 is the critical depth. Taking this into account, four still water depths (h=4.0m, 6.8m, 10.0m and 12.0m) were used in the simulation, covering filling levels from 20% to 60%.

[0064] As shown in Table 2, the present invention selects 9 earthquake events and 13 earthquake records, which are arranged from small to large according to the peak acceleration PGA. There are 13 sets of earthquake data, 4 filling levels, and a total of 52 sets of data sets, which are divided into two groups: training set and test set.

[0065] Table 2. Relevant parameters of earthquakes

[0066]

[0067] During the training phase, 80% of the actual data points are used to build the model. For the test set, 20% of the data points are used to check the accuracy of the model. The prediction results are as follows: Figure 4 As shown, Figure 4 (a) is the prediction result of the improved RBF neural network on the training set; Figure 4(b) is the prediction result of the improved RBF neural network on the test set. From the results, we can see that the improved RBF neural network has a good performance on both the training set and the test set. The root mean square error of the training set is 0.18739, and the root mean square error of the test set is 0.31874.

[0068] The embodiments described above are only preferred specific implementation modes of the present invention, and the protection scope of the present invention is not limited thereto. Any simple changes or equivalent replacements of the technical solutions that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belong to the protection scope of the present invention.

Claims

1. A method for predicting sloshing wave height of a fluid under random load, characterized in that: The steps include: The optimal value of each particle in the world, the optimal value of all particles in the world and random numbers are added to the speed update formula of the particle swarm algorithm, and the speed update formula of the particle swarm algorithm is improved to obtain the improved particle swarm algorithm. The parameters of the RBF neural network are used as particles of the improved particle swarm algorithm, and the improved particle swarm algorithm is used to optimize the parameters of the RBF neural network to obtain the improved RBF neural network; Multiple sets of random load data and multiple sets of parameter data of the fluid in the tank are obtained, and a data set is constructed using the multiple sets of load data and multiple sets of parameter data of the fluid in the tank. The data in the data set are input into the improved RBF neural network to obtain the sloshing wave height of the fluid in the tank under random load.

2. The method for predicting sloshing wave height of fluid under random load according to claim 1, characterized in that: The improved particle swarm algorithm is used to optimize the parameters of the RBF neural network, including the following steps: Initialize the improved particle swarm algorithm and determine the parameters of the improved particle swarm algorithm; Initialize the RBF neural network, determine the number of hidden layer nodes of the BF neural network and the type of radial basis function; The parameters of RBF neural network are used as the initial population particles of the improved particle swarm algorithm; Calculate the fitness value of the initial population particles; Use the improved particle swarm algorithm to update the position and velocity of the initial population particles to obtain the updated fitness value; Repeat the steps of calculating the fitness value of the initial population particles and updating the positions and velocities of the initial population particles until the predetermined number of iterations is reached, then stop the condition and obtain the best fitness value, that is, the best parameters of the RBF neural network.

3. The method for predicting sloshing wave height of fluid under random load according to claim 1, characterized in that: The random load data includes the peak acceleration PGA, the peak velocity PGV, PGA / PGV and duration of the random load.

4. The method for predicting sloshing wave height of fluid under random load according to claim 1, characterized in that: The parameter data of the fluid in the tank are different filling level data of the tank.

5. The method for predicting sloshing wave height of fluid under random load according to claim 1, characterized in that: Before inputting the data in the data set into the improved RBF neural network, the method also includes preprocessing the data set, specifically processing the missing values, abnormal values ​​and repeated values ​​in the data set.

6. The method for predicting sloshing wave height of fluid under random load according to claim 1, characterized in that: The improved speed update formula is: In the formula, v i is the speed of the particle, w is the inertia weight that affects the updated particle speed, g i is the optimal value of each particle in the world, g gb is the optimal value of all particles in the world, r1 and r2 represent random numbers between 0 and 1, r0 is a newly added random number, c1 represents the individual acceleration constant, c2 represents the social acceleration constant, p i is the optimal position of each particle, s i represents the current position of the particle, p gd represents the best position of all particles in the world.

7. The method for predicting sloshing wave height of fluid under random load according to claim 1, characterized in that: The parameters of the improved particle swarm algorithm include: number of particles, number of iterations, inertia weight and acceleration coefficient.

8. A sloshing wave height prediction system for fluid under random loads, characterized in that: include: The algorithm optimization module is used to add the optimal value of each particle in the world, the optimal value of all particles in the world and the random number to the speed update formula of the particle swarm algorithm, improve the speed update formula of the particle swarm algorithm, and obtain the improved particle swarm algorithm; The RBF neural network optimization module is used to use the parameters of the RBF neural network as particles of the improved particle swarm algorithm, and use the improved particle swarm algorithm to optimize the parameters of the RBF neural network to obtain the improved RBF neural network; The data processing module is used to obtain multiple sets of random load data and multiple sets of parameter data of the fluid in the tank, and use the multiple sets of load data and multiple sets of parameter data of the fluid in the tank to construct a data set, input the data in the data set into the improved RBF neural network, and obtain the sloshing wave height of the fluid in the tank under random load.

9. A computer device, characterized in that: It comprises a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to execute the sloshing wave height prediction method of the fluid under random load according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the method for predicting the sloshing wave height of a fluid under random load according to any one of claims 1 to 7.