Method and device for predicting wind-sand erosion rate of composite insulator

Abstract

The invention discloses a method and device for predicting the wind sand erosion rate of a composite insulator, and the method comprises the steps: building a three-dimensional numerical model of the composite insulator according to the physical parameters of the composite insulator; setting simulation parameters for the three-dimensional numerical model; the simulation parameters are input into simulation software for simulation calculation, the erosion rate Ver of the composite insulator under wind and sand erosion is obtained, and the simulation software is provided with the three-dimensional numerical model; obtaining a dimensionless group based on the erosion rate Ver and the simulation parameters, and performing dimension reduction processing to obtain a training data set; and correcting a pre-acquired first Gaussian model based on the training data set to obtain a second Gaussian model, and predicting the erosion rate of the composite insulator based on the second Gaussian model to obtain a predicted erosion rate. According to the invention, the rate of the composite insulator eroded by the wind and sand can be predicted more accurately and quickly.

Description

Technical Field

[0001] The invention relates to the technical field of high-voltage power transmission and transformation engineering, and in particular to a method and device for predicting the wind and sand erosion rate of a composite insulator. Background Art

[0002] Compared with traditional ceramic composite insulators, composite insulators have the advantages of strong anti-pollution flashover capability, hydrophobic migration, and light weight. In recent years, they have been widely used in UHV transmission projects. Although composite insulators in operation on the grid have excellent performance, their performance changes during long-term operation in various complex natural environments still need to be studied.

[0003] The deserts, Gobi and wasteland areas in western and northern my country have high-quality solar and wind energy resources, and the demand for ultra-high voltage power transmission is rigid and continuous. However, operating experience shows that the special natural conditions such as strong wind and sand in deserts, Gobi and wasteland areas are very harmful to power transmission and transformation equipment, and have caused many accidents. Strong wind and sand severe weather will cause surface wear, increased roughness, decreased hydrophobicity, thinning of the shed skirt, formation of holes and even tearing of composite external insulation materials of power equipment. At the same time, strong winds will cause the composite insulators to deform, resulting in a decrease in their insulation and mechanical properties. For composite insulators operating in strong wind and sand areas, their resistance to wind and sand erosion has more stringent requirements. Therefore, it is particularly important to make a reasonable prediction of wind and sand erosion resistance before the actual use of composite insulators.

[0004] At present, the erosion rate prediction method for composite insulators in strong wind and sand environment mainly simulates the wind and sand environment by setting up a test wind tunnel in actual tests, and makes the sand particles hit the surface of the composite insulator at a certain wind speed to obtain test data. This is not suitable in actual engineering design. The main problems are: there are many types of composite insulators, unclear test conditions, long test cycle, and poor economic efficiency of actual tests; the size of composite insulators of UHV transmission lines is large, and the capacity of the test device is limited, which cannot meet the test requirements of composite insulators of various sizes; the existing prediction model for calculating the erosion of composite insulators by strong wind and sand has poor accuracy and cannot accurately predict the erosion of composite insulators. Therefore, it is very necessary to design a method for simulation and prediction analysis of composite insulators eroded by wind and sand. Summary of the invention

[0005] The object of the present invention is to provide a method and device for predicting the wind and sand erosion rate of a composite insulator, so as to predict the erosion condition of the composite insulator.

[0006] In a first aspect, an embodiment of the present invention provides a method for predicting the wind and sand erosion rate of a composite insulator, wherein the composite insulator is used for an ultra-high voltage transmission line, and the prediction method comprises: S100: establishing a three-dimensional numerical model of the composite insulator according to the physical parameters of the composite insulator, wherein the physical parameters include: data of an inlet, an outlet, a wall surface, and an edge of an shed of the composite insulator after grid division; S200: setting simulation parameters for the three-dimensional numerical model, wherein the simulation parameters include: airflow parameters, wind and sand parameters, and system parameters, wherein the airflow parameters include gas density, wind speed, and fluid viscosity, the wind and sand parameters include a diameter of sand and dust particles, and the system parameters include a windward angle of the composite insulator, injection source data, material data, and boundary condition data; S300: inputting the simulation parameters into a simulation software for simulation calculation to obtain an erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model; S400: based on the erosion rate V er and the simulation parameters to obtain a dimensionless group and perform dimensionality reduction processing to obtain a training data set; S500: based on the training data set, the pre-acquired first Gaussian model is corrected to obtain a second Gaussian model, and the erosion rate of the composite insulator is predicted based on the second Gaussian model to obtain a predicted erosion rate.

[0007] Furthermore, S100 includes: obtaining physical parameters of the composite insulator based on a preset grid; locally encrypting the grid of the shed edge of the composite insulator that is severely eroded by wind and sand; treating the airflow in the wind and sand as a continuous phase, and treating the dust as a discrete phase, wherein if the proportion of dust is less than the airflow, the influence of the dust on the continuous phase and the interaction between particles are ignored; and establishing the three-dimensional numerical model based on the continuous phase and the discrete phase.

[0008] Further, S200 includes: the airflow parameter and the wind and sand parameter are set according to actual conditions; the injection source data is the data of injection in the direction of the surface normal; the material data is the silicone rubber parameter; the boundary condition data is the boundary condition of the wall of the composite insulator, which is set to escape.

[0009] Further, S400 includes: the training data set is represented as follows: Among them, X is the input of the training data set, Y is the output of the training data set, and d p is the diameter of dust particles, D is the windward angle of the composite insulator, ρ f is the gas density, V f is the wind speed, μ f is the fluid viscosity, V er is the erosion rate, and n is the number of data sets in the training data set.

[0010] Furthermore, S500 includes: based on the training data set, modifying the hyperparameters in the covariance function of the first Gaussian model to obtain a second Gaussian model, wherein the hyperparameters include: length parameter, signal variance and noise variance; inputting the collected real-time data into the second Gaussian model, predicting the erosion rate of the composite insulator, and obtaining a predicted erosion rate.

[0011] In a second aspect, an embodiment of the present invention provides a device for predicting the wind and sand erosion rate of a composite insulator, wherein the composite insulator is used for an ultra-high voltage transmission line, and the prediction device comprises: a first prediction module, which is used to establish a three-dimensional numerical model of the composite insulator according to the physical parameters of the composite insulator, wherein the physical parameters include: data of the inlet, outlet, wall, and shed edge of the composite insulator after grid division; a second prediction module, which is used to set simulation parameters for the three-dimensional numerical model, wherein the simulation parameters include: airflow parameters, wind and sand parameters, and system parameters, wherein the airflow parameters include gas density, wind speed, and fluid viscosity, the wind and sand parameters include the diameter of sand and dust particles, and the system parameters include the windward angle of the composite insulator, injection source data, material data, and boundary condition data; a third prediction module, which is used to input the simulation parameters into the simulation software for simulation calculation to obtain the erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model; a fourth prediction module for predicting the erosion rate V er and the simulation parameters to obtain a dimensionless group and perform dimensionality reduction processing to obtain a training data set; a fifth prediction module is used to correct the pre-acquired first Gaussian model based on the training data set to obtain a second Gaussian model, and predict the erosion rate of the composite insulator based on the second Gaussian model to obtain a predicted erosion rate.

[0012] Furthermore, the first prediction module is also used to obtain physical parameters of the composite insulator based on a preset grid; locally densify the grid of the shed edge of the composite insulator that is severely eroded by wind and sand; treat the airflow in the wind and sand as a continuous phase, and the dust as a discrete phase, wherein if the proportion of dust is less than the airflow, the influence of the dust on the continuous phase and the interaction between particles are ignored; and establish the three-dimensional numerical model based on the continuous phase and the discrete phase.

[0013] Furthermore, the second prediction module is also used to set the airflow parameters and the wind and sand parameters according to actual conditions; the injection source data is data for injection in the direction of the surface normal; the material data is silicone rubber parameters; and the boundary condition data is the boundary condition of the wall of the composite insulator, which is set to escape.

[0014] Furthermore, the fourth prediction module is also used for the training data set to be expressed as follows: Among them, X is the input of the training data set, Y is the output of the training data set, and d p is the diameter of dust particles, D is the windward angle of the composite insulator, ρ f is the gas density, V f is the wind speed, μ f is the fluid viscosity, V er is the erosion rate, and n is the number of data sets in the training data set.

[0015] Furthermore, the fifth prediction module is also used to modify the hyperparameters in the covariance function in the first Gaussian model based on the training data set to obtain a second Gaussian model, wherein the hyperparameters include: length parameter, signal variance and noise variance; input the collected real-time data into the second Gaussian model, predict the erosion rate of the composite insulator, and obtain the predicted erosion rate.

[0016] Beneficial effects of the embodiments of the present invention:

[0017] The present invention discloses a method and device for predicting the wind and sand erosion rate of a composite insulator, comprising: establishing a three-dimensional numerical model of the composite insulator according to the physical parameters of the composite insulator; setting simulation parameters for the three-dimensional numerical model; inputting the simulation parameters into simulation software for simulation calculation, and obtaining the erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model, based on the erosion rate V er The simulation parameters are used to obtain a dimensionless group and perform dimensionality reduction processing to obtain a training data set; based on the training data set, the first Gaussian model obtained in advance is corrected to obtain a second Gaussian model, and the erosion rate of the composite insulator is predicted based on the second Gaussian model to obtain a predicted erosion rate. The present application can more accurately and quickly predict the rate at which a composite insulator is eroded by wind and sand. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of a method for predicting wind-sand erosion rate of composite insulators;

[0019] Figure 2 The figure is a flow chart of obtaining the second Gaussian model from the first Gaussian model. DETAILED DESCRIPTION

[0020] The present invention is further described in detail below through the accompanying drawings and specific embodiments.

[0021] like Figure 1 , Figure 2As shown, an embodiment of the present invention provides a method for predicting the wind and sand erosion rate of a composite insulator, wherein the composite insulator is used for an ultra-high voltage transmission line. The above prediction method mainly includes the following contents:

[0022] S100: Establishing a three-dimensional numerical model of the composite insulator according to physical parameters of the composite insulator, wherein the physical parameters include: data of an inlet, an outlet, a wall surface, and an edge of a shed of the composite insulator after meshing.

[0023] S100 includes: obtaining physical parameters of the composite insulator based on a preset grid; locally encrypting the grid at the edge of the shed of the composite insulator that is severely eroded by wind and sand; treating the airflow in the wind and sand as a continuous phase and the dust as a discrete phase, wherein if the proportion of dust is less than the airflow, the influence of the dust on the continuous phase and the interaction between particles are ignored. In this application, it is considered that the proportion of dust is less than the airflow. The three-dimensional numerical model is established based on the continuous phase and the discrete phase.

[0024] More specifically, S100 includes: obtaining the actual size of the composite insulator, establishing a three-dimensional numerical model of the composite insulator according to the actual size, namely, a composite insulator model, naming the inlet, outlet, wall, and composite insulator, and meshing the fluid calculation area, locally encrypting the mesh of the edge of the shed of the composite insulator that may be severely eroded, treating the gas phase in the wind and sand two-phase flow as a continuous phase, and treating the dust particles as a discrete phase, among which the dust particles account for a small proportion, and therefore, their influence on the continuous phase and the interaction between particles are ignored.

[0025] The continuity equation is established as:

[0026]

[0027] The momentum conservation equation is:

[0028]

[0029] In the formula, u is the time-averaged velocity of the airflow, x is the displacement of the airflow, ρ is the airflow density, and μ is the dynamic viscosity coefficient of the airflow. is the Reynolds stress component; p is the pressure of the fluid, x i and x j Represents the coordinate of the i-th spatial direction, which can be x, y, z, x j Represents the coordinate of the jth spatial direction, which can also be x, y, z, u i and u j is the fluid at x i and x j The component of velocity in the direction.

[0030] Due to the large curvature of the flow field around the composite insulator, the RNG k-ε turbulence model is selected based on Formula 2, and its control equation is:

[0031]

[0032] Where k is the turbulent kinetic energy, ε is the dissipation rate, μ e is the effective viscosity coefficient, μ e =μ+μ t , α k is the inverse of the Prandtl constant corresponding to the turbulent kinetic energy, α ε is the inverse of the Prandtl constant corresponding to the dissipation rate, C k is the generation term of turbulent kinetic energy k caused by the time-averaged velocity, and C1 and C2 are empirical constants.

[0033] Erosion rate V er The size is positively correlated with the particle diameter Dp, particle velocity v, collision angle θ, and particle hardness h, as follows:

[0034] V er ∝y(D p )y(v)y(θ)y(h) Formula 5;

[0035] Based on the actual needs of composite insulator erosion and wear simulation, in order to reduce the amount of calculation and obtain effective results, a custom erosion model is established as follows:

[0036]

[0037] Where, mp is the particle mass flow rate, unit is kg / s, set according to the actual dust parameters, f(D p ) is the particle size function, f(θ) is the impact angle function, and f n (v) is the velocity exponential function, A is the erosion wall area; Np represents the total number of particles.

[0038] The particle size function f(D p )for:

[0039] f(D p )=C(BH) -0.59 Formula 7;

[0040] Where C = 2.17 × 10 -7 , BH is the Brinell hardness of the material;

[0041] The impact angle function f(θ) is:

[0042] f(θ)=5.40θ-10.11θ 2 +10.93θ 3 -6.33θ 4+1.42θ 5 Formula 8;

[0043] Speed ​​exponential function f n (v) for;

[0044] f n (v) = V p n Formula 9;

[0045] At this point, through S100, a “three-dimensional numerical model” is obtained.

[0046] S200: Setting simulation parameters for the three-dimensional numerical model, wherein the simulation parameters include: airflow parameters, wind and sand parameters and system parameters, the airflow parameters include gas density, wind speed and fluid viscosity, the wind and sand parameters include dust particle diameter, and the system parameters include windward angle of composite insulator, injection source data, material data and boundary condition data.

[0047] S200 includes: the airflow parameters and the wind and sand parameters are set according to actual conditions; the injection source data is the data of injection in the direction of the surface normal; the material data is the silicone rubber parameters; setting the boundary condition data is setting the boundary condition of the wall, which is generally set to escape.

[0048] Specifically, in S200, the airflow parameters and wind and sand parameters can be set according to research needs or actual conditions such as the environment and region where the composite insulator is located.

[0049] More specifically, in S200, considering gravity, along the vertical downward direction of the composite insulator, the gravity acceleration g=9.81m / s 2 , open the discrete phase, open the interaction between fluid and discrete phase, open the erosion / deposition physical model, establish a surface injection source, select the particle diameter according to actual needs, set the random starting point of the particles, use the surface normal direction to inject, set the solid material to silicone rubber, set the fluid material to air, modify the density of the inert particles according to the actual situation of the dust particles, set the inlet type to velocity inlet, the inlet wind speed is the same as the initial velocity of the particle injection, the outlet type is pressure outlet, the wall boundary condition is set to escape, and the turbulence condition at the inlet is set to intensity and hydraulic diameter (Intensity and Hydraulic Diameter), the hydraulic diameter D hd The calculation formula is:

[0050]

[0051] Where S is the inlet area of ​​the fluid domain, and C is the circumference of the inlet of the fluid domain.

[0052] The discrete phase rebound coefficients are set for the composite insulator surface, including the normal rebound coefficient and the tangential rebound coefficient, which are:

[0053] Normal rebound coefficient: ε1 = 0.993-0.0307θ+0.000475θ 2 -2.61×10 -6 θ 3 Formula 11;

[0054] Tangential rebound coefficient: ε2 = 0.998-0.029θ+0.000643θ 2 -3.56×10 -6 θ 3 Formula 12;

[0055] Where the collision angle θ is in radians.

[0056] The above S200 is the process of setting parameters of the simulation.

[0057] S300: Input the simulation parameters into the simulation software for simulation calculation to obtain the erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model.

[0058] On this basis, we continue to use the erosion rate V er Calculate the total mass loss E due to wind and sand erosion total .

[0059] Specifically, S300 includes determining a calculation method and parameters, performing simulation calculations, and obtaining an erosion rate V of the composite insulator under wind and sand erosion. er and the total mass loss due to wind and sand erosion E total Specifically, in order to reduce the storage space occupied, the result file is automatically saved every 10 iterations, and the entry inlet is used as a reference for initialization. After the initialization is completed, the calculation method is set, the calculation type is Fixed, the method is User-Specified, and the time step should be set in accordance with: its value should be one to two orders of magnitude smaller than the ratio of the characteristic length to the characteristic velocity. The present invention sets the time step t to 0.001s, the number of time steps n to 2000 steps, and the maximum number of iterations / time steps to 20. After iterative calculation, the erosion simulation results are obtained, and the DPMErosion discrete phase model is used to color the variables, that is, to display the erosion result cloud map under transient conditions, and then analyze the erosion distribution under the actual erosion test of the composite insulator, extract and save the erosion result data, and obtain the erosion rate V of each eroded part of the composite insulator from the simulation results. er , unit is kg / (m 2 ·s), which is the mass of material lost by erosion per unit area per unit time.

[0060] According to the composite insulator erosion rate V obtained from the above simulation er , the material loss mass of a single grid under wind and sand erosion is calculated as:

[0061]

[0062] Where E is the material loss mass of a single grid of composite insulator under wind and sand erosion, t is the time step, n is the number of time steps, A is the area of ​​a single grid, and Veri is the erosion rate solved in a single transient calculation step.

[0063] In the erosion simulation, the composite insulator model is divided into a grid number m, and the material loss mass of a single grid under wind and sand erosion is E. j , then the total mass loss of the composite insulator due to wind and sand erosion within the preset time is:

[0064]

[0065] Among them, the total mass loss can characterize the wear condition of the composite insulator.

[0066] Specifically, S300 inputs the wind and sand related data into the simulation software to obtain the erosion rate V er and the total amount of mass loss during the process.

[0067] S400: Based on the erosion rate V er The simulation parameters are obtained into a dimensionless group and subjected to dimensionality reduction processing to obtain a training data set.

[0068] The training data set is represented as follows:

[0069]

[0070] Wherein, X is the input of the training data set, Y is the output of the training data set, dp is the dust particle size, D is the windward angle of the composite insulator, ρ f is the gas density, V f is the wind speed, μ f is the fluid viscosity, V er is the erosion rate, and n is the number of data sets in the training data set.

[0071] S400 and 500 are generally based on the erosion rate calculation results V er, taking the composite insulator windward angle, dust particle diameter, gas density, wind speed and fluid viscosity as input parameters, the input parameters are combined into dimensionless groups for dimension reduction. Based on the Gaussian process regression method, the composite insulator erosion rate under any dust particle size, composite insulator windward angle, gas density and wind speed conditions is predicted, and the total mass loss under this condition can be obtained, which is specifically:

[0072] Based on the above S300, the composite insulator erosion rate V er , the erosion rate of composite insulators under arbitrary conditions is predicted.

[0073] In order to establish the relationship between the output variables and the input variables, a training data set T = {T i |i=1,2,...,n}={(x i ,y i )i=1,2,…,n}, where x i ∈R m , m≥1 is the m-dimensional input vector, y i is the corresponding output, x i The vector is represented by the matrix X = {x i} i＝1:n The input matrix y is expressed in the form of i The values ​​form the output vector;

[0074] In the prediction of erosion rate, the definitions are as follows:

[0075] Among them, d p is the diameter of dust particles, D is the windward angle of the composite insulator, ρ f is the gas density, V f is the wind speed, μ f is the fluid viscosity, V er is the erosion rate.

[0076] Specifically, S400 is the process of obtaining the training data set, that is, combining the input and output of the simulation software, and then making dimensionless groups and performing dimensionality reduction processing. The input parameters (or input data) of the training data set are formula 15, and the output parameters (or output data) are formula 16.

[0077] S500: Based on the training data set, a pre-acquired first Gaussian model is modified to obtain a second Gaussian model, and based on the second Gaussian model, an erosion rate of the composite insulator is predicted to obtain a predicted erosion rate.

[0078] S500 predicts the erosion rate of composite insulators based on the defined X and Y and the Gaussian process regression method. S500 includes:

[0079] S501: Based on the training data set, the first Gaussian model obtained in advance is modified to obtain a second Gaussian model, such as Figure 2 shown.

[0080] Specifically, S501 is: first, based on the S300 simulation method, the erosion rate V of the composite insulator is obtained. er Then, based on the S400 method, the erosion rate V er and simulation parameters into a dimensionless group and perform dimensionality reduction processing to obtain a training data set. Finally, the training data set is substituted into the first Gaussian model, and the hyperparameters of the covariance function in the first Gaussian model are modified to obtain a second Gaussian model, wherein the hyperparameters include: length parameter, signal variance and noise variance.

[0081] S502: Input the collected real-time data including dust particle size, windward angle of composite insulator, gas density, wind speed, etc. into the second Gaussian model, predict the erosion rate of the composite insulator, obtain the predicted erosion rate, and then obtain the total mass loss.

[0082] More specifically, the steps of training the Gaussian model in S501 include:

[0083] S501 -1. Determine the first covariance function between any two input data x, that is, Formula 17.

[0084] Specifically, in the first Gaussian model obtained in advance, the covariance function is one of the important indicators in statistical analysis. It is a standard for measuring the mutual influence between input data points. This function represents the coordinated change rate between two variables. The covariance function is the square exponent of the x and x' variables and is defined as:

[0085]

[0086] In the formula, represents the maximum value of the signal variance, l is the length parameter of the covariance function, the length parameter l determines whether two input features x that are a certain distance apart are considered to be close, and x and x' represent any two different input data in x of Formula 15.

[0087] S501-2. According to the Gaussian distribution and the noise distribution, the first covariance function is corrected to a second covariance function, that is, Formula 18 is obtained.

[0088] Specifically, each y i The value can be considered as a function f(x i ) and the random variable simulating noise (ε i ) and y i =f(x i )+ε i, assuming that the noise follows a Gaussian distribution with a mean of zero, N represents Gaussian distribution, p(ε i |x) is the probability of noise given x. Considering the noise distribution, the first covariance function of Formula 17 can be written as Formula 18, which is the second covariance function.

[0089]

[0090] Where δ(x,x') is the preset Kronecker function, σ f (Signal variance) represents the variance of the signal in the first Gaussian model; σ n (Noise Variance) represents the noise variance in the observed data (ie, x in Formula 15).

[0091] S501 - 3. Based on the second covariance function, calculate the covariance matrix K for all different combinations of training data points (input data x), that is, Formula 19.

[0092]

[0093] S501-4. Based on the covariance matrix K, the second covariance is calculated (i.e., Formula 18) by super parameter Θ={σ f ,σ n ,l} optimization.

[0094] Considering that the model prediction mainly depends on the parameters of the covariance function, the correct choice of these parameters has an important impact on the accuracy of the model. These parameters can be expressed as a vector Θ = {σ f ,σ n ,l}, which is called a hyperparameter.

[0095] The optimal values ​​of these parameters are obtained when deriving the maximum of the probability distribution of P(y|X,Θ), where P(y|X,Θ) is the probability of y given X, Θ. According to Bayes' theorem, P(y|X,Θ) is maximized when the following log-likelihood function is maximized:

[0096] logP(y|X,Θ)=-0.5y T K -1 y-0.5log|K|-0.5nlog2π Formula 20;

[0097] Among them, Θ: hyperparameters of the model, including σf, σn, l; K: covariance matrix; K -1 : The inverse of the covariance matrix K; y T K - 1 y: is a quadratic form that represents the degree of fit between the output of the data point and the prediction of the kernel function K. log|K|:

[0098] The natural logarithm of the determinant of the kernel function matrix K, which measures the volume of the kernel function matrix and is related to the correlation between data points. n: the number of data points. log2π: a constant term used for normalization. K is the result of formula 19.

[0099] The conjugate gradient method can be used to optimize hyperparameters using the partial derivatives of the log-likelihood function with respect to Θ:

[0100]

[0101] Where α = K -1 y, Θ j It is a hyperparameter element. By introducing this optimization method, the accuracy of the prediction model can be improved and the problem of excessive deviation of the prediction results can be solved. Θ contains three hyperparameters σf, σn2 and l, so Θ1 represents σ f , Θ2 represents σ n , Θ3 represents l.

[0102] At this point, the optimized hyperparameters can be obtained based on Formula 20-Formula 21.

[0103] S501-5. The optimized hyperparameter Θ={σ f ,σ n ,l} is substituted into formula 18 to obtain the optimized covariance function, and then the second Gaussian model is determined based on the optimized covariance function.

[0104] Furthermore, the present application can also verify the second Gaussian model, and the process is as follows:

[0105] In simple terms, the verification process is as follows: collect a verification set, including x* and y (i.e. Figure 2 Xtext, Ytext), input x* into the second Gaussian model to obtain the predicted value y*, compare the y* output by the second Gaussian model with the standard y in the validation set, and infer the accuracy of the second Gaussian model by the size of p.

[0106]

[0107] Where N represents Gaussian distribution, K represents the covariance matrix obtained based on x* in the validation set (the steps are the same as formula 15-formula 19), and K * K -1 y is the mean of the Gaussian distribution; K ** -K * K -1 K * T is the variance of the Gaussian distribution (indicating the uncertainty of the prediction), K -1 represents the inverse of the covariance matrix of the validation set, K *T Represents the transpose of the covariance matrix of the validation set.

[0108] S502 includes:

[0109] The collected real-time data is input into the second Gaussian model to predict the erosion rate of the composite insulator to obtain the predicted erosion rate, and then the total mass loss of the current composite insulator is obtained according to Formula 13 and Formula 14. The real-time collected data includes one or more of the following data: physical parameters include: data of the composite insulator inlet, outlet, wall, and shed edge after grid division; airflow parameters include gas density, wind speed, and fluid viscosity; the wind and sand parameters include the diameter of sand and dust particles; system parameters include the windward angle of the composite insulator, injection source data, material data, and boundary condition data.

[0110] This embodiment can reduce the test cost, obtain reasonable test conditions, accelerate the test cycle, and solve the defect that the existing model cannot predict the erosion rate according to the actual parameters of the composite insulator.

[0111] The present invention discloses a method for predicting the wind and sand erosion rate of a composite insulator, comprising: establishing a three-dimensional numerical model of the composite insulator according to the physical parameters of the composite insulator; setting simulation parameters for the three-dimensional numerical model; inputting the simulation parameters into simulation software for simulation calculation, and obtaining the erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model, based on the erosion rate V er The simulation parameters are used to obtain a dimensionless group and perform dimensionality reduction processing to obtain a training data set; based on the training data set, the first Gaussian model obtained in advance is corrected to obtain a second Gaussian model, and the erosion rate of the composite insulator is predicted based on the second Gaussian model to obtain a predicted erosion rate. The present application can more accurately and quickly predict the rate at which a composite insulator is eroded by wind and sand.

[0112] Embodiment 2

[0113] Corresponding to the prediction method of the wind and sand erosion rate of the composite insulator of the first embodiment, the embodiment of the present invention provides a prediction device for the wind and sand erosion rate of the composite insulator, the composite insulator is used for the ultra-high voltage transmission line, and the prediction device comprises:

[0114] The first prediction module is used to establish a three-dimensional numerical model of the composite insulator according to the physical parameters of the composite insulator, wherein the physical parameters include: data of the composite insulator inlet, outlet, wall surface, and shed edge after grid division; the second prediction module is used to set simulation parameters for the three-dimensional numerical model, wherein the simulation parameters include: airflow parameters, wind and sand parameters, and system parameters, wherein the airflow parameters include gas density, wind speed, and fluid viscosity, the wind and sand parameters include the diameter of sand and dust particles, and the system parameters include the windward angle of the composite insulator, injection source data, material data, and boundary condition data; the third prediction module is used to input the simulation parameters into the simulation software for simulation calculation to obtain the erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model; a fourth prediction module for predicting the erosion rate V er and the simulation parameters to obtain a dimensionless group and perform dimensionality reduction processing to obtain a training data set; a fifth prediction module is used to correct the pre-acquired first Gaussian model based on the training data set to obtain a second Gaussian model, and predict the erosion rate of the composite insulator based on the second Gaussian model to obtain a predicted erosion rate.

[0115] The first prediction module is also used to obtain the physical parameters of the composite insulator based on a preset grid; locally encrypt the grid at the edge of the shed of the composite insulator that is severely eroded by wind and sand; treat the airflow in the wind and sand as a continuous phase, and the dust as a discrete phase, wherein if the proportion of dust is less than the airflow, the influence of the dust on the continuous phase and the interaction between particles are ignored; and establish the three-dimensional numerical model based on the continuous phase and the discrete phase.

[0116] The second prediction module is also used to set the airflow parameters and the wind and sand parameters according to actual conditions; the injection source data is the data of injection in the direction of the surface normal; the material data is the silicone rubber parameters; the boundary condition data is the boundary condition of the wall of the composite insulator, which is set to escape.

[0117] The fourth prediction module, also used for the training data set, is expressed as follows: Among them, X is the input of the training data set, Y is the output of the training data set, and d p is the diameter of dust particles, D is the windward angle of the composite insulator, ρ f is the gas density, V f is the wind speed, μ f is the fluid viscosity, V er is the erosion rate, and n is the number of data sets in the training data set.

[0118] The fifth prediction module is also used to modify the hyperparameters in the covariance function of the first Gaussian model based on the training data set to obtain a second Gaussian model, wherein the hyperparameters include: length parameter, signal variance and noise variance; input the collected real-time data into the second Gaussian model, predict the erosion rate of the composite insulator, and obtain the predicted erosion rate.

[0119] The beneficial effects of the device for predicting wind-sand erosion rate of composite insulators of the present invention are the same as those of the above-mentioned prediction method embodiment, and will not be described in detail in this embodiment.

[0120] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

[0121] The above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.

Claims

1. A method for predicting the wind and sand erosion rate of a composite insulator, wherein the composite insulator is used for an ultra-high voltage transmission line, characterized in that: The prediction method comprises: S100: establishing a three-dimensional numerical model of the composite insulator according to physical parameters of the composite insulator, wherein the physical parameters include: data of an inlet, an outlet, a wall surface, and an edge of a shed of the composite insulator after meshing; S200: setting simulation parameters for the three-dimensional numerical model, wherein the simulation parameters include: airflow parameters, wind and sand parameters and system parameters, the airflow parameters include gas density, wind speed and fluid viscosity, the wind and sand parameters include dust particle diameter, and the system parameters include windward angle of composite insulator, injection source data, material data and boundary condition data; S300: Input the simulation parameters into the simulation software for simulation calculation to obtain the erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model; S400: Based on the erosion rate V er and the simulation parameters to obtain a dimensionless group and perform dimensionality reduction processing to obtain a training data set; S500: Based on the training data set, a pre-acquired first Gaussian model is modified to obtain a second Gaussian model, and based on the second Gaussian model, an erosion rate of the composite insulator is predicted to obtain a predicted erosion rate.

2. The method for predicting wind and sand erosion rate of composite insulators according to claim 1, characterized in that: S100 includes: Based on the preset grid, the physical parameters of the composite insulator are obtained; The mesh is locally densified at the edge of the shed of the composite insulator which is severely eroded by wind and sand; Treat the airflow in the wind and sand as a continuous phase and the dust as a discrete phase. If the proportion of dust is smaller than the airflow, the influence of the dust on the continuous phase and the interaction between particles are ignored. The three-dimensional numerical model is established based on the continuous phase and the discrete phase.

3. The method for predicting wind and sand erosion rate of composite insulators according to claim 2, characterized in that: S200 includes: The airflow parameters and the wind and sand parameters are set according to actual conditions; The spray source data is data sprayed in the direction of the surface normal; The material data are silicone rubber parameters; The boundary condition data is the boundary condition of the wall of the composite insulator, which is set to escape.

4. The method for predicting wind and sand erosion rate of composite insulators according to claim 3, characterized in that: S400 includes: The training data set is represented as follows: Where X is the input of the training data set, Y is the output of the training data set, and d p is the diameter of dust particles, D is the windward angle of the composite insulator, ρ f is the gas density, V f is the wind speed, μ f is the fluid viscosity, V er is the erosion rate, and n is the number of data sets in the training data set.

5. The method for predicting wind and sand erosion rate of composite insulators according to claim 4, characterized in that: S500 includes: Based on the training data set, the hyperparameters in the covariance function of the first Gaussian model are modified to obtain a second Gaussian model, wherein the hyperparameters include: a length parameter, a signal variance, and a noise variance; The collected real-time data is input into the second Gaussian model to predict the erosion rate of the composite insulator to obtain a predicted erosion rate.

6. A device for predicting wind and sand erosion rate of composite insulators, wherein the composite insulators are used in ultra-high voltage transmission lines, characterized in that: The prediction device comprises: A first prediction module is used to establish a three-dimensional numerical model of the composite insulator according to physical parameters of the composite insulator, wherein the physical parameters include: data of an inlet, an outlet, a wall surface, and an edge of a shed of the composite insulator after meshing; a second prediction module, used for setting simulation parameters for the three-dimensional numerical model, wherein the simulation parameters include: airflow parameters, wind and sand parameters and system parameters, the airflow parameters include gas density, wind speed and fluid viscosity, the wind and sand parameters include dust particle diameter, and the system parameters include windward angle of composite insulator, injection source data, material data and boundary condition data; The third prediction module is used to input the simulation parameters into the simulation software for simulation calculation to obtain the erosion rate V of the composite insulator under wind and sand erosion. er , wherein the simulation software is provided with the three-dimensional numerical model; The fourth prediction module is used to predict the erosion rate V er and the simulation parameters to obtain a dimensionless group and perform dimensionality reduction processing to obtain a training data set; The fifth prediction module is used to modify the pre-acquired first Gaussian model based on the training data set to obtain a second Gaussian model, and predict the erosion rate of the composite insulator based on the second Gaussian model to obtain a predicted erosion rate.

7. The device for predicting wind and sand erosion rate of composite insulators according to claim 6, characterized in that: The first prediction module is further used to obtain physical parameters of the composite insulator based on a preset grid; The mesh is locally densified at the edge of the shed of the composite insulator which is severely eroded by wind and sand; Treat the airflow in the wind and sand as a continuous phase and the dust as a discrete phase. If the proportion of dust is smaller than the airflow, the influence of the dust on the continuous phase and the interaction between particles are ignored. The three-dimensional numerical model is established based on the continuous phase and the discrete phase.

8. The device for predicting wind and sand erosion rate of composite insulators according to claim 7, characterized in that: The second prediction module is also used to set the airflow parameter and the wind and sand parameter according to actual conditions; The spray source data is data sprayed in the direction of the surface normal; The material data are silicone rubber parameters; The boundary condition data is the boundary condition of the wall of the composite insulator, which is set to escape.

9. The device for predicting wind and sand erosion rate of composite insulators according to claim 8, characterized in that: The fourth prediction module is also used for the training data set to be expressed as follows: Where X is the input of the training data set, Y is the output of the training data set, and d p is the diameter of dust particles, D is the windward angle of the composite insulator, ρ f is the gas density, V f is the wind speed, μ f is the fluid viscosity, V er is the erosion rate, and n is the number of data sets in the training data set.

10. The device for predicting wind and sand erosion rate of composite insulators according to claim 9, characterized in that: The fifth prediction module is further used to modify the hyperparameters in the covariance function in the first Gaussian model based on the training data set to obtain a second Gaussian model, wherein the hyperparameters include: a length parameter, a signal variance, and a noise variance; The collected real-time data is input into the second Gaussian model to predict the erosion rate of the composite insulator to obtain a predicted erosion rate.

Images