A method for predicting the internal corrosion rate of a gas transmission pipeline

Abstract

The present invention provides a method for predicting the corrosion rate in a gas pipeline, comprising: inputting acquired data to be tested into a trained corrosion rate prediction model to obtain a corrosion rate prediction result; a method for training the corrosion rate prediction model comprising: obtaining a test data set consisting of pipeline test sample data; screening the test data set based on a gray correlation analysis method to obtain an environmental impact factor data set, and constructing a sample data set; training a support vector machine model based on the sample data set, and optimizing the support vector machine model using a vulture search algorithm to obtain a corrosion rate prediction model. The present invention can effectively address the problems of existing gas pipeline corrosion rate prediction models, such as large basic data requirements and insufficient prediction reliability, and obtain a prediction model with higher accuracy, smaller errors, and better fit, thereby significantly reducing the economic expenditure of gas pipelines and improving the safe operation and maintenance of gas pipelines.

Description

Technical Field

[0001] The present invention relates to the technical field of gas pipeline corrosion and protection, and in particular to a method for predicting the corrosion rate in a gas pipeline. Background Art

[0002] Predicting corrosion rates within gas pipelines has long been a hot topic of research for scholars both domestically and internationally. Research on corrosion rate prediction models is of great significance to the basic design of pipelines and their future safe operation. Currently, the main methods for predicting corrosion rates within gas pipelines include empirical models, semi-empirical models, grey theory, and neural network models.

[0003] However, some neural network modeling processes still have disadvantages such as large amount of required information and low learning rate, which results in poor prediction effect and slow prediction rate of corrosion rate. Summary of the Invention

[0004] In order to overcome the deficiencies of the prior art, the present invention aims to provide a method for predicting the corrosion rate in a gas pipeline.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for predicting corrosion rate in a gas pipeline, comprising:

[0007] Obtain the data to be tested;

[0008] Inputting the data to be detected into a trained corrosion rate prediction model to obtain a corrosion rate prediction result;

[0009] The training method of the corrosion rate prediction model includes:

[0010] Acquire a detection data set consisting of pipeline detection sample data; the pipeline detection sample data includes relevant environmental influencing factors and corrosion rate detection results on internal corrosion behavior of the gas pipeline;

[0011] The detection data set is screened based on a grey relational analysis method to obtain an environmental influencing factor data set, and a sample data set is constructed based on the environmental influencing factor data set and the corrosion rate detection results;

[0012] A support vector machine model is trained according to the sample data set, and the support vector machine model is optimized using a vulture search algorithm to obtain the corrosion rate prediction model.

[0013] Preferably, after inputting the sample data set into a support vector machine model for training and optimizing the support vector machine model using a vulture search algorithm to obtain the corrosion rate prediction model, the method further includes:

[0014] An error calculation is performed according to the prediction result of the corrosion rate prediction model to obtain a calculation result; the calculation result includes: a root mean square error, a mean absolute error and a correlation coefficient.

[0015] Model evaluation is performed based on the calculation results.

[0016] Preferably, the environmental influencing factors include pipeline pressure, temperature, CO2 partial pressure, liquid holdup, fluid flow rate and pH.

[0017] Preferably, the grey relational analysis method is used to screen the detection data set to obtain a sample data set, including:

[0018] Determine a parent sequence consisting of corrosion rate test results and a child sequence consisting of environmental influencing factors;

[0019] Normalizing each index in the parent sequence and the child sequence to obtain a normalized index value of the parent sequence and a normalized index value of the child sequence;

[0020] Calculating a correlation coefficient between each of the normalized index values in the subsequence and the normalized index value of the parent sequence;

[0021] Calculating the correlation between each environmental influencing factor and the corrosion rate detection result according to each correlation coefficient;

[0022] The correlation degrees are sorted, and the environmental influencing factors with the smallest correlation degree are removed according to the sorting result, and the remaining environmental influencing factors are determined as the environmental influencing factor data set.

[0023] Preferably, the support vector machine model is trained according to the sample data set, and the support vector machine model is optimized using a vulture search algorithm to obtain the corrosion rate prediction model, including:

[0024] Based on the nonlinear mapping relationship, determining an initial regression function according to the sample data set;

[0025] Based on multivariate statistical analysis theory, determining an SVM regression function and constraints of the SVM regression function according to the initial regression function;

[0026] Using the Lagrange multiplier method to perform constrained optimization on the duality variables of the SVM regression function to obtain a final regression function;

[0027] Constructing the vector machine model using the final regression function;

[0028] The penalty factor and kernel function of the vector machine model are optimized according to the vulture search algorithm to obtain the corrosion rate prediction model.

[0029] Preferably, after constructing the sample data set according to the environmental influencing factor data set and the corrosion rate detection result, the method further includes:

[0030] The sample data set is divided into a training set and a test set according to a preset ratio; the training set is used to construct the support vector machine model; and the test set is used to test the corrosion rate prediction model.

[0031] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0032] The present invention provides a method for predicting the corrosion rate in a gas pipeline, comprising: obtaining data to be tested; inputting the data to be tested into a trained corrosion rate prediction model to obtain a corrosion rate prediction result; a method for training the corrosion rate prediction model comprising: obtaining a test data set consisting of pipeline test sample data; the pipeline test sample data comprising environmental factors related to the internal corrosion behavior of the gas pipeline and corrosion rate test results; screening the test data set based on a grey correlation analysis method to obtain an environmental factor data set, and constructing a sample data set based on the environmental factor data set and the corrosion rate test results; training a support vector machine model based on the sample data set, and optimizing the support vector machine model using a vulture search algorithm to obtain the corrosion rate prediction model. The present invention can effectively solve the problems of existing gas pipeline corrosion rate prediction models in requiring a large amount of basic data and having insufficient prediction reliability, and obtain a prediction model with higher accuracy, smaller error, and better fitting, thereby significantly reducing the economic expenditure of the gas pipeline and improving the safe operation and maintenance of the gas pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, 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 paying any creative work.

[0034] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0035] Figure 2 Flowchart of the GRA-BES-SVM prediction model provided by an embodiment of the present invention;

[0036] Figure 3 A structural diagram of the GRA-SVM prediction model provided by an embodiment of the present invention;

[0037] Figure 4 A convergence curve diagram of the BES-SVM prediction model provided by an embodiment of the present invention;

[0038] Figure 5 A comparison chart of the test set prediction results of the four models provided in the embodiments of the present invention. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0040] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0041] The terms "first," "second," "third," and "fourth," etc., in the specification, claims, and drawings of this application are used to distinguish between different items, not to describe a particular order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a statement that a sequence of steps, a process, or a method is included is not limited to the listed steps but may optionally include steps not listed, or may optionally include other steps inherent to the process, method, product, or apparatus.

[0042] The purpose of the present invention is to provide a method for predicting the corrosion rate in gas pipelines, which can effectively solve the problems of existing gas pipeline corrosion rate prediction models requiring a large amount of basic data and insufficient prediction reliability, and obtain a prediction model with higher accuracy, smaller error and better fitting degree, thereby greatly reducing the economic expenditure of gas pipelines and improving the safe operation and maintenance of gas pipelines.

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] Figure 1 A flow chart of the method provided in the embodiment of the present invention is shown in FIG. Figure 1 As shown, this embodiment provides a method for predicting the corrosion rate in a gas pipeline, comprising:

[0045] Step 100: Obtaining data to be detected;

[0046] Step 200: inputting the data to be detected into a trained corrosion rate prediction model to obtain a corrosion rate prediction result;

[0047] The training method of the corrosion rate prediction model includes:

[0048] Step 201: Acquire a test data set consisting of pipeline test sample data; the pipeline test sample data includes relevant environmental factors affecting internal corrosion behavior of the gas pipeline and corrosion rate test results;

[0049] Step 202: screening the detection data set based on a Grey Relation Analysis (GRA) method to obtain an environmental influencing factor data set, and constructing a sample data set based on the environmental influencing factor data set and the corrosion rate detection results;

[0050] Step 203: training a support vector machine (SVM) model according to the sample data set, and optimizing the SVM model using a bald eagle search (BES) algorithm to obtain the corrosion rate prediction model.

[0051] Figure 2 The flowchart of the GRA-BES-SVM prediction model provided by the embodiment of the present invention is as follows: Figure 2 As shown, the implementation process in this embodiment is as follows:

[0052] Process 1: Obtain the relevant environmental influencing factors and actual corrosion rate test results for the internal corrosion behavior of the gas pipeline to be evaluated.

[0053] Process 2: Taking the gas pipeline as the research object, a data model is established by combining the detection data of all environmental factors related to internal corrosion and the actual corrosion rate values. Then, through the GRA (grey relational analysis) method, the influencing factors are ranked according to their importance to the actual corrosion rate, and a data set of environmental factors with a correlation coefficient of not less than 0.7 is screened.

[0054] Step 3: Based on the sample data set established in Step 2, divide the training set and test set into a ratio of 0.7. Substitute the divided training set and test set into the established SVM (support vector machine) model to obtain the initialized training set and test set. Figure 3 This is the structure diagram of the GRA-SVM prediction model.

[0055] Process 4: c and g obtained by BES algorithm optimization are substituted into the initialized SVM model to establish a BES-SVM corrosion rate prediction model.

[0056] Process 5: Evaluate the model prediction effect based on relevant evaluation indicators.

[0057] Preferably, the environmental influencing factors include pipeline pressure, temperature, CO2 partial pressure, liquid holdup, fluid flow rate and pH.

[0058] Optionally, in this embodiment, the correlation coefficient between the environmental influencing factors and the actual corrosion rate is calculated by using a grey correlation analysis method, as shown in Table 1.

[0059] Table 1 Correlation between influencing factors and actual corrosion rate

[0060] <![CDATA[Influencing factor x i > <![CDATA[Degree of association R ij > temperature 0.7265 pressure 0.8004 Liquid holdup 0.7891 <![CDATA[Partial pressure of CO2]]> 0.8002 Fluid flow rate 0.6313 pH 0.8046

[0061] Preferably, the detection data set is screened based on the grey relational analysis method to obtain an environmental influencing factor data set, including:

[0062] Determine a parent sequence consisting of corrosion rate test results and a child sequence consisting of environmental influencing factors;

[0063] Normalizing each index in the parent sequence and the child sequence to obtain a normalized index value of the parent sequence and a normalized index value of the child sequence;

[0064] Calculating a correlation coefficient between each of the normalized index values in the subsequence and the normalized index value of the parent sequence;

[0065] Calculating the correlation between each environmental influencing factor and the corrosion rate detection result according to each correlation coefficient;

[0066] The correlation degrees are sorted, and the environmental influencing factors with the smallest correlation degree are removed according to the sorting result, and the remaining environmental influencing factors are determined as the environmental influencing factor data set.

[0067] Specifically, in this embodiment, the actual corrosion rate is used as the output variable of the data set. The correlation between the environmental influencing factors and the actual corrosion rate is calculated using the grey correlation analysis method. The influencing factors with a correlation coefficient of not less than 0.7 are selected as the input variables of the data set, including the following steps:

[0068] Determine the parent sequence x consisting of actual corrosion rates j (j=1, 2, 3, ..., p) and the subsequence x consisting of influencing factors i (i=1, 2, 3, ..., n):

[0069] Parent sequence: x j ={x j (k)}={x j (1), x j (2), x j (3),…,x j (m)};

[0070] Subsequence: x i ={x i (k)}={x i (1), x i (2), x i (3),…,x i (m)};

[0071] Normalize the various indicators of the parent sequence and the child sequence:

[0072]

[0073] Calculate the correlation coefficient between each indicator in the subsequence and the parent sequence:

[0074]

[0075] Calculate the correlation between the influencing factors and the actual corrosion rate:

[0076]

[0077] Sort the correlations, remove the factors with the smallest correlation according to the sorting results, and determine the remaining environmental factors as input variables.

[0078] Preferably, a support vector machine model is trained according to the sample data set, and the support vector machine model is optimized using a vulture search algorithm to obtain the corrosion rate prediction model, including:

[0079] Based on the nonlinear mapping relationship, determining an initial regression function according to the sample data set;

[0080] Based on multivariate statistical analysis theory, determining an SVM regression function and constraints of the SVM regression function according to the initial regression function;

[0081] The Lagrange multiplier method is used to optimize the duality variable constraints of the SVM regression function to obtain a final regression function;

[0082] Constructing the vector machine model using the final regression function;

[0083] The penalty factor and kernel function of the vector machine model are optimized according to the vulture search algorithm to obtain the corrosion rate prediction model.

[0084] Optionally, in this embodiment, the data in process 2 is substituted into the SVM model to initialize the data. The support vector machine model can be used to perform nonlinear mapping. Mapping the sample space to a high-dimensional or even infinite-dimensional feature space (Hilbert space) allows linear learning machine methods to be applied in the feature space to solve high-dimensional nonlinear classification and regression problems in the sample space. The specific steps are as follows:

[0085] Assume that the sample is (x1, x2, x3, ..., x n )∈R n , (y1, y2, y3, ..., y n ), x∈R is the input parameter, y∈R is the output parameter, n is the number of samples, the transformation process of SVM is to establish a nonlinear mapping Φ, mapping the data X to a high-dimensional feature space, and the regression function is:

[0086]

[0087] According to the theory of multivariate statistical analysis, the SVM regression function can be determined by minimizing the following target number:

[0088]

[0089] The constraints are:

[0090]

[0091] Lagrange multiplier method for dual variable constrained optimization:

[0092]

[0093]

[0094] The constraints are:

[0095]

[0096] Where: a i , a i * is the Lagrange coefficient; α i , β i is the Lagrangian operator; b is the critical value; ω is the weight vector; ξ, ξ * is a non-negative slack variable; C is a penalty variable; ε is an insensitive loss function parameter; K(x i , x j ) is the kernel function of SVM.

[0097] The kernel function is to calculate the inner product function of two vectors in the implicit mapping space, and the vector in the low-dimensional space is transformed to obtain the inner product value of the vector in the high-dimensional space. The Gaussian radial basis function is selected, that is: K(x i , x j )=exp(-g|x i -x j | 2 ), g is the parameter width of the kernel function.

[0098] The final regression function is:

[0099]

[0100] For a new input parameter x, the corresponding output value can be calculated using this formula.

[0101] Furthermore, in process 4 of this embodiment, the vulture search algorithm is used to optimize the SVM model. The specific steps are as follows:

[0102] BES search algorithm process:

[0103] Step 1. Initialize the vulture algorithm parameters and initialize the population

[0104] Step 2. Calculate the fitness value

[0105] Step 3. The vulture selects the search space. The vulture randomly selects the search area and determines the best search position by judging the number of prey to facilitate the search of prey. The vulture position P in this stage i，new The update is determined by multiplying the prior information of the random search by α, and the mathematical model of this behavior is described as:

[0106] P i，new =P best +α*r(P meαn -P i );

[0107] Where: α represents the control position change parameter, and its range is (1.5, 2); r is a random number between (0, 1); P best The best search location determined for the current condor search; P mmean is the average distribution position of vultures after the previous search; P i is the i-th vulture.

[0108] Step 4: The vulture searches for prey in the search space: The vulture flies in a spiral pattern within the selected search space, accelerating the search process and finding the optimal diving capture position. The spiral flight mathematical model uses polar coordinate equations to update the position, using the following formula:

[0109] θ(i)=α*π*rand;

[0110] r(i)=θ(i)+R*rand;

[0111] xr(i)=r(i)*sin(θ(i));

[0112] xr(i)=r(i)*coS(θ(i));

[0113] x(i) = xr(i) / max(|xr|);

[0114] y(i)=yr(i) / max(|yr|);

[0115] Where: θ i and r(i) are the polar angle and polar radius of the spiral equation, respectively; a and R are parameters that control the spiral trajectory, with a range of (0, 5) and (0.5, 2), respectively; rand is a random number in the range (0, 1), and x(i) and y(i) represent the vulture's position in polar coordinates, both with a value of (-1, 1). The vulture's position is updated as follows:

[0116] P i，new =P i +x(i)*(P i -P mean )+y(i)*(P i -P i+1 );

[0117] Step 5: The vulture dives: The vulture quickly dives from the optimal position in the search space to the target prey. The other individuals in the population also move to the optimal position and attack the prey. The motion state is still described by the polar coordinate equation as follows: Use the following formula to update the position:

[0118] θ(i)=α*π*rand;

[0119] r(i) = θ(i);

[0120] xr(i)=r(i)*sin h(θ(i));

[0121] xr(i)=r(i)*cos h(θ(i));

[0122] x1(i)=xr(i) / max(|xr|);

[0123] y1(i)=yr(i) / max(|yr|);

[0124] The formula for updating the vulture's position during a dive is:

[0125]

[0126] Pi，new =rand*P best +δ x +δ y ;

[0127] Where: c1 and c2 represent the vulture's movement intensity towards the optimal and central positions, and both have values of (1, 2).

[0128] Step 6: Determine whether the end condition is met. If so, output the optimal result. Otherwise, repeat steps 2-step 6.

[0129] Specifically, by Figure 4 It can be seen that the BES algorithm optimizes the penalty factor and kernel function of SVM, and the optimized C and g are 25.0099 and 8.311 respectively.

[0130] Preferably, after inputting the sample data set into a support vector machine model for training and optimizing the support vector machine model using a vulture search algorithm to obtain the corrosion rate prediction model, the method further includes:

[0131] An error calculation is performed according to the prediction result of the corrosion rate prediction model to obtain a calculation result; the calculation result includes: a root mean square error, a mean absolute error and a correlation coefficient.

[0132] The model is evaluated based on the calculation results.

[0133] Furthermore, in this embodiment, the errors of the prediction results are calculated, and the errors are the root mean square error, the mean absolute error and the correlation coefficient, respectively, and the calculation method is as follows:

[0134] Root mean square error:

[0135]

[0136] Mean absolute error:

[0137]

[0138] Correlation coefficient:

[0139]

[0140] Among them, y i is the true value, is the predicted value, is the average value, and n is the number of test samples in the test set.

[0141] Furthermore, this embodiment evaluates the model prediction effect according to relevant evaluation indicators. The following is a data comparison of the corrosion rate prediction model trained by this solution, where the MAE and RMSE are set to be as small as possible, and R2 The setting requirement is that the closer to 1, the better. After comparison, the prediction data of the model of this embodiment and other models are shown in Table 2.

[0142] Table 2 Data comparison of corrosion rate prediction models

[0143] Model MAE RMSE <![CDATA[R 2 ]]> Support Vector Machine 0.073342 0.093278 0.894 PSO-SVM 0.021059 0.030568 0.90612 BES-SVM 0.014578 0.020199 0.93459 GRA-BES-SVM 0.0057949 0.0061693 0.97999

[0144] from Figure 5 As can be seen from Table 2, compared with the BES-SVM, PSO-SVM and SVM models, the mean absolute error and root mean square error of the GRA-BES-SVM prediction model are smaller than those of the other three prediction models, while the correlation coefficient is greater than that of the other three prediction models, indicating that the prediction model based on GRA-BES-SVM has better prediction performance in predicting the corrosion rate in gas pipelines.

[0145] The beneficial effects of the present invention are as follows:

[0146] In a specific implementation of the GRA-BES-SVM-based method for predicting the internal corrosion rate of a gas transmission pipeline, the present invention analyzes a test data set using a grey correlation analysis method to remove influencing factors with the lowest correlation. A vulture search algorithm is used to optimize the penalty factor c and kernel function g of the SVM neural network. The test set is then input into the trained prediction model to obtain a prediction result. If the error between the predicted value and the measured value is small, it indicates that the trained corrosion rate prediction model has achieved good prediction results. During actual prediction, the test data to be predicted can be input into the trained prediction model to obtain the internal corrosion rate of the gas transmission pipeline, with high prediction accuracy.

[0147] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0148] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A method for predicting the corrosion rate in a gas pipeline, characterized in that: include: Obtain the data to be tested; Inputting the data to be detected into a trained corrosion rate prediction model to obtain a corrosion rate prediction result; The training method of the corrosion rate prediction model includes: Acquire a detection data set consisting of pipeline detection sample data; the pipeline detection sample data includes relevant environmental influencing factors on internal corrosion behavior of the gas pipeline and corrosion rate detection results; The detection data set is screened based on a grey relational analysis method to obtain an environmental influencing factor data set, and a sample data set is constructed based on the environmental influencing factor data set and the corrosion rate detection results; Training a support vector machine model according to the sample data set, and optimizing the support vector machine model using a vulture search algorithm to obtain the corrosion rate prediction model; The detection data set is screened based on the grey relational analysis method to obtain a sample data set, including: Determine a parent sequence consisting of corrosion rate test results and a child sequence consisting of environmental influencing factors; Normalizing each index in the parent sequence and the child sequence to obtain a normalized index value of the parent sequence and a normalized index value of the child sequence; Calculating a correlation coefficient between each of the normalized index values in the subsequence and the normalized index value of the parent sequence; Calculating the correlation between each environmental influencing factor and the corrosion rate detection result according to each correlation coefficient; sorting the correlation degrees, and removing the environmental influencing factors with the smallest correlation degree according to the sorting results, and determining the remaining environmental influencing factors as the environmental influencing factor data set; The support vector machine model is trained according to the sample data set, and the support vector machine model is optimized using the vulture search algorithm to obtain the corrosion rate prediction model, including: Based on the nonlinear mapping relationship, determining an initial regression function according to the sample data set; Based on multivariate statistical analysis theory, determining an SVM regression function and constraints of the SVM regression function according to the initial regression function; Using the Lagrange multiplier method to perform constrained optimization on the duality variables of the SVM regression function to obtain a final regression function; Constructing the vector machine model using the final regression function; The penalty factor and kernel function of the vector machine model are optimized according to the vulture search algorithm to obtain the corrosion rate prediction model.

2. The method for predicting the corrosion rate in a gas pipeline according to claim 1, characterized in that: After inputting the sample data set into a support vector machine model for training and optimizing the support vector machine model using a vulture search algorithm to obtain the corrosion rate prediction model, the method further includes: Performing error calculation based on the prediction results of the corrosion rate prediction model to obtain calculation results; the calculation results include: root mean square error, mean absolute error and correlation coefficient; Model evaluation is performed based on the calculation results.

3. The method for predicting the corrosion rate in a gas pipeline according to claim 1, wherein: The environmental influencing factors include pipeline pressure, temperature, CO2 partial pressure, liquid holdup, fluid flow rate and pH.

4. The method for predicting the corrosion rate in a gas pipeline according to claim 1, wherein: After constructing a sample data set based on the environmental influencing factor data set and the corrosion rate detection results, the method further includes: The sample data set is divided into a training set and a test set according to a preset ratio; the training set is used to construct the support vector machine model; and the test set is used to test the corrosion rate prediction model.

Images