Pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation method and system

Abstract

The invention discloses a pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation method and system in the technical field of water supply pipeline corrosion evaluation. The method comprises the following steps: acquiring pipeline corrosion products in a plurality of pipelines; acquiring microscopic characterization data of the pipeline corrosion product, standardizing the microscopic characterization data to obtain microscopic standard data, and screening the microscopic standard data to obtain microscopic verification data; common factor data capable of representing the corrosion state of the pipeline and a corresponding factor load matrix are extracted from the microscopic verification data meeting the requirements; dividing the common factor into a forward factor and a reverse factor based on a preset factor direction; and calculating comprehensive evaluation data according to the factor load matrix and the forward factor and reverse factor direction rule, and judging the corrosion degree of the pipeline by combining the comprehensive evaluation data and a preset value, so that comprehensive detection and accurate quantitative evaluation of the corrosion state of the pipeline are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water supply pipeline corrosion evaluation, and particularly relates to a pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation method and system. BACKGROUND

[0002] Water supply pipeline inner wall corrosion is a common problem in water supply systems, often leading to pipeline leakage, water quality deterioration, and reduced water supply efficiency. Existing pipeline corrosion evaluation methods, such as a water supply network stability index system and its evaluation method (authorized publication number CN108764594B), provide an index system combining water quality and pipeline surface characteristics, and segment threshold judgment of corresponding indexes to determine pipeline health, but the evaluation system is single and the classification principle is rough, which cannot accurately define the pipeline condition; a long-distance pipeline internal corrosion detection method and system (publication number CN119400271A) constructs a multi-level corrosion evaluation system, and determines the corrosion degree grade by combining factor weight and membership matrix, but the weight distribution has subjective bias and limited applicability.

[0003] However, the current determination method for water supply pipeline corrosion state is not clear, and the existing detection and evaluation methods often cannot comprehensively and accurately reflect the corrosion state of the pipeline. Traditional detection methods usually rely on single physical or chemical indicators, such as pipeline wall thickness measurement, potential current monitoring, water quality parameter analysis, etc., which are difficult to fully reveal the complexity and multi-dimensional characteristics of pipeline corrosion. SUMMARY

[0004] The purpose of the present application is to provide a pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation method and system to solve the problem that the determination method for water supply pipeline corrosion state is not clear, and the existing detection and evaluation methods often cannot comprehensively and accurately reflect the corrosion state of the pipeline.

[0005] To achieve the above purpose, the present application provides the following technical solutions:

[0006] In a first aspect, the present application provides a pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation method, characterized in that it comprises:

[0007] Obtaining pipeline corrosion products in a plurality of pipelines;

[0008] Obtaining micro-characterization data of the pipeline corrosion products, standardizing the micro-characterization data to obtain micro-standard data, and screening the micro-standard data based on factor analysis suitability test to obtain micro-verification data;

[0009] The microcosmic verification data is subjected to dimension reduction analysis and a common factor of several targets is extracted to construct a factor loading matrix, the factor loading matrix comprising loadings on the common factor of several targets, and the common factor is divided into a positive factor and a reverse factor based on a preset factor direction;

[0010] The factor loading matrix is coupled with the positive factor and the reverse factor to calculate comprehensive evaluation data of the pipeline corrosion state, and if the comprehensive evaluation data is less than a preset value, it is output that the pipeline has a corrosion risk, otherwise, it is output that the pipeline is in a healthy state.

[0011] As a further scheme of the present application: the microcosmic characterization data of the pipeline corrosion product comprises:

[0012] The microcosmic characterization data comprises element index data, pore volume data, average pore size data, specific surface area data and crystal structure data of the pipeline corrosion product;

[0013] The element index data comprises relative content data of iron, calcium, magnesium, aluminum, silicon, phosphorus, sulfur, manganese and chlorine in the pipeline corrosion product;

[0014] The element index data is obtained by at least one of X-ray fluorescence spectroscopy, X-ray photoelectron spectroscopy, inductively coupled plasma spectroscopy and scanning electron microscope-energy spectrometer;

[0015] The pore volume data, the average pore size data, the specific surface area data and the crystal structure data are obtained by at least one of full-automatic specific surface and porosity analysis, X-ray diffraction and scanning electron microscope.

[0016] As a further scheme of the present application: the element index data, the pore volume data, the average pore size data, the specific surface area data and the crystal structure data are subjected to standardization processing to obtain microcosmic standard data;

[0017] The standardization processing comprises data transformation and data standardization;

[0018] The data transformation comprises center logarithmic ratio transformation processing of the relative content data to eliminate pseudo-correlation among iron, calcium, magnesium, aluminum, silicon, phosphorus, sulfur, manganese and chlorine in the pipeline corrosion product caused by constant total amount;

[0019] The data standardization comprises Z-score standardization of the microcosmic standard data, and the Z-score standardization formula is:

[0020]

[0021] Wherein x is: microcosmic characterization original data, x' is: standardized data, mean(x) is: sample mean, and sigma is: sample standard deviation.

[0022] As a further scheme of the present application: the suitability test based on factor analysis screens the micro-standard data to obtain micro-verification data, comprising;

[0023] The data correlation matrix is calculated from the micro-standard data, and the suitability test based on factor analysis screens the micro-standard data to obtain micro-verification data;

[0024] The standardized data is constructed into a data matrix, and the data matrix is converted to obtain a standardized data matrix X, and the data correlation matrix is obtained according to the standardized data matrix;

[0025] The data correlation matrix is:

[0026]

[0027] Where X T is the transpose of the standardized data matrix X, and n is the sample size;

[0028] The suitability test based on factor analysis includes KMO test and Bartlett's sphericity test, and the micro-verification data is screened according to the preset values of KMO test and Bartlett's sphericity test.

[0029] As a further scheme of the present application: the micro-verification data is subjected to dimensionality reduction analysis and a number of target common factors are extracted to construct a factor loading matrix, the factor loading matrix includes the loadings of the common factors of a number of targets, comprising;

[0030] The dimensionality reduction analysis method includes principal component analysis, and the principal component analysis includes;

[0031] The data correlation matrix R is subjected to eigenvalue decomposition:

[0032] R = P ∧ P T ;

[0033] Where P is the eigenvector matrix, P T is the transpose of the eigenvector matrix, ∧ is the diagonal eigenvalue matrix, and the eigenvalues in ∧ are eigenroots;

[0034] According to the standard that the eigenroot is greater than 1, a number of target common factors are selected, the number of common factors is K, and a factor loading matrix L of the first number of target common factors is constructed:

[0035]

[0036] The explained variance ratio of the i-th common factor in the K common factors:

[0037]

[0038] wherein λ i is: the eigenvalue corresponding to the i-th common factor, k is the number of extracted common factors.

[0039] As a further scheme of the present application: the common factors of the several targets include a first common factor, a second common factor and a third common factor;

[0040] Among them, the iron element content is the dominant variable in the first common factor, and the load signs of calcium element and silicon element are opposite to that of iron element, which is used to reflect the reverse relationship between metal corrosion and surface protection layer, and is used to characterize the inner wall corrosion degree and protection layer state.

[0041] The pore size parameter is the dominant variable in the second common factor, which is used to reflect the pore characteristics of the corrosion product structure, and is used to indicate the compactness and integrity of the corrosion layer.

[0042] The manganese element content is the dominant variable in the third common factor, and is positively correlated with the phosphorus element, which reflects the enrichment degree of exogenous elements in the corrosion product, and is used to evaluate the composition migration characteristics in the corrosion process.

[0043] As a further scheme of the present application: the common factors are divided into positive factors and reverse factors based on the preset factor direction, including:

[0044] When the load value of the iron element content in the first common factor is less than -0.5, and the load value of calcium and silicon element content in the first common factor is greater than 0.5, the common factor corresponding to the iron element is a positive factor.

[0045] When the load value of the pore size data in the second common factor is less than -0.5, the common factor corresponding to the pore size data is a positive factor.

[0046] When the load value of the manganese element and phosphorus element content in the third common factor is less than -0.5, the common factor corresponding to the manganese element and phosphorus element is a positive factor.

[0047] If the load direction of the main contribution characteristics in the common factor is inconsistent with the preset factor direction, the corresponding common factor is divided into a reverse factor according to the sign of the load value in the common factor.

[0048] As a further scheme of the present application: the factor loading matrix is coupled with the positive factor and the reverse factor, and the comprehensive evaluation data of the pipeline corrosion state is calculated, including:

[0049] The common factor data is calculated by regression method to obtain a factor score matrix F:

[0050] F = R -1 L(L T R -1 L) -1 X

[0051] Wherein R is: data correlation matrix, L is: factor loading matrix, X is: standardized data matrix:

[0052] According to each common factor term F1, F2, F3 of the factor score matrix F and the corresponding common factor weight, the comprehensive evaluation score Y of the pipeline corrosion state is calculated:

[0053] Y = Y1 + Y2 + Y3

[0054] Wherein, the common factor weight is the proportion of the explained variance corresponding to the common factor, Y1, Y2 and Y3 respectively represent the weighted contribution value corresponding to the first common factor, the second common factor and the third common factor;

[0055] If the comprehensive evaluation score of the corrosion state is less than 0, the pipeline exists corrosion risk, otherwise, the pipeline is in healthy state.

[0056] As a further scheme of the application: further comprising: the comprehensive evaluation score of the corrosion state and the existing pipeline health evaluation parameter are subjected to Spearman analysis or Pearson analysis to obtain correlation coefficient data, if the correlation coefficient data is higher than 0.9, it indicates that the dimensionality reduction analysis is effective, wherein the existing pipeline health evaluation parameter is obtained by pipeline internal real image, rust area ratio, corrosion nodule thickness, surface finish and four dimensions of pipe wall integrity for pipeline health score.

[0057] Secondly, the application provides a pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation system, the system comprises:

[0058] The sampling module is used for acquiring pipeline corrosion products in a plurality of pipelines.

[0059] The screening module acquires micro-characteristic data of the pipeline corrosion products, standardizes the micro-characteristic data to obtain micro-standard data, and screens the micro-standard data based on suitability test of factor analysis to obtain micro-verification data.

[0060] The analysis module performs dimensionality reduction analysis on the micro-verification data, extracts common factors of a plurality of targets, and constructs a factor loading matrix, the factor loading matrix comprising loadings on the common factors of a plurality of targets, and the common factors are divided into positive factors and negative factors based on a preset factor direction.

[0061] An output module is coupled with the factor loading matrix and forward factors and reverse factors to calculate comprehensive evaluation data of the pipeline corrosion state, and if the comprehensive evaluation data is less than a preset value, the pipeline has a corrosion risk, otherwise, the pipeline is in a healthy state.

[0062] Compared with the prior art, the present application has the following advantages:

[0063] 1、In the present application, by obtaining and processing the microcosmic characterization data of the pipeline corrosion product, the microcosmic standard data is screened to obtain microcosmic verification data, and the data meeting the requirements is preferably selected, the microcosmic verification data obtained by processing is extracted to obtain common factor data and factor loading matrix which can represent the pipeline corrosion state, the comprehensive evaluation data is calculated according to the direction rule of the factor loading matrix and forward factors and reverse factors, and the comprehensive evaluation data is used to realize the comprehensive detection and accurate evaluation of the pipeline corrosion state, the pipeline corrosion state is quantitatively calculated through the corrosion product feature extraction and data dimensionality reduction analysis, the subjective deviation of the traditional evaluation method is avoided, the scientificity and accuracy of the corrosion monitoring are improved, and the use effect is good.

[0064] 2、In the present application, by using the data dimensionality reduction and factor analysis means, the key corrosion features are quickly extracted and the score is obtained, the efficient evaluation of the pipeline corrosion state is realized, the pipeline network maintenance strategy is optimized, the operation risk is reduced, and efficient analysis and optimized maintenance are facilitated.

[0065] 3、The present application has strong applicability and can be applied to different pipeline materials and environmental media, can be integrated with the existing pipeline network management system, improves the intelligent level of pipeline corrosion prevention and control, and realizes efficient operation and maintenance of the water supply system. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 It is a method flow structure schematic diagram of the present application;

[0067] Figure 2 It is a principal component analysis common factor extraction diagram provided by example one of the present application;

[0068] Figure 3 It is a pipeline corrosion score and subjective health score correlation analysis diagram based on a pipeline actually photographed image provided by example one of the present application.

[0069] Figure 4 It is a pipeline corrosion score and subjective health score correlation analysis diagram based on a pipeline actually photographed image provided by comparative example one of the present application. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0071] Embodiments:

[0072] Please refer to Figure 1 In the embodiments of the present application, a pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation method has the characteristics that it comprises the following steps.

[0073] S1: Obtain pipeline corrosion products in a plurality of pipelines.

[0074] S2: Obtain micro-characterization data of the pipeline corrosion products, perform standardization processing on the micro-characterization data to obtain micro-standard data, perform screening on the micro-standard data based on suitability inspection of factor analysis to obtain micro-verification data.

[0075] S3: Perform dimensionality reduction analysis on the micro-verification data, extract a plurality of target common factors, and construct a factor loading matrix including the loadings on the common factors of the plurality of targets, divide the common factors into positive factors and negative factors based on a preset factor direction.

[0076] S4: Couple the factor loading matrix with the positive factors and the negative factors, and calculate comprehensive evaluation data of the pipeline corrosion state, if the comprehensive evaluation data is less than a preset value, output that the pipeline has a corrosion risk, otherwise, output that the pipeline is in a healthy state.

[0077] Specifically, the micro-characterization data of the pipeline corrosion products is obtained and processed to obtain micro-standard data, the micro-standard data is screened to obtain micro-verification data, and the data meeting the requirements is preferably selected, the micro-verification data obtained by processing is extracted to obtain common factor data and a factor loading matrix representing the pipeline corrosion state, the comprehensive evaluation data is calculated according to the direction rules of the factor loading matrix, the positive factors and the negative factors, and the comprehensive evaluation data is used to realize comprehensive detection and accurate evaluation of the pipeline corrosion state, the corrosion product feature extraction and data dimensionality reduction analysis are used to quantitatively calculate the pipeline corrosion state, subjective bias of the traditional evaluation method is avoided, the scientificity and accuracy of the corrosion monitoring are improved, and the use effect is good.

[0078] Further, by using data dimensionality reduction and factor analysis means, key corrosion features are quickly extracted and scores are obtained, efficient evaluation of the pipeline corrosion state is realized, pipeline network maintenance strategies are optimized, operation risks are reduced, and efficient analysis and optimized maintenance are facilitated.

[0079] Further, the application has strong applicability and can be applied to different pipeline materials and environmental media, can be integrated with existing pipeline management systems, can improve the intelligent level of pipeline corrosion prevention and control, and can realize efficient operation and maintenance of the water supply system

[0080] Preferably, the micro-characterization data of the pipeline corrosion product is obtained, including;

[0081] The micro-characterization data includes element index data, pore volume data, average pore size data, specific surface area data (BET), and crystal structure data of the pipeline corrosion product.

[0082] The element index data includes the relative content data of iron, calcium, magnesium, aluminum, silicon, phosphorus, sulfur, manganese, and chlorine in the pipeline corrosion product.

[0083] The element index data is obtained by at least one of X-ray fluorescence spectroscopy, X-ray photoelectron spectroscopy, inductively coupled plasma spectroscopy, and scanning electron microscope-energy spectrometer.

[0084] The pore volume data, average pore size data, specific surface area data, and crystal structure data are obtained by at least one of full-automatic specific surface and porosity analysis, X-ray diffraction, and scanning electron microscope.

[0085] Specifically, the element index data, pore volume data, average pore size data, specific surface area data, and crystal structure data of the pipeline corrosion product together constitute a comprehensive and detailed micro-description of the pipeline corrosion product. The standardization processing step ensures that data obtained by different sources and different measurement methods can be compared and analyzed on the same scale, improving the comparability of the data and the accuracy of the analysis results. Through the suitability test of factor analysis, the micro-characterization data most related to the pipeline corrosion state and with the highest information content are further selected. By using data dimension reduction and factor analysis means, key corrosion features are quickly extracted and scores are obtained, realizing efficient evaluation of the pipeline corrosion state, optimizing pipeline maintenance strategy, and reducing operation risk.

[0086] Specifically, X-ray fluorescence spectroscopy can quickly and accurately analyze the element composition in the pipeline corrosion product, while X-ray photoelectron spectroscopy can provide more in-depth element chemical state information. X-ray diffraction is mainly used to analyze the crystal phase structure of the corrosion product, and scanning electron microscope-energy spectrometer can directly observe the micro-morphology of the corrosion product and perform element quantitative analysis; full-automatic specific surface and porosity analysis is a common method for measuring pore volume, average pore size, and specific surface area, which uses gas adsorption principle to accurately measure these key parameters. X-ray diffraction and scanning electron microscope also play an important role in obtaining crystal structure data, which can reveal the crystal structure and micro-morphology characteristics of the corrosion product; the micro-characterization data of the pipeline corrosion product can be comprehensively and accurately obtained.

[0087] Preferably, the elemental index data, pore volume data, average pore size data, specific surface area data and crystal structure data are standardized to obtain micro-standard data;

[0088] The standardization process includes data transformation and data standardization;

[0089] The data transformation includes centering log-ratio transformation of the relative content data to eliminate the pseudo-correlation between iron, calcium, magnesium, aluminum, silicon, phosphorus, sulfur, manganese and chlorine in the pipeline corrosion products caused by constant total amount;

[0090] The data standardization includes Z-score standardization of the micro-standard data, and the Z-score standardization formula is:

[0091]

[0092] Wherein x is: micro-characterization raw data, x' is: standardized data, mean(x) is: sample mean, and σ is: sample standard deviation.

[0093] Specifically, the data correlation matrix is used to analyze the correlation between each micro-characterization data and reveal the internal relationship between them. By standardizing these data, the dimensional differences between different data can be eliminated, making the data more comparable.

[0094] Z-score standardization is a commonly used data standardization method, which converts the original data into standard normal distribution data with a mean of 0 and a standard deviation of 1, so that the data is more concentrated and easy to analyze. Data transformation includes centering log-ratio transformation of the relative content data to eliminate the pseudo-correlation between iron, calcium, magnesium, aluminum, silicon, phosphorus, sulfur, manganese and chlorine in the pipeline corrosion products caused by constant total amount. Data transformation also involves appropriate conversion of specific types of data to meet the requirements of subsequent analysis models.

[0095] Further, the standardization process also includes at least one of the minimum-maximum standardization, centering and scaling methods.

[0096] Preferably, the micro-verification data is obtained by screening the micro-standard data based on the suitability test of factor analysis, including:

[0097] The data correlation matrix is obtained from the micro-standard data, and the micro-verification data is obtained by screening the micro-standard data based on the suitability test of factor analysis;

[0098] The data matrix is constructed from the standardized micro-standard data, the standardized data matrix X is obtained by converting the data matrix, and the data correlation matrix is obtained from the standardized data matrix;

[0099] The data correlation matrix is:

[0100]

[0101] where X T is the transpose of the standardized data matrix X, and n is the number of samples;

[0102] The suitability test of factor analysis includes KMO test and Bartlett's test of sphericity, and the micro-standard data is filtered to obtain micro-verification data according to the preset values of the KMO test and the Bartlett's test of sphericity.

[0103] Specifically, the suitability test of factor analysis includes KMO test and Bartlett's test of sphericity.

[0104] KMO test:

[0105]

[0106] where is the sum of squares of Pearson correlation coefficients between each pair of index variables, is the sum of squares of partial correlation coefficients.

[0107] Bartlett's test of sphericity:

[0108]

[0109] where χ 2 is the chi-square statistic, n is the number of samples, and p is the number of index variables.

[0110] KMO test: KMO>0.5 and Bartlett's test of sphericity: P<0.05 meet the requirements of factor analysis, and micro-verification data is obtained.

[0111] The dimension reduction analysis method includes principal component analysis (PCA), maximum likelihood method, and principal axis factor method; the principal component analysis method extracts common factors, including:

[0112] Eigenvalue decomposition is performed on the data correlation matrix R:

[0113] R=P∧P T ;

[0114] where P is the eigenvector matrix, P T is the transpose of the eigenvector matrix, and ∧ is the diagonal eigenvalue matrix, and the eigenvalues in ∧ are eigenroots.

[0115] The first k common factors are selected according to the standard that the eigenroot is greater than 1, and the factor loading matrix L representing the contribution size of each original variable on the common factors is calculated:

[0116]

[0117] The factor loading matrix can also be rotated by an orthogonal rotation method to enhance the attribution clarity between the dominant variables and the common factors.

[0118] The explained variance proportion corresponding to the i-th common factor:

[0119]

[0120] wherein λ i is the eigenvalue corresponding to the i-th common factor, and k is the number of extracted common factors.

[0121] Preferably, the micro-verification data is subjected to dimensionality reduction analysis and a plurality of common factors of the target are extracted and a factor loading matrix is constructed, the factor loading matrix including the loadings on the common factors of the plurality of targets, including:

[0122] The dimensionality reduction analysis method includes principal component analysis, and the principal component analysis includes:

[0123] The data correlation matrix R is subjected to eigenvalue decomposition:

[0124] R = P ∧ P T ;

[0125] wherein P is a characteristic vector matrix, P T is the transpose of the characteristic vector matrix, and ∧ is a diagonal eigenvalue matrix, and the eigenvalues in ∧ are eigenroots;

[0126] The first plurality of common factors of the target are selected according to the standard that the eigenroot is greater than 1, the number of common factors is K, and a factor loading matrix L of the first plurality of common factors of the target is constructed:

[0127]

[0128] The explained variance proportion corresponding to the i-th common factor in the K common factors:

[0129]

[0130] wherein λ i is the eigenvalue corresponding to the i-th common factor, and k is the number of extracted common factors.

[0131] Preferably, the plurality of common factors of the target include a first common factor, a second common factor, and a third common factor;

[0132] wherein the first common factor takes the iron element content as the dominant variable, and the calcium element and the silicon element loadings are in opposite directions to the iron element, so as to reflect the inverse relationship between metal corrosion and the surface protective layer, and to characterize the inner wall rusting degree and the protective layer state.

[0133] The second common factor takes the pore size parameter as a dominant variable to reflect the pore characteristics of the corrosion product structure, and is used to indicate the compactness and integrity of the corrosion layer;

[0134] The third common factor takes the manganese element content as a dominant variable and is positively correlated with the phosphorus element, reflecting the enrichment degree of exogenous elements in the corrosion product, and is used to evaluate the composition migration characteristics in the corrosion process.

[0135] Preferably, the common factors are divided into positive factors and negative factors based on the preset factor direction, including:

[0136] When the load value of the iron element content in the first common factor is less than -0.5 and the load value of the calcium and silicon element content in the first common factor is greater than 0.5, the common factor corresponding to the iron element is a positive factor;

[0137] When the load value of the pore size data in the second common factor is less than -0.5, the common factor corresponding to the pore size data is a positive factor;

[0138] When the load value of the manganese element and phosphorus element content in the third common factor is less than -0.5, the common factor corresponding to the manganese element and phosphorus element is a positive factor;

[0139] If the load direction of the main contribution characteristics in the common factor is inconsistent with the preset factor direction, the corresponding common factor is divided into a negative factor according to the sign corresponding to the load value in the common factor.

[0140] Specifically, if the load direction of the main contribution characteristics in a certain common factor is opposite to the preset factor direction, that is, the load value is negative and the absolute value is greater than a preset threshold, then the common factor is considered as a negative factor. The negative factor indicates that the change trend of the characteristics represented by the common factor in the corrosion product is opposite to the expected or conventional understanding, revealing the influence of special corrosion mechanism or environmental factors on the corrosion process. By comprehensively understanding the multi-dimensional characteristics of the pipeline corrosion product through the information of the positive factor and the negative factor, the corrosion state can be accurately evaluated.

[0141] Preferably, the factor loading matrix is coupled with the positive factor and the negative factor to calculate the comprehensive evaluation data of the pipeline corrosion state, including:

[0142] The common factor data is calculated by a regression method to obtain a factor score matrix F:

[0143] F = R -1 L(L T R -1 L) -1 X

[0144] Wherein R is a data correlation matrix, L is a factor loading matrix, and X is a standardized data matrix:

[0145] According to the coupling of each common factor term F1, F2, F3 of the factor score matrix F and the corresponding common factor weight, the comprehensive evaluation score Y of the pipeline corrosion state is calculated:

[0146] Y=Y1+Y2+Y3

[0147] Wherein, the common factor weight is the proportion of the explained variance corresponding to the common factor, Y1, Y2 and Y3 respectively represent the weighted contribution value corresponding to the first common factor, the second common factor and the third common factor;

[0148] If the comprehensive evaluation score of the corrosion state is less than 0, it is output that the pipeline has corrosion risk, otherwise, it is output that the pipeline is in a healthy state;

[0149] For newly introduced sample data, the existing standardization method, factor loading matrix and direction rule can be directly substituted without reconstructing the factor structure to calculate the comprehensive evaluation score of the corrosion state.

[0150] Preferably, it also includes: performing spearman analysis or pearson analysis on the comprehensive evaluation score of the corrosion state and the existing pipeline health evaluation parameter to obtain correlation coefficient data, if the correlation coefficient data is higher than 0.9, it indicates that the dimensionality reduction analysis is effective, wherein the existing pipeline health evaluation parameter is obtained by pipeline internal shooting image, rust area ratio, corrosion nodule thickness, surface finish and pipe wall integrity four dimensions to score the pipeline health, that is, the subjective health score of the pipeline.

[0151] Obtain pipeline corrosion products in a plurality of pipelines, comprising:

[0152] The pipeline corrosion products are obtained by scraper, forceps or sampling hammer on the inner wall of the pipeline, wherein the sampling time of the pipeline corrosion products is limited to within 2 hours after the inner wall of the pipeline is separated from the water, and the pipeline corrosion products are stored by low-temperature transportation after being collected and dried at 60 DEG C.

[0153] Specifically, when sampling, it is necessary to ensure that the tools such as scraper, forceps or sampling hammer are clean and free of pollution to avoid secondary pollution of the corrosion products. After sampling, the corrosion products are marked and recorded for subsequent analysis and research. Low-temperature transportation can slow down the chemical change of the corrosion products and maintain their original state, and drying at 60 DEG C. can remove water to facilitate subsequent chemical analysis and testing.

[0154] The present application provides a pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation system, as shown in FIG. 1, the system comprises:

[0155] A sampling module is used to obtain pipeline corrosion products in a plurality of pipelines;

[0156] The screening module obtains micro-characterization data of the pipeline corrosion products, standardizes the micro-characterization data to obtain micro-standard data, and screens the micro-standard data based on suitability inspection of factor analysis to obtain micro-verification data.

[0157] The analysis module performs dimensionality reduction analysis on the micro-verification data, extracts common factors of a plurality of targets, and constructs a factor loading matrix including loadings on the common factors of the plurality of targets. The common factors are divided into positive factors and negative factors based on a preset factor direction.

[0158] The output module couples the factor loading matrix with the positive factors and the negative factors, and calculates comprehensive evaluation data of the pipeline corrosion state. If the comprehensive evaluation data is less than a preset value, it is output that the pipeline has a corrosion risk, otherwise, it is output that the pipeline is in a healthy state.

[0159] Example 1

[0160] In this embodiment, a water supply network in a certain place is taken as an implementation background, and a commonly used principal component analysis method is used for dimensionality reduction operation based on the method established in the embodiment. The conditions for verifying the establishment of factor analysis include KMO statistics>0.5 and Bartlett test p value<0.05.

[0161] First, the original data of the micro-characterization of the corrosion products on the inner wall of the 18 groups of pipelines is standardized.

[0162]

[0163] The correlation matrix or the covariance matrix of the variables is calculated.

[0164]

[0165] The suitability of factor analysis is verified. The correlation between variables is evaluated by KMO test (Kaiser-Meyer-Olkin test) to determine whether it is suitable for factor analysis. KMO>0.5 indicates that it is suitable for factor analysis.

[0166]

[0167] The Bartlett spherical test is used to determine whether the correlation matrix is an identity matrix (no correlation). If the p value is less than 0.05, it indicates that it is suitable for factor analysis.

[0168]

[0169] After screening the index data that can pass the test, the factors are extracted by the eigenvalue decomposition method, and the correlation matrix R is decomposed by eigenvalue:

[0170] R=P∧P T

[0171] After the operation, the KMO statistics of the group of data = 0.563, and the p value of Bartlett's sphericity test < 0.001, which meets the requirements of factor analysis. After principal component analysis, the eigenvalues of the first three principal components are greater than 1 (as shown in the scatter plot, the component number is the component in the micro characterization data, and the part number 1 represents the first principal component (iron), the component number 2 represents the second principal component (average pore size) …), and the cumulative variance contribution rate of the first three principal components = 83.6%>80%, indicating that the first three principal components can be extracted to describe the level of pipeline corrosion. Figure 2

[0172] The principal component coefficient matrix is obtained after eigenvalue decomposition based on the extracted common factors, and the results are shown in Table 1;

[0173] Corrosion characteristic index PC1 PC2 PC3 Fe -0.843 0.208 -0.472 Mg 0.738 -0.61 -0.036 P 0.241 -0.433 0.829 Si 0.829 -0.054 0.385 Specific surface area 0.236 -0.827 0.193 S -0.11 0.743 -0.222 Average pore size 0.328 0.804 0.113 Ca 0.864 0.164 -0.156 Mn 0.023 0.036 0.952

[0174] Table 1

[0175] Among them, the Fe load in the principal component PC1 is the largest, indicating that this component is a positive factor, and the low score on the principal component 1 reflects more corrosion products; the average pore size in the principal component PC2 has the largest positive load, indicating that this component is a reverse factor, and the high score on the principal component 2 reflects higher looseness, and the sediment is easy to fall off; the Mn load in the principal component PC3 is the largest, indicating that this component is a reverse factor, and the high score on the principal component 3 reflects that the deposition of manganese in the water quality environment is stronger. Overall, the principal components extracted from the group of data can analyze the corrosion state characteristics of the pipeline from the corrosion and mineralization characteristics, the structural looseness of the sediment, and the environmental chemical characteristics.

[0176] Based on the extracted common factors, the specific value of each sample on the extracted factor is calculated by regression method to obtain the factor score matrix F:

[0177] F = R -1 L(L T R -1 L) -1 X

[0178] The element content distribution and microstructure characterization of 18 groups of pipeline corrosion product samples are calculated according to the variance contribution rate of each common factor as the weight to obtain the comprehensive evaluation score of the corrosion state of each pipeline, and the subjective pipeline health score based on the internal shooting image is verified by Spearman correlation analysis, as shown in Figure 3 The correlation coefficient is 0.91, indicating that the score obtained by the operation can well represent the pipeline corrosion state, and the subjective pipeline health score is the corrosion grade score according to the pipeline photos.

[0179] Comparative example 1

[0180] ​The method of directly weighting and adding the corrosion characterization indicators is selected as a comparative example, other application conditions are kept the same as the foregoing examples, the pipeline corrosion product characterization data in Example 1 is weighted according to the following principles: among the main corrosion related elements (Fe, Mn, S), iron is the main corrosion product, manganese can promote localized corrosion, and sulfur can form H2S under anaerobic conditions to accelerate corrosion, so the weight is higher, and all are 0.15; the deposition related elements (Ca, Mg, Si, P) are usually related to mineral fouling, which can affect the formation and stability of the corrosion product, and are given a medium weight of 0.06-0.10; the specific surface area (BET) and average pore size affect the structural properties of the corrosion product, such as porosity, compactness, etc., and are given a lower weight of 0.05. The score calculated by directly weighting the normalized data according to the above weight is subjected to Spearman correlation analysis with the subjective pipeline health score based on the actual internal image of the pipeline, as shown in Figure 4 The existing pipeline health evaluation parameters are obtained by expert rating method weighting calculation based on the actual internal image of the pipeline

[0181] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can make equivalent replacement or change within the technical range disclosed by the present application according to the technical solution and inventive concept of the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for multi-dimensional feature extraction and corrosion state evaluation of pipeline corrosion products, characterized in that, The method comprises the following steps: obtaining corrosion products of a pipeline in a plurality of pipelines; obtaining micro-characterization data of the pipeline corrosion products, standardizing the micro-characterization data to obtain micro-standard data, and screening the micro-standard data based on a suitability test of factor analysis to obtain micro-verification data; performing dimensionality reduction analysis on the micro-verification data, extracting a plurality of target common factors, and constructing a factor loading matrix, wherein the factor loading matrix comprises loadings on the common factors of the plurality of targets, and the common factors are divided into positive factors and negative factors based on a preset factor direction; coupling the factor loading matrix with the positive factors and the negative factors, and calculating comprehensive evaluation data of the pipeline corrosion state, wherein if the comprehensive evaluation data is less than a preset value, it is output that the pipeline has a corrosion risk, otherwise, it is output that the pipeline is in a healthy state.

2. The method according to claim 1, wherein the method is characterized by: The method comprises the following steps: The micro-characterization data comprises element index data, pore volume data, average pore size data, specific surface area data, and crystal structure data of the pipeline corrosion products. The element index data comprises relative content data of iron, calcium, magnesium, aluminum, silicon, phosphorus, sulfur, manganese, and chlorine in the pipeline corrosion products. The element index data is obtained by at least one of X-ray fluorescence spectroscopy, X-ray photoelectron spectroscopy, inductively coupled plasma spectroscopy, and scanning electron microscopy-energy dispersive spectrometer. The pore volume data, average pore size data, specific surface area data, and crystal structure data are obtained by at least one of full-automatic specific surface and porosity analysis, X-ray diffraction, and scanning electron microscopy.

3. The method according to claim 2, wherein the method is characterized by: The element index data, pore volume data, average pore size data, specific surface area data, and crystal structure data are standardized to obtain micro-standard data. The standardization process comprises data transformation and data standardization. The data transformation comprises center log ratio transformation of the relative content data to eliminate pseudo-correlation between iron, calcium, magnesium, aluminum, silicon, phosphorus, sulfur, manganese, and chlorine in the pipeline corrosion products caused by constant total amount. The data standardization comprises Z-score standardization of the micro-standard data, and the Z-score standardization formula is: wherein x is the micro-characterization original data, x' is the standardized data, mean(x) is the sample mean, and σ is the sample standard deviation.

4. The method according to claim 3, characterized in that: The method comprises the following steps: The micro-standard data is screened based on a suitability test of factor analysis to obtain micro-verification data. The micro-standard data is standardized to obtain a data matrix, the data matrix is converted to obtain a standardized data matrix X, and a data correlation matrix is obtained based on the standardized data matrix. The data correlation matrix is: where X T is the transpose of the standardized data matrix X and n is the number of samples; The suitability test of the factor analysis includes KMO test and Bartlett's test, and the microcosmic standard data is filtered to obtain microcosmic verification data according to preset values of the KMO test and the Bartlett's test.

5. The method according to claim 4, wherein the method is characterized by: The microcosmic verification data is subjected to dimension reduction analysis, and a plurality of target common factors are extracted to construct a factor loading matrix, the factor loading matrix including the loadings on the common factors of the plurality of targets, including: The dimension reduction analysis method includes principal component analysis, and the principal component analysis includes: The data correlation matrix R is subjected to eigenvalue decomposition: R = P ^ P T ; where P is: the eigenvector matrix, P T is: the transpose of the eigenvector matrix, and A is: the diagonal eigenvalue matrix, the eigenvalues in A are the eigenroots; A plurality of target common factors are selected according to the standard that the eigenvalues are greater than 1, the number of the common factors is K, and a factor loading matrix L of the plurality of target common factors is constructed: The proportion of the explained variance corresponding to the i-th common factor in the K common factors: where λ i is: the eigenvalue corresponding to the ith common factor, k is the number of extracted common factors.

6. The method according to claim 5, wherein the method is characterized by: The common factors of the plurality of targets include a first common factor, a second common factor and a third common factor; In the first common factor, the content of iron element is the dominant variable, and the loadings of calcium element and silicon element are in the opposite direction of the iron element, reflecting the inverse relationship between metal corrosion and surface protective layer, and being used to represent the corrosion degree of inner wall and the state of protective layer; In the second common factor, the pore size parameter is the dominant variable, reflecting the pore characteristics of the structure of corrosion products, and being used to indicate the compactness and integrity of the corrosion layer; In the third common factor, the content of manganese element is the dominant variable, and is positively correlated with the content of phosphorus element, reflecting the enrichment degree of exogenous elements in the corrosion products, and being used to evaluate the composition migration characteristics in the corrosion process.

7. The method according to claim 6, wherein the method is characterized by: The common factors are divided into positive factors and negative factors based on preset factor directions, including: When the loading value of the content of iron element in the first common factor is less than -0.5, and the loading values of the contents of calcium and silicon element in the first common factor are greater than 0.5, the common factor corresponding to the iron element is a positive factor; When the loading value of the pore size data in the second common factor is less than -0.5, the common factor corresponding to the pore size data is a positive factor; When the loading values of the contents of manganese element and phosphorus element in the third common factor are less than -0.5, the common factor corresponding to the contents of manganese element and phosphorus element is a positive factor; If the loading direction of the main contribution characteristics in the common factor is inconsistent with the preset factor direction, the common factor is divided into a negative factor according to the sign corresponding to the loading value in the common factor.

8. The method according to claim 7, wherein the method is characterized by: The factor loading matrix is coupled with the positive factors and the negative factors to calculate the comprehensive evaluation data of the pipeline corrosion state, including: The factor score matrix F is calculated by using a regression method on the common factor data: F = R -1 L(L T R -1 L) -1 X Wherein R is a data correlation matrix, L is a factor loading matrix, and X is a standardized data matrix: The comprehensive evaluation score Y of the pipeline corrosion state is calculated according to the coupling of each common factor item F1, F2, F3 of the factor score matrix F and the corresponding common factor weight: Y=Y1+Y2+Y3 Wherein the common factor weight is the proportion of the explained variance corresponding to the common factor, Y1, Y2 and Y3 respectively represent the weighted contribution values corresponding to the first common factor, the second common factor and the third common factor. If the comprehensive evaluation score of the corrosion state is less than 0, it is output that the pipeline has a corrosion risk, otherwise, it is output that the pipeline is in a healthy state.

9. The method according to claim 8, characterized in that: Also included are: The comprehensive evaluation score of the corrosion state is subjected to Spearman analysis or Pearson analysis with existing pipeline health evaluation parameters to obtain correlation coefficient data, if the correlation coefficient data is higher than 0.9, it indicates that the dimensionality reduction analysis is effective, wherein the existing pipeline health evaluation parameters are obtained by pipeline internal real image, rust area ratio, corrosion nodule thickness, surface finish and four dimensions of pipeline wall integrity for pipeline health scoring.

10. A pipeline corrosion product multi-dimensional feature extraction and corrosion state evaluation system, characterized in that, The system comprises: A sampling module for obtaining pipeline corrosion products in a plurality of pipelines; A screening module for obtaining micro-characterization data of the pipeline corrosion products, standardizing the micro-characterization data to obtain micro-standard data, and screening the micro-standard data based on suitability test of factor analysis to obtain micro-validation data; An analysis module for performing dimensionality reduction analysis on the micro-validation data, extracting common factors of a plurality of targets and constructing a factor loading matrix, the factor loading matrix comprising loadings on the common factors of the plurality of targets, and dividing the common factors into positive factors and negative factors based on a preset factor direction; An output module for coupling the factor loading matrix with the positive factors and the negative factors and calculating comprehensive evaluation data of the pipeline corrosion state, if the comprehensive evaluation data is less than a preset value, it is output that the pipeline has a corrosion risk, otherwise, it is output that the pipeline is in a healthy state.

Images