Pipeline damage assessment method and corrosion prevention method based on multi-dimensional data fusion

Abstract

The invention relates to the technical field of chemical production, and discloses a pipeline damage assessment method and a corrosion prevention method based on multi-dimensional data fusion. The pipeline damage assessment method comprises the steps of configuring a trained lightweight neural network model, inputting a pipeline damage picture into the network model, and obtaining probability distribution of damage types after feature extraction, feature enhancement, feature mapping and output; the corrosion related index data set is used as input, a correlation coefficient method is applied to evaluate the correlation degree between the corrosion factors and the corrosion rate, and an initial strong correlation data set is obtained; expanding the initial strong correlation data set to obtain a strong correlation data set; the strong correlation data set is used as input, and an Apriori correlation algorithm is applied to evaluate a strong correlation rule between corrosion factors; and performing qualitative and quantitative evaluation on the damage of the chemical pipeline based on the probability distribution of the damage type and a strong association rule. According to the method, the mechanism and influence factors of pipeline corrosion can be deeply known, and a basis is provided for formulating a targeted corrosion protection strategy.

Description

Technical Field

[0001] The present invention belongs to the technical field of chemical production, and particularly relates to a pipeline damage assessment method and an anti-corrosion method based on multi-dimensional data fusion. Background Art

[0002] In the field of chemical production, equipment is often exposed to a series of complex working conditions, and these stresses include but are not limited to corrosion, wear, thermal stress, and pressure cycling, etc. The combined action of these factors may cause structural damage to metal materials, thereby affecting the reliability and safety of equipment. The damage of metal materials will not only shorten the expected service life of equipment, increase the economic burden of maintenance and replacement, but also, in extreme cases, may trigger serious safety accidents, posing a threat to the life safety of operators and environmental health. Therefore, it is particularly necessary to accurately identify and evaluate the damage of steel structures in chemical equipment, not only to extend the service life of equipment and reduce maintenance costs, but also to prevent potential safety risks and ensure personnel safety and environmental protection.

[0003] There are various influencing factors for pipeline corrosion, and there have been many studies on its correlation analysis in data mining. However, in the field of data mining, at present, domestic and foreign scholars' correlation analysis of corrosion influencing factors is mainly limited to the analysis of individual corrosion factors of specific long-distance pipelines, such as the corrosion analysis of pipelines transporting sour natural gas, while there are few studies on the correlation of corrosion influencing factors of process pipelines in industrial installations, and the correlation of process pipeline corrosion factors is not clear, making the anti-corrosion measures for process pipelines lack pertinence and the anti-corrosion efficiency is low; and at present, the data sources of some studies are not real pipeline corrosion data, but data obtained from simulation experiments or tests, which cannot fully reflect the laws contained in real pipeline corrosion data and there will be a slight deviation from the actual laws; in addition, each study only studies the correlation analysis of each corrosion influencing factor, and does not give how to optimize the existing pipeline anti-corrosion measures based on the current laws.

[0004] In summary, the existing technologies have deficiencies such as low accuracy of pipeline damage detection, insufficient or uneven distribution of research data samples, and failure to propose effective anti-corrosion measures. Summary of the Invention

[0005] The purpose of the present invention is to propose a pipeline damage assessment method and an anti-corrosion method based on multi-dimensional data fusion in view of the deficiencies of the existing technologies.

[0006] To achieve the above objectives, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides a pipeline damage assessment method based on multi-dimensional data fusion, including the following steps:

[0008] Configure the trained lightweight neural network model, input the pipeline damage pictures into the lightweight neural network model, and after feature extraction, feature enhancement, feature mapping and output, obtain the probability distribution of the damage types;

[0009] Taking the corrosion-related index data set as the input, apply the correlation coefficient method to evaluate the degree of association between the corrosion factors and the corrosion rate, and obtain the initial strongly correlated data set, and the initial strongly correlated data set is a small sample data set;

[0010] Expand the initial strongly correlated data set to obtain the strongly correlated data set;

[0011] Taking the strongly correlated data set as the input, apply the Apriori association algorithm to evaluate the strong association rules between the corrosion factors;

[0012] Qualitatively and quantitatively evaluate the chemical pipeline damage based on the probability distribution of the damage types and the strong association rules.

[0013] As a possible implementation, the backbone network of the lightweight neural network model is MobileNetV2, and an attention mechanism module is placed after MobileNetV2; the feature extraction specifically includes:

[0014] S10. The pipeline damage pictures expand their channels through dimensionality-increasing convolution, and the dimensionality-increasing convolution is depth convolution, that is, each channel is convolved separately;

[0015] S11. Apply an activation function to enhance the non-linear characteristics of MobileNetV2, and limit the output value between [0, 6];

[0016] S12. Reduce the number of channels from the expanded number of channels to the original number of channels through dimensionality-reducing convolution, and the dimensionality-reducing convolution is pointwise convolution, and apply pointwise convolution to integrate the information between channels;

[0017] After S10 to S12, the features of the pipeline damage pictures are extracted and output as the feature map Y0.

[0018] As a possible implementation, the attention mechanism module includes channel attention and spatial attention, and the feature enhancement specifically includes:

[0019] The channel attention generates the attention weights of the channels through average pooling and max pooling, which are used to highlight the important channel information;

[0020] Perform a fully connected network process on the pooled result to obtain the channel attention weights;

[0021] Multiply the calculated channel attention weights by the feature map Y0 to obtain the feature map Y1 with enhanced channel attention.

[0022] As a possible implementation, feature enhancement further includes:

[0023] Perform global average pooling and global max pooling on the feature map Y1 to obtain the pooling results at each position;

[0024] After concatenating the pooling at each position, generate spatial attention weights through convolution;

[0025] Multiply the spatial attention weights by the feature map Y1 to highlight the important spatial regions in the feature map Y1 and obtain the spatially attention-enhanced feature map Y2.

[0026] As a possible implementation, feature mapping and output specifically include:

[0027] Flatten the feature map Y2 into a vector, and input the flattened vector into a fully connected layer to obtain the probability distribution of the damage categories.

[0028] As a possible implementation, the initial strongly correlated dataset is obtained through the following method:

[0029] S20. Obtain the original dataset, and through preprocessing, obtain the corrosion-related index dataset. The preprocessing includes at least data cleaning, outlier removal, and missing value filling; there are multiple corrosion factors X and corresponding corrosion rates Y in the corrosion-related index dataset;

[0030] S21. For data with a normal distribution, calculate the Pearson product-moment correlation coefficient τ between the corrosion factor X and the corrosion rate Y xy ; for data with a non-normal distribution, calculate the Spearman rank correlation coefficient τ between the corrosion factor X and the corrosion rate Y s ; for each corrosion factor X in the corrosion-related index dataset, obtain the corresponding correlation coefficient

[0031] S22. Configure the correlation coefficient threshold, and retain the corrosion factor X whose correlation coefficient is greater than or equal to the correlation coefficient threshold, and discard the corrosion factor X whose correlation coefficient is less than the correlation coefficient threshold. The strongly correlated dataset is composed of all the retained corrosion factors X and their corresponding corrosion rates Y.

[0032] As a possible implementation, the strong association rules are determined through the following method:

[0033] The Apriori association algorithm gradually scans the strongly correlated dataset and generates frequent item sets by using the properties of frequent item sets;

[0034] For X and Y in the frequent item sets, if and X ∪ Y is a frequent item set, then an association rule can be formed Its confidence level is:

[0035]

[0036] As a possible implementation, the chemical pipeline damage assessment method further includes: formulating an anti-corrosion mechanism based on strong association rules, specifically including:

[0037] S30. Obtain the recommended value of the corrosion rate, and sequentially obtain the recommended values of corrosion influencing factors and the corrosion rate based on strong association rules;

[0038] S31. Obtain the monitored value of the corrosion rate, monitor the corrosion rate and corrosion factors to obtain the monitored value of the corrosion rate;

[0039] S32. Determine whether the monitored value exceeds the recommended value. If so, execute S30; otherwise, loop and execute S31 - S32;

[0040] S33. Alarm and decision support;

[0041] S34. Execute decision adjustment according to strong association rules;

[0042] S35. Determine whether the alarm is eliminated. If so, loop and execute S32 - S35; otherwise, perform process anti-corrosion inspection.

[0043] As a possible implementation, apply a generative adversarial network to expand the initial strongly correlated dataset to obtain a strongly correlated dataset.

[0044] In a second aspect, the present invention provides a pipeline anti-corrosion method based on multi-dimensional data fusion, including the following steps:

[0045] Apply the configured and trained lightweight neural network model to evaluate the damaged parts and damage types of chemical pipelines, and establish an inspection ledger for dangerous points; the damage types include mechanical damage and chemical corrosion damage;

[0046] Apply the mined strong association rules to improve the existing chemical pipeline anti-corrosion strategy.

[0047] Compared with the prior art, the beneficial effects produced by the present invention are as follows:

[0048] 1. The pipeline damage assessment method based on multi-dimensional data fusion provided by the present invention proposes to use a lightweight neural network model to process pipeline damage pictures. This model uses MobileNetV2 as the backbone network, and an attention mechanism module is placed after MobileNetV2. The backbone network MobileNetV2 itself takes depthwise separable convolution as the core, featuring lightweight and efficient computing. After introducing the attention mechanism module, the features are dynamically weighted and optimized through the channel attention mechanism and the spatial attention mechanism, significantly enhancing the model's ability to focus on key features.

[0049] 2. For the pipeline damage assessment method based on multi-dimensional data fusion provided by the present invention, its initial strongly correlated dataset is a small sample dataset, and then the generative adversarial network and the Apriori association rule algorithm are combined to evaluate the strong association rules among corrosion factors. Such strong association rules based on small samples can solve the problems of insufficient corrosion data samples and low rule mining efficiency, thus more accurately mining the association rules among corrosion influencing factors and providing theoretical support for corrosion prediction and protection.

[0050] 3. The present invention combines the correlation coefficient method, the GAN adversarial neural network, and the Apriori association rule algorithm to implement a data mining algorithm for studying the correlation between various corrosion factors and the corrosion rate. The correlation coefficient method is used to evaluate the degree of association between each factor in the original dataset and the corrosion rate, obtaining a small sample dataset; the highly correlated corrosion influencing factors are selected, and the small sample dataset is expanded through the GAN adversarial neural network. Finally, the Apriori association rule mining algorithm is used to discover potential association rules, which can ensure the accuracy of the evaluation of the correlation between various corrosion factors and the corrosion rate, help to deeply understand the mechanism and influencing factors of pipeline corrosion, and provide a scientific basis for formulating targeted corrosion protection strategies.

[0051] 4. The present invention also proposes a pipeline anti-corrosion method based on multi-dimensional data fusion. By conducting pipeline corrosion damage inspection and assessment based on multi-dimensional data, a pipeline anti-corrosion method is formed to optimize existing maintenance measures, and ultimately keep the safety of process pipelines at a high level at a relatively low cost. These improvement measures will help improve the safety and reliability of pipelines, extend their service life, and reduce maintenance and repair costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0053] Figure 1 It is a flowchart of the pipeline damage assessment method based on multi-dimensional data fusion provided by an embodiment of the present invention;

[0054] Figure 2 It is the curve graph of the training iteration process of the CBAM-MobileNet V2 model provided by the embodiment of the present invention;

[0055] Figure 3 It is the flowchart of data mining for corrosion factors of the GAN-Apriori association rule pipeline provided by the embodiment of the present invention;

[0056] Figure 4 It is the flowchart of pipeline damage assessment and anti-corrosion method based on multi-dimensional data fusion provided by the embodiment of the present invention. Detailed implementation manners

[0057] For the convenience of clearly describing the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, terms such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and roles. For example, the first threshold and the second threshold are only used to distinguish different thresholds, and do not limit their sequence. Those skilled in the art can understand that terms such as "first" and "second" do not limit the quantity and execution order, and terms such as "first" and "second" do not necessarily limit being different.

[0058] It should be noted that in the present invention, words such as "exemplary" or "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, using words such as "exemplary" or "for example" aims to present relevant concepts in a specific manner.

[0059] In the present invention, "at least one" means one or more, and "a plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the front and rear associated objects. The following at least one (item) or its similar expression refers to any combination of these items, including any combination of single item (item) or plural items (items). For example, at least one (item) of a, b or c can represent: a, b, c, the combination of a and b, the combination of a and c, the combination of b and c, or the combination of a, b and c, where a, b and c can be single or multiple.

[0060] An embodiment of the present invention aims to provide a pipeline damage assessment method and an anti-corrosion method based on multi-dimensional data fusion, and proposes a trained lightweight neural network model. An attention mechanism module is placed after the backbone network MobileNetV2, enabling the model to accurately extract the features of pipeline damage pictures with almost no increase in computational resource consumption and running time.

[0061] In a first aspect, an embodiment of the present invention provides a chemical pipeline damage assessment method based on multi-dimensional data fusion. Refer to Figure 1 , including the following steps:

[0062] Configure a trained lightweight neural network model, input the pipeline damage picture into the lightweight neural network model, and after feature extraction, feature enhancement, feature mapping and output, obtain the probability distribution of the damage type;

[0063] As a possible implementation, the backbone network of the lightweight neural network model is MobileNetV2, and an attention mechanism module is placed after MobileNetV2; the backbone network MobileNetV2 itself takes depthwise separable convolution as the core, with the characteristics of lightweight and efficient computing, but its ability to model the dependence on global features is limited. After introducing the attention mechanism module (CBAM), the features are dynamically weighted and optimized through the channel attention mechanism and the spatial attention mechanism, significantly enhancing the model's ability to focus on key features.

[0064] As an example, the structure of the lightweight neural network model (CBAM-MobileNet) is shown in Table 1:

[0065] Table 1 Structure of the lightweight neural network model

[0066]

[0067]

[0068] As a possible implementation, the feature extraction specifically includes:

[0069] S10. The pipeline damage picture expands its channels through dimensionality-increasing convolution, and the dimensionality-increasing convolution is depth convolution, that is, each channel is convolved separately;

[0070] As an example, MobileNetV2 is a lightweight deep neural network, and the core modules are the inverted residual structure and depthwise separable convolution. Assuming the pipeline damage picture is X, the input image X first passes through the inverted residual module of MobileNetV2, and this module consists of multiple depthwise separable convolutions and inverted residual structures. First, the input image X passes through the dimensionality-increasing convolution Wexpand Expand its number of channels. At this time, the convolution operation belongs to depth convolution, that is, each channel is convolved separately.

[0071] S11. Apply an activation function to enhance the non-linear characteristics of MobileNetV2 and limit the output value between [0, 6];

[0072] As an example, on the result after dimension elevation in step S10, apply the ReLU6 activation function to enhance the non-linear characteristics of MobileNetV2 and limit the output value between [0, 6], thus avoiding the situation of excessive or too small values.

[0073] S12. Reduce the number of channels from the expanded number of channels to the original number of channels through depthwise separable convolution. The depthwise separable convolution is pointwise convolution, and apply pointwise convolution to integrate the information between channels;

[0074] As an example, use the depthwise separable convolution W proj Reduce the number of channels from the expanded number of channels to the original number of channels. At this time, the operation is also pointwise convolution, that is, 1×1 convolution, and its purpose is to integrate the information between channels through 1×1 convolution. After S10 to S12, the features of the pipeline damage picture are extracted and output as the feature map Y0 as follows:

[0075] Y0 = W proj ·ReLU6(W expand ·X)

[0076] where W expand is the dimension elevation weight, used to expand the number of channels, and W proj is the dimension reduction weight, used to reduce the number of channels. ReLU6(z) = min(max(0, z), 6) is the activation function, used to limit the numerical range.

[0077] Next, the above feature map Y0 is input into the attention mechanism module. The attention mechanism module includes channel attention and spatial attention, and further enhances the feature representation through channel attention and spatial attention.

[0078] As a possible implementation, feature enhancement specifically includes:

[0079] Channel attention generates the attention weights of channels through average pooling and max pooling, used to highlight important channel information;

[0080] Perform a fully connected network process on the result after pooling to obtain the channel attention weights, as shown in the following formula:

[0081] M c (Y0) = σ(MLP(AvgPool(Y0) + MLP(MaxPool(Y0))));

[0082] Among them, σ is the Sigmoid activation function, MLP is the fully connected layer, AvgPool is the global average pooling, and MaxPool is the global max pooling.

[0083] Multiply the calculated channel attention weight by the feature map Y0 to obtain the feature map Y1 with enhanced channel attention, as shown in the following formula:

[0084] Y1 = M c (Y0)·Y0.

[0085] Next, the spatial attention generates the attention weight of the spatial position to highlight the important regions in the image. Specifically, the feature enhancement also includes:

[0086] Perform global average pooling and global max pooling on the feature map Y1 to obtain the pooling results at each position;

[0087] After concatenating the pooling at each position, generate the spatial attention weight through convolution;

[0088] Multiply the spatial attention weight by the feature map Y1 to highlight the important spatial regions in the feature map Y1 and obtain the feature map Y2 with enhanced spatial attention, as shown in the following formula:

[0089] M s (Y1) = σ(Conv2D([AvgPool(Y1), MaxPool(Y1)])),

[0090] Y2 = M s (Y1)·Y1.

[0091] Among them, Conv2D is the 7×7 convolution operation, AvgPool is the global average pooling, MaxPool is the global max pooling, and [·,·] is the feature concatenation.

[0092] Next, flatten the feature map Y2 into a vector, and input the flattened vector into the fully connected layer to obtain the probability distribution of the damage category, as shown in the following formula:

[0093] P = Softmax(W cls ·Flatten(Y2));

[0094] Among them, W cls is the weight matrix of the classification head, Flatten(Y2) is the vector obtained by flattening the feature map, and Softmax is to convert the output into the category probability.

[0095] See Figure 2, is a schematic diagram of the training process of the lightweight neural network model proposed in the embodiment of the present invention. It can be observed from the figure that the training and validation losses of the model decrease rapidly in the initial stage and then tend to be stable, indicating that the model effectively learns the data features during the training process and avoids overfitting to a certain extent. The training accuracy and validation accuracy increase rapidly in the initial stage and then remain at a high level, showing that the model has good generalization ability. In addition, the average precision of training and validation also shows the same trend, further confirming the performance of the model in the steel defect classification task. Adam was used for training with an initial learning rate of 0.001 to help the model converge better. It can be seen from the figure that the model shows good performance during the training process, and the reduction of the loss function and the increase of the accuracy both indicate the effectiveness of the model training.

[0096] Table 2 shows the performance indicators of the lightweight neural network model during training and validation:

[0097] Table 2 Performance indicators of the lightweight neural network model

[0098]

[0099] It can be seen from Table 2 that the model achieved an accuracy of 99% on the validation set, showing its excellent classification ability to accurately identify samples of different categories. During the training process, the model reached this optimal accuracy after 90 training epochs. The number of parameters of the model is 2.44M, which is relatively small while maintaining a high accuracy, helping to reduce the storage and computational costs of the model. Its computational complexity is 266.9MMac, which is acceptable in resource-constrained scenarios such as mobile devices. The overall training time is 729.0776489 seconds, and the time for the best validation is 1.352496624 seconds, indicating that the model can find the optimal solution relatively quickly during the training process. This reflects the demand for computational resources during the training of the model and has certain reference value for the hardware configuration during actual deployment.

[0100] See Figure 1 and Figure 3 , next, in this embodiment, the corrosion-related index dataset is used as the input, and the correlation coefficient method is applied to evaluate the degree of association between the corrosion factors and the corrosion rate, and the initial strongly correlated dataset is obtained.

[0101] Corrosion phenomena widely exist in industrial equipment, materials, and the environment. The influencing factors are complex and diverse, including material properties, environmental factors (such as humidity, temperature, chemical composition), time, and human factors, etc. Accurately mining the association rules between corrosion influencing factors is of great significance for preventing and controlling corrosion. However, corrosion data often have problems such as insufficient samples and uneven distribution, resulting in difficulties for traditional association rule mining methods (such as the Apriori algorithm) to effectively extract high-quality rules. In this embodiment, by combining a generative adversarial network (GAN) and the Apriori association rule algorithm, the problems of insufficient corrosion data samples and low rule mining efficiency are solved, so as to more accurately mine the association rules between corrosion influencing factors and provide theoretical support for corrosion prediction and protection.

[0102] See Figure 3 , as a possible implementation, the initial strongly correlated data set is obtained through steps such as preprocessing the original data set, calculating the correlation coefficient, and corrosion factor analysis, as follows:

[0103] S20. Obtain the original data set, and obtain the corrosion-related index data set through preprocessing. The preprocessing includes at least data cleaning, removing outliers, and filling missing values; there are multiple corrosion factors X and corresponding corrosion rates Y in the corrosion-related index data set;

[0104] As an example, assume that the original data set is D0. After processing steps such as data cleaning, removing outliers, and filling missing values, the preprocessed data set D is obtained. Pearson product-moment correlation coefficient analysis is performed on the data, and Spearman rank correlation coefficient verification is carried out. The Pearson product-moment correlation coefficient is used to calculate the strength of the linear relationship between each factor and the corrosion rate. Assume that the sample data of the corrosion factor X and the corresponding corrosion rate Y are X = {x1, x2,..., x n} and Y = {y1, y2,..., y n}.

[0105] S21. For data with a normal distribution, calculate the Pearson product-moment correlation coefficient τ xy between the corrosion factor X and the corrosion rate Y, as follows:

[0106]

[0107] where is the sample mean of the corrosion factor X, is the sample mean of the corrosion rate Y.

[0108] For data with a non-normal distribution, calculate the Spearman rank correlation coefficient τ s between the corrosion factor X and the corrosion rate Y; as follows:

[0109]

[0110] where d i is the difference in ranks between the i-th sample of the corrosion factor X and the corrosion rate Y, i.e., d i = R(x i ) - R(y i ), where R(x i ) and R(y i ) are the ranks of x i and y i in their respective data, and n is the number of samples.

[0111] For each corrosion factor X in the corrosion-related index dataset, its corresponding correlation coefficient is obtained

[0112] S22. Configure the correlation coefficient threshold, and retain the corrosion factor X whose correlation coefficient is greater than or equal to the correlation coefficient threshold, and discard the corrosion factor X whose correlation coefficient is less than the correlation coefficient threshold. The strongly correlated dataset is composed of all the retained corrosion factors X and their corresponding corrosion rates Y.

[0113] As an example, configure the correlation coefficient threshold to be θ. For the calculated correlation coefficients between each corrosion factor and the corrosion rate, if then X i can be used as an input feature, otherwise it is discarded. The initial strongly correlated dataset F is composed of all the retained corrosion factors X and their corresponding corrosion rates Y.

[0114] To solve the problem of insufficient samples, refer to Figure 3 . As a possible implementation, in this embodiment, a generative adversarial network is applied to expand the initial strongly correlated dataset to obtain a strongly correlated dataset.

[0115] As an example, the generative adversarial network (GAN) consists of a generator and a discriminator. The goal of the generator is to generate synthetic data similar to the real data. Let z be the random noise and G(z) be the synthetic data generated by the generator. Then the logic of the generator can be expressed as:

[0116] G: z → G(z)

[0117] The parameters of the generator are updated by minimizing the misclassification probability of the discriminator for the generated data, i.e.:

[0118]

[0119] Among them, D(G(z)) is the classification probability of the discriminator for the generated data, and p(z) is the random noise distribution.

[0120] The parameters of the discriminator are updated by maximizing the correct classification probabilities for the real data and the generated data, that is:

[0121]

[0122] Among them, p(F) is the real data distribution.

[0123] See Figure 3 , taking the strongly correlated data set as the input, applying the Apriori association algorithm to evaluate the strong association rules between corrosion factors; as a possible implementation, the strong association rules are determined by the following method:

[0124] The Apriori association algorithm gradually scans the strongly correlated data set and generates frequent item sets by using the properties of frequent item sets;

[0125] As an example, the property of a frequent item set is that if X is a frequent item set, then all subsets of X are also frequent item sets. Each element in the item set I is 1 item. For the item set If its support meets the following conditions, then X is a frequent item set:

[0126]

[0127] For X and Y in the frequent item set, if and X ∪ Y is a frequent item set, then an association rule can be formed Its confidence is:

[0128]

[0129] The confidence of the strong association rules screened in this embodiment is all 100%, and they are sorted in descending order of support. Generally speaking, the higher the support and confidence, the more credible the strong association rules are, and the more reasonable it is to use these strong association rules.

[0130] In the embodiments of the present invention, the correlation coefficient method, the GAN adversarial neural network, and the Apriori association rule algorithm are combined to implement a data mining algorithm for studying the correlation between various corrosion factors and the corrosion rate. The correlation coefficient method is used to evaluate the degree of association between each factor in the original dataset and the corrosion rate, select highly correlated corrosion influencing factors, and expand the dataset through the GAN adversarial neural network. Finally, the Apriori association rule mining algorithm is used to discover potential association rules. It helps to deeply understand the mechanism and influencing factors of pipeline corrosion and provides a scientific basis for formulating targeted corrosion protection strategies. The corrosion sample data of 20# steel material used in this embodiment is shown in Table 3, the parameter settings of the GAN adversarial neural network model are shown in Table 4, the dataset expanded by GAN is encoded according to Table 5, and the association rules are obtained through the Apriori algorithm, as shown in Table 6.

[0131] Table 3 Corrosion sample data of 20# steel material

[0132] Serial number <![CDATA[CO2 partial pressure MPa]]> <![CDATA[Partial pressure of H2S, MPa]]> <![CDATA[Cl - Content mg / L]]> Temperature °C Pitting rate mm / a 1 0.01 0.0003 52500 50 0.0626 2 0.5 0.0003 52500 50 0.3076 3 0.01 0.05 52500 50 0.2034 4 0.5 0.05 52500 50 0.524 5 0.255 0.02515 5000 20 0.1121 6 0.255 0.02515 100000 20 0.1486 7 0.255 0.02515 5000 80 0.3206 8 0.255 0.02515 100000 80 0.4693 9 0.01 0.02515 52500 20 0.0913 10 0.5 0.02515 52500 20 0.2086 11 0.01 0.02515 52500 80 0.3206 12 0.5 0.02515 52500 80 0.5475 13 0.255 0.0003 5000 50 0.2738 14 0.255 0.05 5000 50 0.3076 15 0.255 0.0003 100000 50 0.3076 16 0.255 0.05 100000 50 0.3598 17 0.01 0.02515 5000 50 0.1981 18 0.5 0.02515 5000 50 0.4093 19 0.01 0.02515 100000 50 0.2815 20 0.5 0.02515 100000 50 0.4901 21 0.255 0.0003 52500 20 0.0938 22 0.255 0.05 52500 20 0.1772 23 0.255 0.0003 52500 80 0.3285 24 0.255 0.05 52500 80 0.4119 25 0.255 0.02515 52500 50 0.2476

[0133] Table 4 GAN model parameter settings

[0134]

[0135]

[0136] Table 5 Dataset encoding rule table

[0137]

[0138] Table 6 Corrosion influencing factor association rule table

[0139]

[0140]

[0141] By mining strong association rules, the corrosion factors can be quantitatively regulated, so as to maintain the corrosion rate at a low level. In practical applications, according to the specific situation of the factory, a pipeline anti-corrosion mechanism based on strong association rules should be adopted for anti-corrosion treatment.

[0142] In practical applications, the damage of chemical pipelines can also be qualitatively and quantitatively evaluated based on the probability distribution of damage types and strong association rules.

[0143] As a possible implementation, see Figure 4 ., the chemical pipeline damage assessment method further includes: formulating an anti-corrosion mechanism based on strong association rules, specifically including:

[0144] S30. Obtain the recommended value of the corrosion rate, and sequentially obtain the recommended values of corrosion influencing factors and the corrosion rate based on strong association rules;

[0145] S31. Obtain the monitored value of the corrosion rate, monitor the corrosion rate and corrosion factors, and obtain the monitored value of the corrosion rate;

[0146] S32. Determine whether the monitored value exceeds the recommended value. If so, execute S30; otherwise, loop and execute S31 - S32;

[0147] S33. Alarm and decision support;

[0148] As an example, if the monitored value exceeds the recommended value, the data monitoring system will issue an alarm to remind personnel to come and conduct corrosion condition analysis and corrosion condition log recording in a timely manner, or it can also be automatically recorded by the system.

[0149] S34. Execute decision adjustment according to strong association rules;

[0150] As an example, the executed decision adjustment is to judge which values exceed the range required by the strong association rules according to the strong association rules, and adopt specific numerical adjustment strategies according to the types of values.

[0151] S35. Determine whether the alarm is eliminated. If so, loop and execute S32 - S35; otherwise, execute process anti-corrosion inspection.

[0152] In a second aspect, an embodiment of the present invention provides a chemical pipeline anti-corrosion method based on multi-dimensional data fusion. Refer to Figure 4 , including the following steps:

[0153] Apply the configured and trained lightweight neural network model to evaluate the damaged parts and damage types of chemical pipelines, and establish an inspection ledger for dangerous points; the damage types include mechanical damage and chemical corrosion damage;

[0154] Apply the mined strong association rules to improve the existing chemical pipeline anti-corrosion strategy.

[0155] Refer to Figure 4 , as an example, train the CBAM-MobileNetV2 model according to the pipeline damage pictures in the factory historical inspection data to improve its accuracy. Formulate a point inspection and patrol inspection system according to the equipment operation status, including the determination of inspection parts, the determination of inspection cycles, and the formulation of inspection strategies. After using the edge device deployed with CBAM-MobileNetV2 to inspect the damaged parts of the pipeline, conduct damage assessment on it and establish an inspection ledger for dangerous points.

[0156] On the other hand, based on the mined strong association rules, improving the existing anti-corrosion strategies can reduce the risk of pipeline corrosion by adjusting the Cl - content, CO2 partial pressure, H2S partial pressure, and temperature. In practical applications, it is also necessary to combine the corrosion factors of specific pipeline operating conditions to improve the anti-corrosion strategy. According to the basic data and operating parameters of the process pipeline, conduct corrosion detection and monitoring on the operating conditions of the process pipeline, mainly monitor the conditions of various correlation factors of pipeline corrosion and the pipeline corrosion rate, and at the same time record the corrosion status log and analyze the corrosion status. Finally, carry out process anti-corrosion according to the strong association rules, optimize the existing operations, and conduct anti-corrosion effect analysis to verify the conclusions of the existing strong association rules.

[0157] Through pipeline corrosion damage inspection and assessment based on multi-dimensional data, form pipeline anti-corrosion maintenance decisions, optimize the existing maintenance measures, and finally keep the safety of the process pipeline at a high level at a relatively low cost. These improvement measures will help improve the safety and reliability of the pipeline, extend its service life, and reduce maintenance and repair costs.

[0158] Although the present invention has been described in connection with various embodiments, however, in the process of implementing the claimed invention, those skilled in the art can understand and realize other variations of the disclosed embodiments by viewing the drawings, the disclosure content, and the like. In the specification, the term "comprising" does not exclude other components or steps, and "a" or "one" does not exclude the case of multiple. A single processor or other unit can implement several functions listed in the specification. Certain measures are described in different embodiments, but this does not mean that these measures cannot be combined to produce good results.

[0159] Although the present invention has been described in connection with specific features and their embodiments, it is obvious that various modifications and combinations can be made without departing from the spirit and scope of the present invention. Accordingly, the present specification and the drawings are merely exemplary descriptions of the present invention and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present invention. Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. A pipeline damage assessment method based on multi-dimensional data fusion, characterized in that, It includes the following steps: Configure a trained lightweight neural network model, input the pipeline damage pictures into the lightweight neural network model, and obtain the probability distribution of the damage type after feature extraction, feature enhancement, feature mapping and output; Taking the corrosion-related index data set as the input, apply the correlation coefficient method to evaluate the correlation degree between the corrosion factors and the corrosion rate, and obtain the initial strongly correlated data set, and the initial strongly correlated data set is a small sample data set; Expand the initial strongly correlated data set to obtain a strongly correlated data set; Taking the strongly correlated data set as the input, apply the Apriori association algorithm to evaluate the strong association rules between the corrosion factors; Qualitatively and quantitatively evaluate the chemical pipeline damage based on the probability distribution of the damage type and the strong association rules.

2. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 1, wherein The backbone network of the lightweight neural network model is MobileNetV2, and an attention mechanism module is placed after MobileNetV2; the feature extraction specifically includes: S10. The pipeline damage picture expands its channels through a dimensionality-increasing convolution, and the dimensionality-increasing convolution is a depth convolution, that is, each channel is convolved separately; S11. Apply an activation function to enhance the non-linear characteristics of MobileNetV2, and limit the output value between [0, 6]; S12. Reduce the number of channels from the expanded number of channels to the original number of channels through a dimensionality-reducing convolution. The dimensionality-reducing convolution is a pointwise convolution, and the pointwise convolution is used to integrate the information between channels; After S10 to S12, the features of the pipeline damage picture are extracted and output as the feature map Y0.

3. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 2, characterized in that The attention mechanism module includes channel attention and spatial attention, and the feature enhancement specifically includes: The channel attention generates the attention weights of the channels through average pooling and max pooling, which is used to highlight important channel information; Perform a fully connected network process on the pooled result to obtain the channel attention weights; Multiply the calculated channel attention weights by the feature map Y0 to obtain the feature map Y1 with enhanced channel attention.

4. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 3, characterized in that The feature enhancement also includes: Perform global average pooling and global max pooling on the feature map Y1 to obtain the pooling results at each position; After splicing the pooling at each position, generate spatial attention weights through convolution; Multiply the spatial attention weights by the feature map Y1 to highlight the important spatial regions in the feature map Y1, and obtain the feature map Y2 with enhanced spatial attention.

5. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 4, characterized in that, The feature mapping and output specifically include: Flatten the feature map Y2 into a vector, and input the flattened vector into the fully connected layer to obtain the probability distribution of the damage category.

6. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 1, characterized in that The initial strongly correlated data set is obtained through the following method: S20. Obtain the original data set, and obtain the corrosion-related index data set through preprocessing. The preprocessing includes at least data cleaning, removing outliers and filling missing values; there are multiple corrosion factors X and the corresponding corrosion rate Y in the corrosion-related index data set; S21. For data with a normal distribution, calculate the Pearson product-moment correlation coefficient τ between the corrosion factor X and the corrosion rate Y xy ; for data with a non-normal distribution, calculate the Spearman rank correlation coefficient τ between the corrosion factor X and the corrosion rate Y s ; for each corrosion factor X in the corrosion-related index dataset, obtain the corresponding correlation coefficient S22. Configure the correlation coefficient threshold, and for the correlation coefficient retain the corrosion factor X greater than or equal to the correlation coefficient threshold, and discard the correlation coefficient for the corrosion factor X less than the correlation coefficient threshold. The strongly correlated data set is composed of all the retained corrosion factors X and their corresponding corrosion rates Y.

7. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 6, characterized in that The strong association rules are determined through the following method: The Apriori association algorithm scans the strongly correlated data set step by step, and generates frequent itemsets by using the properties of frequent itemsets; For X and Y in the frequent itemset, if and X ∪ Y is a frequent itemset, then an association rule can be formed Its confidence is:

8. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 1, characterized in that The chemical pipeline damage assessment method also includes: formulating an anti-corrosion mechanism based on the strong association rules, specifically including: S30. Obtain the recommended value of the corrosion rate, and sequentially obtain the recommended values of corrosion influencing factors and the corrosion rate based on strong association rules; S31. Obtain the monitored value of the corrosion rate, monitor the corrosion rate and corrosion factors, and the monitored value of the corrosion rate has been obtained; S32. Determine whether the monitored value exceeds the recommended value. If so, execute S30; otherwise, loop and execute S31 - S32; S33. Alarm and decision support; S34. Execute decision adjustment according to strong association rules; S35. Determine whether the alarm is eliminated. If so, loop and execute S32 - S35; otherwise, perform process anti-corrosion inspection.

9. The pipeline damage assessment method based on multi-dimensional data fusion according to claim 1, characterized in that, Apply a generative adversarial network to expand the initial strongly correlated dataset to obtain a strongly correlated dataset.

10. A pipeline anti-corrosion method based on multi-dimensional data fusion, characterized in that, It includes the following steps: Apply the configured and trained lightweight neural network model to evaluate the damaged parts and damage types of chemical pipelines, and establish an inspection ledger for dangerous points; the damage types include mechanical damage and chemical corrosion damage; Apply the mined strong association rules to improve the existing anti-corrosion strategy for chemical pipelines.

Images