Search

CN-122015019-A - Pipeline corrosion risk early warning method, electronic equipment and storage medium

CN122015019ACN 122015019 ACN122015019 ACN 122015019ACN-122015019-A

Abstract

The embodiment of the application provides a pipeline corrosion risk early warning method, electronic equipment and a storage medium. The method comprises the steps of obtaining corrosion monitoring data collected by the sensor array based on the deployed sensor array on a target pipeline, carrying out data preprocessing on the corrosion monitoring data according to preset data preprocessing rules to obtain corrosion characteristic data, obtaining corrosion health state information according to the corrosion characteristic data by adopting a pre-trained corrosion health state assessment model, obtaining corrosion prediction information according to the corrosion characteristic data by adopting a pre-trained corrosion prediction model, calculating and obtaining the safe operation period of the target pipeline according to the current pipeline wall thickness and the corrosion prediction information of the target pipeline, and obtaining and outputting corrosion risk early warning information of the target pipeline according to the corrosion health state information, the corrosion prediction information and the safe operation period of the pipeline by means of early warning judgment rules. The method and the device improve the accuracy, reliability and timeliness of corrosion risk early warning of the target pipeline.

Inventors

  • HU JINQIU
  • ZHANG LAIBIN
  • MA MINGJUN
  • DONG JIAYAN

Assignees

  • 中国石油大学(北京)

Dates

Publication Date
20260512
Application Date
20251230

Claims (10)

  1. 1. A method for early warning of risk of corrosion of a pipeline, comprising: based on a deployed sensor array on a target pipeline, acquiring corrosion monitoring data acquired by the sensor array, and performing data preprocessing on the corrosion monitoring data according to a preset data preprocessing rule to acquire corrosion characteristic data; According to the corrosion characteristic data, a pre-trained corrosion health state evaluation model is adopted to acquire corrosion health state information of the target pipeline; According to the corrosion characteristic data, a pre-trained corrosion prediction model is adopted to obtain corrosion prediction information of the target pipeline, wherein the corrosion prediction model is determined after super-parameter optimization is carried out through a self-adaptive weight gray wolf optimization algorithm; According to the current pipeline wall thickness of the target pipeline and the corrosion prediction information, calculating and obtaining the safe operation period of the target pipeline; and acquiring and outputting corrosion risk early warning information of the target pipeline according to the corrosion health state information, the corrosion prediction information and the safe operation period of the target pipeline by a preset early warning judgment rule.
  2. 2. The method of claim 1, wherein the corrosion health assessment model comprises a gating feature handling layer, a self-encoding feature handling layer, and a hierarchical classification layer; correspondingly, according to the corrosion characteristic data, a pre-trained corrosion health state evaluation model is adopted to acquire corrosion health state information of the target pipeline, and the method comprises the following steps: Performing feature mapping linear transformation and gating mapping linear transformation on the corrosion feature data through the gating feature processing layer to obtain a feature mapping result and a gating mapping result, processing the gating mapping result by adopting an activation function to obtain a gating coefficient, and multiplying the gating coefficient and the feature mapping result element by element to obtain a gating feature; The self-coding feature processing layer is used for carrying out nonlinear coding processing on the gating features by adopting a preset encoder to obtain low-dimensional latent variable features; Mapping the low-dimensional latent variable characteristics into classification scores corresponding to all preset corrosion health state categories through the hierarchical classification layer, and carrying out index normalization processing on the classification scores to obtain corrosion health state probability distribution corresponding to all preset corrosion health state categories; and determining the corrosion health status grade of the target pipeline according to the corrosion health status probability distribution.
  3. 3. The method of claim 1, wherein the corrosion prediction model comprises a mild corrosion sub-model, a moderate corrosion sub-model, and a severe corrosion sub-model; correspondingly, according to the corrosion characteristic data, a pre-trained corrosion prediction model is adopted to obtain corrosion prediction information of the target pipeline, and the method comprises the following steps: according to the corrosion characteristic data, a first corrosion prediction rate and a first corrosion grade corresponding to the first corrosion prediction rate are obtained through a pre-trained mild corrosion sub-model; according to the corrosion characteristic data, a second corrosion prediction rate and a second corrosion grade corresponding to the second corrosion prediction rate are obtained through a pre-trained moderate corrosion sub-model; According to the corrosion characteristic data, a third corrosion prediction rate and a third corrosion grade corresponding to the third corrosion prediction rate are obtained through a pre-trained severe corrosion sub-model; Acquiring a target corrosion grade corresponding to the target pipeline by adopting a preset grade voting rule according to the first corrosion grade, the second corrosion grade and the third corrosion grade; and acquiring one or more corrosion prediction rates corresponding to the target corrosion level based on the first corrosion prediction rate, the second corrosion prediction rate and the third corrosion prediction rate, and taking the maximum corrosion prediction rate as the target corrosion prediction rate corresponding to the target pipeline.
  4. 4. A method according to claim 3, wherein said calculating a safe operating period for the target pipe based on the current pipe wall thickness of the target pipe and the corrosion prediction information comprises: Acquiring the current pipeline wall thickness of the target pipeline and acquiring a preset safety thickness threshold corresponding to the target pipeline; And obtaining a target corrosion prediction rate according to the corrosion prediction information, and calculating the safe operation period of the target pipeline according to the current pipeline wall thickness, the preset safe thickness threshold value and the target corrosion prediction rate.
  5. 5. The method of claim 1, wherein obtaining the pre-trained corrosion prediction model comprises: Determining a search space of the super parameters to be optimized according to the super parameters to be optimized, and generating a plurality of groups of initial super parameter combinations in the search space of the super parameters to be optimized; determining each initial hyper-parameter combination as an initial position parameter of a corresponding searching individual to obtain an initial wolf population, wherein the initial wolf population comprises a plurality of searching individuals; Marking the initial gray wolf population as a current gray wolf population, and obtaining the current iteration times; when the current iteration number does not reach the preset iteration number, constructing a support vector regression prediction model according to the position parameters corresponding to each searching individual in the current gray wolf population; Calculating a training set mean square error corresponding to the support vector regression prediction model according to a preset training data set, taking the training set mean square error as fitness, and obtaining a fitness set corresponding to the current gray wolf population; Screening the current sirius population based on the fitness set to obtain a preset number of target search individuals, and performing dynamic disturbance processing on the position parameters of a plurality of target search individuals to obtain the position parameters of each target search individual after disturbance; According to the current iteration times and the preset iteration times, a preset nonlinear self-adaptive weight updating rule is adopted to determine weight parameters corresponding to the current iteration times; Updating the position parameters of each searching individual in the current wolf population according to the weight parameters and the position parameters after disturbance of each target searching individual to obtain an updated wolf population; Marking the updated gray wolf population as the current gray wolf population, and updating the current iteration times; When the current iteration number reaches the preset iteration number, screening and obtaining the search individual with the optimal fitness in the updated gray wolf population, and taking the position parameter of the search individual with the optimal fitness as a target super-parameter combination of the support vector regression prediction model; and respectively training and verifying the support vector regression prediction model according to a preset training data set and a preset verification data set based on the target hyper-parameter combination so as to obtain the pre-trained corrosion prediction model passing verification.
  6. 6. The method of claim 5, wherein screening the current sirius population based on the fitness set to obtain a preset number of target search individuals, and dynamically perturbing position parameters of a plurality of target search individuals to obtain post-perturbation position parameters of each target search individual, comprises: according to the fitness collection, performing fitness sequencing on each searching individual in the current sirius population, and determining optimal searching individuals, suboptimal searching individuals and third optimal searching individuals according to the preset quantity; respectively acquiring position parameters corresponding to the optimal searching individual, the suboptimal searching individual and the third optimal searching individual; And respectively superposing random disturbance parameters on the position parameters corresponding to the optimal searching individual, the suboptimal searching individual and the third optimal searching individual to obtain the position parameters after disturbance corresponding to the optimal searching individual, the suboptimal searching individual and the third optimal searching individual.
  7. 7. The method of claim 5, wherein determining the weight parameter corresponding to the current iteration number using a preset nonlinear adaptive weight update rule according to the current iteration number and the preset iteration number comprises: acquiring the current iteration number and the preset iteration number, and adopting a formula: calculating and obtaining a weight parameter alpha (t) corresponding to the current iteration times; wherein T is the current iteration number, T is the preset iteration number, alpha 0 is the initial weight value, k is the adjustment coefficient, and gamma is the convergence rate index.
  8. 8. The method according to any one of claims 1 to 7, wherein the performing data preprocessing on the corrosion monitoring data according to a preset data preprocessing rule to obtain corrosion characteristic data includes: according to the corrosion monitoring data, mapping each corrosion monitoring parameter in the corrosion monitoring data to a preset numerical value interval through normalization processing to obtain standardized monitoring data; Acquiring a plurality of preset health evaluation indexes, and performing level correlation analysis on each standardized monitoring parameter in the standardized monitoring data and the preset health evaluation indexes through level correlation coefficient analysis to acquire correlation coefficients corresponding to each standardized monitoring parameter; And adding standardized monitoring parameters with the correlation number larger than or equal to the correlation threshold value to the corrosion characteristic data according to a preset correlation threshold value.
  9. 9. An electronic device is characterized by comprising a memory and a processor; The memory stores computer-executable instructions; The processor executing computer-executable instructions stored in the memory, causing the processor to perform the method of any one of claims 1 to 8.
  10. 10. A computer readable storage medium having stored therein computer executable instructions which when executed by a processor are adapted to carry out the method of any one of claims 1 to 8.

Description

Pipeline corrosion risk early warning method, electronic equipment and storage medium Technical Field The application relates to the technical field of petrochemical industry, in particular to a pipeline corrosion risk early warning method, electronic equipment and a storage medium. Background In the oil refining and chemical production processes, a material inlet and outlet pipeline of the refining device is used as a material transmission path, and the running state of the material inlet and outlet pipeline is directly related to continuous and stable running of the device and safety of equipment. With the continuous expansion of the scale of the device and the improvement of the process complexity, the pipeline is subjected to more severe working conditions, and particularly, the pipeline is easy to be corroded and damaged under the conditions of high temperature, high pressure, strong corrosive media containing sulfur, chlorine and the like. Corrosion not only causes wall thickness reduction and pressure bearing capacity reduction, but also may cause leakage, interruption of production, and even major accidents. Therefore, the monitoring and accurate early warning of the corrosion state of the pipeline are key for guaranteeing the safe and stable operation of the device. At present, the monitoring and early warning of pipeline corrosion in the refining device mainly adopts a mode of combining manual detection with experience rule judgment. The common detection method comprises the steps of regularly carrying out on-site inspection, acquiring wall thickness data at key parts of a pipeline by using an ultrasonic thickness gauge, or acquiring corrosion rate indexes in a medium environment by means of corrosion hanging pieces, linear polarization probes and the like. These physical detection means form the basis of the raw data of the corrosion analysis. In the early warning judging process, a static threshold value comparison method is generally adopted, namely, the measured wall thickness value is compared with the set minimum safe wall thickness, and an early warning signal is triggered when the data is lower than a threshold value. Furthermore, the risk level preliminary judgment can be completed by carrying out manual experience weighting treatment by combining factors such as service life, medium type and the like. However, the mode relying on manual detection and preset static threshold judgment has stronger subjectivity, not only has the problems of limited coverage and missed detection in the data acquisition process, but also has the problems of unstable identification, judgment deviation and the like which are easy to occur because the judgment standard is usually established on the basis of a single index, so that the accuracy and reliability of the whole corrosion risk early warning result are lower. Disclosure of Invention The embodiment of the application provides a pipeline corrosion risk early warning method, electronic equipment and a storage medium, which are used for achieving the effect of improving the reliability of the accuracy of a corrosion risk early warning result. In a first aspect, an embodiment of the present application provides a method for early warning risk of corrosion of a pipeline, including: Based on a deployed sensor array on a target pipeline, acquiring corrosion monitoring data acquired by the sensor array, and performing data preprocessing on the corrosion monitoring data according to a preset data preprocessing rule to acquire corrosion characteristic data; According to the corrosion characteristic data, a pre-trained corrosion health state evaluation model is adopted to acquire corrosion health state information of the target pipeline; According to the corrosion characteristic data, a pre-trained corrosion prediction model is adopted to obtain corrosion prediction information of a target pipeline, wherein the corrosion prediction model is determined after super-parameter optimization is carried out through a self-adaptive weight gray wolf optimization algorithm; According to the current pipeline wall thickness and corrosion prediction information of the target pipeline, calculating and obtaining the safe operation period of the target pipeline; And acquiring and outputting corrosion risk early warning information of the target pipeline according to corrosion health state information, corrosion prediction information and safe operation period of the target pipeline by a preset early warning judgment rule. In a second aspect, an embodiment of the present application provides a device for early warning risk of corrosion of a pipeline, including: The data processing module is used for acquiring corrosion monitoring data acquired by the sensor array based on the deployed sensor array on the target pipeline, and performing data preprocessing on the corrosion monitoring data according to a preset data preprocessing rule to acquire corrosion characteristic data; the pipe