Vector autoregression model-based ship condenser corrosion dynamic prediction method and system
By combining the vector autoregressive model and the autoencoder, high-precision dynamic prediction of ship condenser corrosion is achieved, which solves the problems of insufficient real-time and accuracy in traditional methods and provides a scientific basis for monitoring and maintenance.
Patent Information
- Application Number
- CN202510538784.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-27
- Publication Date
- 2025-09-16
AI Technical Summary
Traditional corrosion prediction methods cannot achieve real-time updates and dynamic predictions, and are unable to meet the real-time and accuracy requirements of modern ships for condenser corrosion monitoring.
A dynamic prediction method for ship condenser corrosion based on a vector autoregression model is adopted. Autoencoder is used for feature extraction, and the vector autoregression model is combined to model and predict the corrosion dynamics. Error correction is performed by introducing the adjacent pipeline model.
It achieves high-precision dynamic prediction of ship condenser corrosion, provides a scientific basis for condition monitoring and maintenance decision-making, and improves the real-time and accuracy of the prediction.
Smart Images
Figure CN120654315A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of condenser corrosion prediction, and in particular to a method and system for dynamic prediction of ship condenser corrosion based on a vector autoregressive model. Background Art
[0002] As a core component of a ship's cooling system, a ship's condenser is exposed to the complex marine environment for long periods of time and is susceptible to various corrosion factors. Over time, corrosion can reduce the condenser's heat exchange efficiency, weaken its structural strength, and even cause safety incidents such as pipeline leaks or condenser failure. Therefore, accurately predicting the corrosion of a ship's condenser is crucial to proactively implement maintenance measures and ensure safe ship operation.
[0003] However, traditional corrosion prediction methods have numerous limitations. For example, empirical prediction methods, while relying on historical data and manual experience, lack a precise description of the dynamic changes in corrosion. Physical modeling methods, while able to explain the physical and chemical processes of corrosion, often require complex parameter adjustments and struggle to adapt to the actual operating conditions of different ship condensers. Furthermore, most of these traditional methods lack real-time updates and dynamic predictions, making them difficult to meet the real-time and precision requirements of condenser corrosion monitoring on modern ships. Summary of the Invention
[0004] The purpose of the present invention is to propose a dynamic prediction method and system for ship condenser corrosion based on a vector autoregression model, which uses an autoencoder to extract features from data, mine implicit information in the data, and combines the vector autoregression model to model and predict the corrosion dynamics.
[0005] According to a first aspect of an embodiment of the present disclosure, a method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model is provided, comprising the following steps:
[0006] Pre-process the corrosion data of each pipe of the ship's condenser, identify and process abnormal data, and then store it in the database;
[0007] The autoencoder is used to extract features from the processed pipeline corrosion data to obtain a feature array Features, which is then integrated with the processed pipeline corrosion data and saved in a database.
[0008] Obtain the newly added integrated pipeline corrosion data in the database, use the vector autoregression model to model each pipeline individually, and save the obtained autoregression parameters to the database;
[0009] The corrosion data of each integrated pipeline is input into the corresponding pipeline model for corrosion prediction, and the adjacent pipeline model is used to correct the error of the obtained prediction results.
[0010] In one embodiment, the pipeline corrosion data includes key information such as ship number, condenser number, measurement time, corrosion point location information, corrosion depth, measurement voltage, and measurement phase.
[0011] In one embodiment, the autoencoder is constructed as follows:
[0012] Confirm the autoencoder input dimension, intermediate dimension, and output dimension;
[0013] Use the Tanh function as the activation function and the mean square error MSE as the loss function:
[0014] All the pipeline corrosion data of the ship condenser corrosion is selected as input, the number of iteration steps is set, and the trained autoencoder is persisted.
[0015] The autoencoder is called to extract features from the corrosion data of each pipe of the ship condenser to obtain the feature array Features.
[0016] In one embodiment, the vector autoregressive model is:
[0017] Y t =A0+A1Y t-1 +A2Y t-2 +…+A P Y t-p +ε t
[0018] where Y t is a column vector containing all variables, A0 is a constant vector, A1, A2, ..., A P is the coefficient matrix, p is the order of the model, which represents the dynamic relationship between variables, and ε t is the error term vector, which is assumed to be normally distributed with mean 0 and covariance matrix Σ.
[0019] In one embodiment, when calling a vector autoregression model:
[0020] Use unit root test to check the stationarity of time series. Non-stationary series need to be made stationary through difference operation.
[0021] The order of the vector autoregression model was determined by the information criterion AIC;
[0022] Estimation of model parameters based on input pipeline corrosion data;
[0023] Residual tests were used to verify the validity of the model and the causal relationship between variables.
[0024] In one embodiment, each pipeline is modeled individually as follows:
[0025] Asynchronously call the vector autoregression model training program and automatically retrieve the historical data of the corresponding pipeline. The newly added integrated pipeline corrosion data is used as input for model training.
[0026] After training, the regression parameters are saved to the database, including model parameters, pipeline ID, model update time, and pipeline adjacency fields.
[0027] In one embodiment, during prediction, each integrated pipeline corrosion data is input into the corresponding pipeline model to obtain a prediction result Y; at the same time, according to the adjacent relationship of the pipelines, each integrated pipeline corrosion data is input into the adjacent pipeline model to obtain multiple prediction correction data Z1, Z2, ..., Z n , and determine the weight of each predicted correction data according to the distance formula as the basis for correcting Y:
[0028]
[0029] Where: D(Z1) is the distance between pipe Y and pipe Z1.
[0030] According to a second aspect of an embodiment of the present disclosure, a dynamic prediction system for ship condenser corrosion based on a vector autoregressive model is provided, comprising:
[0031] The pre-processing module pre-processes the corrosion data of each pipe of the ship condenser, identifies abnormal data, processes it, and then stores it in the database;
[0032] The feature extraction module uses the autoencoder to extract features from the processed pipeline corrosion data to obtain the feature array Features, integrates the feature array Features with the processed pipeline corrosion data and saves them to the database;
[0033] The model building module obtains the newly added integrated pipeline corrosion data in the database, calls the vector autoregression model to model each pipeline separately, and saves the obtained autoregression parameters to the database;
[0034] The prediction and correction module inputs the corrosion data of each integrated pipeline into the corresponding pipeline model to perform corrosion prediction, and uses the adjacent pipeline model to perform error correction on the obtained prediction results.
[0035] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising a memory, a processor, and a computer program stored and running on the memory, wherein when the processor executes the program, the method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model is implemented.
[0036] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model is implemented.
[0037] The technical solution employed in this invention offers advantages over existing technologies: it utilizes an autoencoder to extract features from corrosion monitoring data for ship condensers, mining the underlying information, and combines this with a vector autoregressive model to model and predict the dynamic corrosion process. Furthermore, by incorporating a model of nearby pipelines to correct errors in the prediction results, prediction accuracy is further improved, ultimately achieving high-precision dynamic prediction of ship condenser corrosion. This method provides a scientific basis for condition monitoring, maintenance decision-making, and lifespan management of ship condensers, possessing significant engineering application value and broad development prospects. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] The drawings in the specification, which constitute a part of this application, are used to provide further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute improper limitations on this application.
[0039] Figure 1 This is a flow chart of a method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model in an embodiment;
[0040] Figure 2 Schematic diagram of the autoencoder structure in the embodiment;
[0041] Figure 3 It is the characteristic diagram of Tanh function in the embodiment;
[0042] Figure 4 This is a graph showing the change in the iterative loss value of the autoencoder in the embodiment;
[0043] Figure 5 This is a graph showing the prediction results of some variables using the vector autoregression model in the embodiment. DETAILED DESCRIPTION
[0044] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.
[0045] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present application belongs.
[0046] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.
[0047] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the methods and systems according to the various embodiments of the present disclosure. It should be noted that each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code can include one or more executable instructions for implementing the logical functions specified in the various embodiments. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, or they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the flowchart and / or block diagram, and the combination of the boxes in the flowchart and / or block diagram, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or can be implemented using a combination of dedicated hardware and computer instructions.
[0048] The vector autoregression (VAR) model, based on time series analysis, can effectively capture the dynamic relationships and implicit information in corrosion data, providing a powerful tool for solving this type of corrosion prediction problem. By combining feature extraction techniques from autoencoders with machine learning methods, it not only enables real-time monitoring and dynamic analysis of ship condenser corrosion data, but also provides highly accurate corrosion prediction results, providing a scientific basis for the maintenance and management of ship condensers.
[0049] Example 1:
[0050] like Figure 1 As shown, this embodiment provides a dynamic prediction method for ship condenser corrosion based on a vector autoregressive model, comprising the following steps:
[0051] S1. Preprocess the corrosion data of each pipe of the ship's condenser, identify and process abnormal data, and then store it in the database;
[0052] Specifically, the Excel file was parsed line by line using the EasyExcel open source database to obtain the corrosion data of each pipe of the ship's condenser. The pipeline corrosion data includes key information such as ship number, condenser number, measurement time, corrosion point location information, corrosion depth, measurement voltage, and measurement phase.
[0053] The Z-score method is used to identify abnormal data, and the score is:
[0054]
[0055] where x i is a single data point on the pipeline corrosion data, u is x i The mean of the sequence, σ is the standard deviation, s i is the score of the data point, and is identified according to the set threshold t. When the score s i When >t, the data is considered an outlier. The outlier is processed using linear interpolation, specifically:
[0056]
[0057] S2. Use an autoencoder to extract features from the processed pipeline corrosion data to obtain a feature array Features, integrate the feature array Features with the processed pipeline corrosion data, and save the results to a database;
[0058] like Figure 2 As shown in Figure 2, the core working principle of the autoencoder is to compress the input pipeline erosion data into a low-dimensional feature representation through an encoder, and then attempt to reconstruct the input pipeline erosion data through a decoder. The training goal of the autoencoder is to minimize the difference between the input data and the reconstructed data, thereby learning the hidden features of the input data, which is generally expressed as:
[0059]
[0060] z=Encoder(x)
[0061] Where z is the feature learned from the original data x.
[0062] The autoencoder is constructed as follows:
[0063] S2.1 confirms that the autoencoder input dimension is 3, the intermediate dimension is 1, and the output dimension is 3;
[0064] S2.2 uses the Tanh function as the activation function, such as Figure 3 As shown; using mean square error MSE as the loss function:
[0065]
[0066] S2.3 selects all the corrosion data of the ship condenser as input, the number of iteration steps is 20, and the trained autoencoder is persisted. The change of the loss value during the iteration process is as follows: Figure 4 As shown;
[0067] S2.4 calls the autoencoder to extract features from the corrosion data of each pipe of the ship condenser to obtain the feature array Features, so as to increase the dimension of the original data and mine the implicit information between variables.
[0068] S3. Obtain the newly added integrated pipeline corrosion data from the database, apply the vector autoregression model to model each pipeline individually, and save the obtained autoregression parameters to the database;
[0069] The vector autoregressive model is expressed as:
[0070] Y t =A0+A1Y t-1 +A2Y t-2 +…+A P Y t-p +ε t
[0071] where Y t is a column vector containing all variables, A0 is a constant vector, A1, A2, ..., A P is the coefficient matrix, p is the order of the model, which represents the dynamic relationship between variables, and ε t is the error term vector, assuming it obeys a normal distribution with mean 0 and covariance matrix Σ; when calling the vector autoregression model:
[0072] S3.1 Use unit root tests (such as ADF tests) to check the stationarity of time series. Non-stationary series need to be made stationary through difference operations.
[0073] S3.2 The order p of the vector autoregression model is determined to be 4 using the information criterion AIC;
[0074] S3.3 Estimate model parameters based on input pipeline corrosion data;
[0075] S3.4 Use residual tests to verify the validity of the model and the causal relationship between variables.
[0076] The way to model each pipeline individually is as follows:
[0077] Asynchronously call the vector autoregression model training program and automatically retrieve the historical data of the corresponding pipeline. The newly added integrated pipeline corrosion data is used as input for model training.
[0078] After training, the regression parameters are saved to the database, including model parameters, pipeline ID, model update time, pipeline adjacency fields, etc.
[0079] S4. Input each integrated pipeline corrosion data into the corresponding pipeline model to perform corrosion prediction, and use the adjacent pipeline model to perform error correction on the obtained prediction results.
[0080] Specifically, when a user wants to check the future corrosion trend of a certain pipeline, they send a message carrying the pipeline identifier. Based on the pipeline identifier, the corrosion data of the pipeline and the prediction model parameters in the database are queried. The corrosion data is input into the corresponding prediction model to obtain the prediction result Y. At the same time, based on the adjacent relationship of the pipelines, the prediction models of nearby pipelines are queried and the corrosion data are respectively input into the nearby pipeline models to obtain multiple prediction correction data Z1, Z2, ..., Zn. The weight of each data is determined according to the distance formula as the basis for correcting Y:
[0081]
[0082] The prediction results of some parameters are as follows Figure 5 shown.
[0083] In the preferred mode, the data displayed on the front-end page includes historical pipeline corrosion data and future corrosion trends, the number of pipeline corrosion points, the maximum corrosion depth, and the corrosion conditions of various areas of the pipeline.
[0084] Example 2:
[0085] This embodiment provides a dynamic prediction system for ship condenser corrosion based on a vector autoregressive model, including:
[0086] The pre-processing module pre-processes the corrosion data of each pipe of the ship condenser, identifies abnormal data, processes it, and then stores it in the database;
[0087] The feature extraction module uses the autoencoder to extract features from the processed pipeline corrosion data to obtain the feature array Features, integrates the feature array Features with the processed pipeline corrosion data and saves them to the database;
[0088] The model building module obtains the newly added integrated pipeline corrosion data in the database, calls the vector autoregression model to model each pipeline separately, and saves the obtained autoregression parameters to the database;
[0089] The prediction and correction module inputs the corrosion data of each integrated pipeline into the corresponding pipeline model to perform corrosion prediction, and uses the adjacent pipeline model to perform error correction on the obtained prediction results.
[0090] Example 3:
[0091] An electronic device includes a memory, a processor, and a computer program stored and running on the memory, wherein the processor, when executing the program, implements the above-mentioned method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model, including:
[0092] Pre-process the corrosion data of each pipe of the ship's condenser, identify and process abnormal data, and then store it in the database;
[0093] The autoencoder is used to extract features from the processed pipeline corrosion data to obtain a feature array Features, which is then integrated with the processed pipeline corrosion data and saved in a database.
[0094] Obtain the newly added integrated pipeline corrosion data in the database, use the vector autoregression model to model each pipeline individually, and save the obtained autoregression parameters to the database;
[0095] The corrosion data of each integrated pipeline is input into the corresponding pipeline model for corrosion prediction, and the adjacent pipeline model is used to correct the error of the obtained prediction results.
[0096] Example 4:
[0097] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model, comprising:
[0098] Pre-process the corrosion data of each pipe of the ship's condenser, identify and process abnormal data, and then store it in the database;
[0099] The autoencoder is used to extract features from the processed pipeline corrosion data to obtain a feature array Features, which is then integrated with the processed pipeline corrosion data and saved in a database.
[0100] Obtain the newly added integrated pipeline corrosion data in the database, use the vector autoregression model to model each pipeline individually, and save the obtained autoregression parameters to the database;
[0101] The corrosion data of each integrated pipeline is input into the corresponding pipeline model for corrosion prediction, and the adjacent pipeline model is used to correct the error of the obtained prediction results.
[0102] Those skilled in the art will appreciate that the modules or steps of the present disclosure described above can be implemented using a general-purpose computer device. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. Alternatively, they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. The present disclosure is not limited to any specific combination of hardware and software.
[0103] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.
[0104] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Those skilled in the art should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.
Claims
1. A dynamic prediction method for ship condenser corrosion based on a vector autoregressive model, characterized in that: The following steps are involved: Pre-process the corrosion data of each pipe of the ship's condenser, identify and process abnormal data, and then store it in the database; The autoencoder is used to extract features from the processed pipeline corrosion data to obtain a feature array Features, which is then integrated with the processed pipeline corrosion data and saved in a database. Obtain the newly added integrated pipeline corrosion data in the database, use the vector autoregression model to model each pipeline individually, and save the obtained autoregression parameters to the database; The corrosion data of each integrated pipeline is input into the corresponding pipeline model for corrosion prediction, and the adjacent pipeline model is used to correct the error of the obtained prediction results.
2. The method for dynamic prediction of ship condenser corrosion based on vector autoregressive model according to claim 1 is characterized in that: The pipeline corrosion data includes key information such as ship number, condenser number, measurement time, corrosion point location information, corrosion depth, measurement voltage, and measurement phase.
3. The method for dynamic prediction of ship condenser corrosion based on vector autoregressive model according to claim 1 is characterized in that: The autoencoder is constructed as follows: Confirm the autoencoder input dimension, intermediate dimension, and output dimension; Use the Tanh function as the activation function and the mean square error MSE as the loss function: All the pipeline corrosion data of the ship condenser corrosion is selected as input, the number of iteration steps is set, and the trained autoencoder is persisted. The autoencoder is called to extract features from the corrosion data of each pipe of the ship condenser to obtain the feature array Features.
4. The method for dynamic prediction of ship condenser corrosion based on vector autoregressive model according to claim 1 is characterized in that: The vector autoregressive model is: Y t =A0+A1Y t-1 +A2Y t-2 +…+A P Y t-p +e t where Y t is a column vector containing all variables, A0 is a constant vector, A1, A2, ..., A P is the coefficient matrix, p is the order of the model, which represents the dynamic relationship between variables, and ε t is the error term vector, which is assumed to be normally distributed with mean 0 and covariance matrix Σ.
5. The method for dynamic prediction of ship condenser corrosion based on vector autoregressive model according to claim 1 is characterized in that: When calling the vector autoregression model: Use unit root test to check the stationarity of time series. Non-stationary series need to be made stationary through difference operation. The order of the vector autoregression model was determined by the information criterion AIC; Estimation of model parameters based on input pipeline corrosion data; Residual tests were used to verify the validity of the model and the causal relationship between variables.
6. The method for dynamic prediction of ship condenser corrosion based on vector autoregressive model according to claim 1 is characterized in that: The way to model each pipeline individually is as follows: Asynchronously call the vector autoregression model training program and automatically retrieve the historical data of the corresponding pipeline. The newly added integrated pipeline corrosion data is used as input for model training. After training, the regression parameters are saved to the database, including model parameters, pipeline ID, model update time, and pipeline adjacency fields.
7. The method for dynamic prediction of ship condenser corrosion based on vector autoregressive model according to claim 1 is characterized in that: During prediction, each integrated pipeline corrosion data is input into the corresponding pipeline model to obtain the prediction result Y; at the same time, according to the adjacent relationship of the pipelines, each integrated pipeline corrosion data is input into the adjacent pipeline model to obtain multiple prediction correction data Z1, Z2, ..., Z n , and determine the weight of each predicted correction data according to the distance formula as the basis for correcting Y: Where: D(Z1) is the distance between pipe Y and pipe Z1.
8. A dynamic prediction system for ship condenser corrosion based on a vector autoregressive model, characterized by: include: The pre-processing module pre-processes the corrosion data of each pipe of the ship condenser, identifies abnormal data, processes it, and then stores it in the database; The feature extraction module uses the autoencoder to extract features from the processed pipeline corrosion data to obtain the feature array Features, integrates the feature array Features with the processed pipeline corrosion data and saves them to the database; The model building module obtains the newly added integrated pipeline corrosion data in the database, calls the vector autoregression model to model each pipeline separately, and saves the obtained autoregression parameters to the database; The prediction and correction module inputs the corrosion data of each integrated pipeline into the corresponding pipeline model to perform corrosion prediction, and uses the adjacent pipeline model to perform error correction on the obtained prediction results.
9. An electronic device comprising a memory, a processor, and a computer program stored and running on the memory, characterized in that: When the processor executes the program, the method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model as described in any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for dynamic prediction of ship condenser corrosion based on a vector autoregressive model as described in any one of claims 1 to 7 is implemented.