Method, device and equipment for evaluating durability of welding node of offshore converter station and medium

Abstract

The invention discloses a method, a device, equipment and a medium for evaluating durability of a welding node of an offshore converter station, and the method comprises the steps: synchronously collecting multi-source environment dynamic data of the offshore converter station and vibration data of the welding node, and carrying out the correlation analysis, and obtaining an environment-vibration correlation data set; according to the environment-vibration associated data set and the local plastic deformation data, welding node environment coupling analysis is carried out, and a fatigue stress feature vector is obtained; according to the fatigue stress feature vector, predicting a crack expansion trend to obtain a crack expansion feature vector, and based on the crack expansion feature vector, calculating a comprehensive damage evaluation parameter so as to identify a stress concentration area of the welding node; and according to the crack expansion rate and the stress force data of the stress concentration area, generating a durability evaluation result of each welding node of the offshore converter station. According to the method, the accuracy of durability evaluation of the welding node of the offshore converter station can be improved by dynamically coupling the influence of the marine environment load and fatigue damage evolution on the welding node.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent offshore converter station, and particularly relates to a method and device for evaluating the durability of a welded joint of an offshore converter station, equipment and a medium. BACKGROUND

[0002] As a key facility of offshore wind power and other clean energy systems, the offshore converter station is long-term exposed to complex marine environment and subjected to the continuous action of wind, wave, tidal current and other dynamic loads. As a weak link in the structure, the welded joint is prone to crack and gradually expand due to stress concentration, fatigue damage and environmental impact, which seriously affects the integrity and service life of the overall structure. Therefore, the evaluation of the durability of the welded joint has important engineering significance and practical value for ensuring the safe operation of offshore energy facilities, preventing sudden structural failure and prolonging the service life of the facilities.

[0003] At present, the durability evaluation method for the welded joint of the offshore converter station mainly relies on static load analysis or simplified fatigue model, which is difficult to accurately reflect the dynamic coupling effect of random and multi-source external loads in the actual marine environment. The existing method has limitations in dealing with the correlation between vibration frequency characteristics, stress distribution and environmental factors, often ignoring the comprehensive influence of local plastic deformation, material degradation and crack path change on durability. In addition, the traditional method relies on empirical formula or single physical model for crack propagation trend prediction, lacks systematic analysis of multi-source data fusion and dynamic damage evolution, resulting in insufficient accuracy and poor adaptability of the evaluation results, which is difficult to meet the actual needs of structural health management in complex marine environment. SUMMARY

[0004] The present application provides a method and device for evaluating the durability of a welded joint of an offshore converter station, which can improve the accuracy of the durability evaluation of the welded joint of the offshore converter station by dynamically coupling the influence of marine environmental loads and fatigue damage evolution on the welded joint.

[0005] In a first aspect, the present application provides a method for evaluating the durability of a welded joint of an offshore converter station, comprising:

[0006] Synchronously collecting multi-source environmental dynamic data of the offshore converter station and vibration data of the welded joint to obtain initial environment-vibration data, and obtaining an environment-vibration correlation data set by performing correlation analysis on the initial environment-vibration data;

[0007] Obtaining local plastic deformation data of the welded joint, and performing welded joint environmental coupling analysis according to the environment-vibration correlation data set and the local plastic deformation data to obtain a fatigue stress feature vector;

[0008] According to the fatigue stress feature vector, a crack propagation trend is predicted to obtain a crack propagation feature vector, and according to the crack propagation feature vector, a comprehensive damage evaluation parameter is calculated;

[0009] According to the comprehensive damage evaluation parameter, a stress concentration area of the welded joint is identified, and crack propagation rate and stress quantization data of the stress concentration area are obtained to generate a durability evaluation result of each welded joint of the offshore converter station according to the crack propagation rate and stress quantization data.

[0010] The embodiment of the present application avoids evaluation deviation caused by disconnection of data by capturing real-time correlation of environmental load (such as wind, wave, and tidal current) and structural vibration response, obtaining key factors affecting durability of the welded joint from multiple dimensions, and then filtering irrelevant noise and extracting significantly relevant features by identifying internal relationship between environmental data and vibration data, thereby improving reliability and representativeness of the data set and laying a foundation for environmental coupling analysis; the physical authenticity of fatigue evaluation is enhanced by introducing a material deformation factor, thereby accurately reflecting material behavior of the welded joint under actual load, and then the coupling effect of environmental load and response of the welded joint is quantified to generate a fatigue stress feature vector, thereby capturing dynamic influence of environmental factors on fatigue stress and solving the problem of neglecting coupling effect of environmental load in the existing method; the generation and expansion direction of the crack are predicted by using the model, thereby identifying potential damage in advance to realize a warning function, and then multiple factors such as crack propagation and material degradation are integrated to provide a unified damage quantization index, thereby simplifying the evaluation process and providing a comprehensive damage state description; the stress concentration area is identified to focus on the key area where the crack starts, thereby improving the pertinence and efficiency of the evaluation, and then the crack propagation rate and stress quantization data of the stress concentration area are obtained to generate the durability evaluation result of the welded joint, thereby supporting long-term reliable operation of offshore facilities. Compared with the prior art, the present application can dynamically couple the influence of marine environmental load and fatigue damage evolution on the welded joint to improve the accuracy of durability evaluation of the welded joint of the offshore converter station.

[0011] Further, the environmental-vibration correlation analysis is performed on the initial environmental-vibration data to obtain an environmental-vibration correlation data set, specifically:

[0012] The vibration displacement peak data in the initial environmental-vibration data are extracted, and the vibration displacement peak data are subjected to frequency domain analysis by Fourier transform to obtain vibration frequency distribution;

[0013] The correlation coefficient between the vibration frequency distribution and each type of environmental data in the environment-vibration initial data is calculated by a Pearson correlation analysis method, and vibration frequency distribution data and environmental data with a correlation coefficient greater than a preset threshold are extracted to obtain an environment-vibration correlation data set; wherein the environment-vibration initial data includes offshore wind load data, offshore wave load data, and offshore tidal flow data.

[0014] The embodiment of the present application converts the time-domain vibration signal into the frequency domain by Fourier transform, identifies the main vibration frequency component, facilitates the analysis of the influence of periodic load, and thus reveals the frequency domain correlation between the vibration characteristics and the environmental load, improves the data analysis capability; by quantifying the linear relationship between the environmental data and the vibration frequency, the significantly correlated features are screened out, the irrelevant variables are removed, the key influencing factors of the environment-vibration correlation data set are focused on, and based on the preset threshold, the high correlation data is retained, the noise interference is reduced, and the accuracy and usability of the data set are improved.

[0015] Further, based on the environment-vibration correlation data set and the local plastic deformation data, the welding node environment coupling analysis is performed to obtain a fatigue stress feature vector, specifically:

[0016] The environment-vibration correlation data set and the local plastic deformation data are input into a preset linear regression model to obtain a corresponding environmental load coefficient set;

[0017] The correlation coefficient between the environmental load coefficient set and a pre-acquired vibration feature set is calculated by a Pearson correlation analysis method to generate a coupling coefficient set; wherein the vibration feature set includes vibration frequency interval distribution and stress amplitude data extracted from the environment-vibration initial data;

[0018] According to a preset coupling threshold, significant coefficients are extracted from the coupling coefficient set to obtain a significant coefficient set, and the change trend of the significant coefficient set is analyzed by a preset time series decomposition method to obtain a change feature vector;

[0019] The vibration feature set and the change feature vector are fused to obtain a fatigue stress feature vector.

[0020] The embodiment of the present application establishes the mathematical relationship between the environmental data and the vibration data through the linear regression model, quantifies the influence of the environment on the welding joint load, and provides a quantitative index of the environmental load; the correlation analysis is further used to screen the correlation between the environmental load coefficient and the vibration characteristics, a coupling coefficient set is generated to identify the key coupling factors, and the interaction effect of the environment and the vibration is highlighted; then, the long-term cumulative effect is captured by analyzing the change trend of the coupling coefficient over time, and the understanding of the dynamic response is enhanced; the vibration characteristics and the trend characteristics are integrated through feature fusion, a comprehensive fatigue stress feature vector is formed to represent the fatigue state of the welding joint under the environmental coupling, and the completeness of the feature representation is improved.

[0021] Further, according to the fatigue stress feature vector, a crack propagation trend is predicted to obtain a crack propagation feature vector, specifically:

[0022] According to the stress amplitude and the spatial position, the fatigue stress feature vector is divided into subsets through the pre-trained K-means clustering algorithm to obtain a stress concentration feature set;

[0023] Through Fourier transform, time series features of stress fluctuations are extracted from the stress concentration feature set to obtain a stress fluctuation sequence;

[0024] The stress fluctuation sequence is input into a pre-trained support vector regression model to predict the crack propagation rate through the support vector regression model to obtain a crack propagation rate set;

[0025] According to the crack propagation rate set, the change trend of the crack propagation over time is analyzed to obtain crack propagation trend description data, and the crack propagation trend description data and the stress concentration feature set are fused to generate a crack propagation feature vector.

[0026] The embodiment of the present application identifies the stress concentration area according to the stress amplitude and the spatial position grouping through K-means clustering, thereby focusing on the key area and improving the evaluation efficiency; the frequency domain features of stress fluctuations are extracted through Fourier transform to identify the periodic fluctuation mode and reveal the potential influence of stress fluctuations on crack driving; the crack propagation rate is predicted based on stress fluctuations through the support vector regression model to handle the nonlinear relationship and provide accurate crack rate quantification; the comprehensive crack propagation feature vector is formed by integrating the crack rate trend and the stress concentration feature, thereby comprehensively describing the crack behavior and enhancing the prediction reliability.

[0027] Further, according to the crack propagation feature vector, a comprehensive damage assessment parameter is calculated, specifically:

[0028] According to the crack propagation feature vector, a crack propagation path prediction is performed to obtain a crack propagation path feature set;

[0029] According to the crack propagation path feature set, a crack path deflection angle is calculated to generate a deflection angle distribution set, and through a preset timing analysis method, an angle change trend is extracted according to the deflection angle distribution set to obtain an angle change trend set;

[0030] According to the angle change trend set and a pre-acquired fatigue damage ratio, a local material degradation index is calculated to generate a material degradation distribution set;

[0031] According to the crack propagation rate set and the material degradation distribution set, a comprehensive damage evaluation parameter is calculated to obtain a comprehensive damage parameter set.

[0032] The embodiment of the present application predicts the crack path by considering the directionality of crack propagation to evaluate the geometric complexity of crack behavior, captures the change trend of crack path deflection to reflect the influence of environmental load on the crack path, quantifies the material degradation degree based on the angle trend and the fatigue damage ratio to introduce the material performance attenuation factor, and provides a unified damage index by integrating the crack rate and material degradation to realize multi-factor fusion damage evaluation.

[0033] Further, according to the crack propagation feature vector, crack propagation path prediction is performed to obtain a crack propagation path feature set, specifically:

[0034] According to the crack propagation feature vector, the stress intensity factor of the crack tip is calculated through linear elastic fracture mechanics theory, and the crack tip opening displacement is calculated through elastic-plastic fracture mechanics theory based on the crack propagation feature vector.

[0035] According to the stress intensity factor of the crack tip and the crack tip opening displacement, crack propagation path prediction is performed to obtain a crack propagation path feature set.

[0036] The embodiment of the present application quantifies the stress field intensity of the crack tip based on linear elastic theory to provide a driving force index for crack propagation, reflects the plastic deformation of the crack tip based on elastic-plastic theory to capture the influence of material plasticity on crack behavior, and combines the two to scientifically predict crack path features, making the prediction more consistent with the physical mechanism and improving reliability.

[0037] Further, according to the crack propagation rate and stress quantization data, durability evaluation results of each welded node of the offshore converter station are generated, specifically:

[0038] From the crack propagation rate set, crack propagation rate data corresponding to the stress concentration area are extracted to obtain a key crack propagation rate set;

[0039] From the stress concentration feature set, stress concentration feature data corresponding to the stress concentration area is extracted to obtain a key stress concentration feature set;

[0040] The key crack propagation rate set and the key stress concentration feature set are input into a preset machine learning model to generate a durability evaluation value of each welding node by the machine learning model.

[0041] The embodiment of the present application improves the pertinence and efficiency of the evaluation by focusing on the stress concentration area and screening the most relevant crack rate and stress features; and realizes automatic, objective and efficient durability evaluation by outputting quantitative evaluation values from the trained model.

[0042] In a second aspect, the embodiment of the present application provides a welding node durability evaluation device for a marine converter station, comprising an original data acquisition module, a stress feature acquisition module, a stress damage evaluation module and a durability evaluation module, wherein,

[0043] The original data acquisition module is configured to synchronously acquire multi-source environmental dynamic data and vibration data of the welding node of the marine converter station to obtain environmental-vibration initial data, and acquire an environmental-vibration correlation data set by performing correlation analysis on the environmental-vibration initial data;

[0044] The stress feature acquisition module is configured to acquire local plastic deformation data of the welding node, and perform welding node environmental coupling analysis according to the environmental-vibration correlation data set and the local plastic deformation data to obtain a fatigue stress feature vector;

[0045] The stress damage evaluation module is configured to predict a crack propagation trend according to the fatigue stress feature vector to obtain a crack propagation feature vector, and calculate a comprehensive damage evaluation parameter according to the crack propagation feature vector;

[0046] The durability evaluation module is configured to identify a stress concentration area of the welding node according to the comprehensive damage evaluation parameter, and acquire crack propagation rate and stress quantification data of the stress concentration area, so as to generate a durability evaluation result of each welding node of the marine converter station according to the crack propagation rate and the stress quantification data.

[0047] The embodiment of the present application captures the real-time correlation between environmental loads (such as wind, wave, and tidal current) and structural vibration responses through the original data acquisition module, obtains key factors affecting the durability of the welded joint from multiple dimensions, avoids evaluation deviation caused by disconnection of data, then filters irrelevant noise by identifying the internal relationship between environmental data and vibration data, extracts significant relevant features, thereby improving the reliability and representativeness of the data set, and laying a foundation for environmental coupling analysis; the stress feature acquisition module introduces the material deformation factor, enhances the physical authenticity of fatigue evaluation, thereby accurately reflecting the material behavior of the welded joint under actual load, then quantifies the coupling effect of environmental load and welded joint response, and generates a fatigue stress feature vector, thereby capturing the dynamic influence of environmental factors on fatigue stress and solving the problem of neglecting the coupling effect of environmental load in the existing method; the durability evaluation module uses the model to predict the generation and expansion direction of the crack, thereby identifying potential damage in advance and realizing the early warning function, then integrates multiple factors such as crack propagation and material degradation to provide a unified damage quantification index, thereby simplifying the evaluation process and providing a comprehensive damage state description; the durability evaluation module identifies the stress concentration area, focuses on the key area where the crack starts, improves the pertinence and efficiency of the evaluation, then obtains the crack propagation rate and stress quantification data of the stress concentration area to generate the durability evaluation result of the welded joint, and supports long-term reliable operation of offshore facilities.

[0048] In a third aspect, an embodiment of the present application provides a terminal device, comprising: a processor, a memory, a communication interface and a communication bus, the processor, the memory and the communication interface complete communication with each other through the communication bus;

[0049] The memory is used to store at least one executable instruction, and the executable instruction makes the processor execute the operation of the offshore converter station welded joint durability evaluation method according to any one of the above.

[0050] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, the computer readable storage medium comprises a stored computer program, wherein when the computer program runs, the computer readable storage medium controls the device or apparatus where the computer readable storage medium is located to execute the offshore converter station welded joint durability evaluation method according to any one of the above.

[0051] The above description is only a summary of the technical scheme of the embodiment of the present application, in order to more clearly understand the technical means of the embodiment of the present application, the content of the specification can be implemented, and in order to make the above and other purposes, characteristics and advantages of the embodiment of the present application more obvious and easy to understand, the specific embodiment of the present application is described as follows. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1A schematic diagram of a durability assessment method for welded joints in an offshore converter station provided by an embodiment of the present invention;

[0053] Figure 2 This is a structural diagram of a device for evaluating the durability of welded joints in an offshore converter station, provided as an embodiment of the present invention. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Example 1:

[0056] like Figure 1 As shown, an embodiment of the present invention provides a method for evaluating the durability of welded joints in an offshore converter station, comprising the following steps:

[0057] S101, synchronously collect multi-source environmental dynamic data and vibration data of welding nodes of the offshore converter station to obtain initial environmental-vibration data, and obtain environmental-vibration correlation dataset by performing correlation analysis on the initial environmental-vibration data;

[0058] In this embodiment, an environmental-vibration associated dataset is obtained by performing correlation analysis on the initial environmental-vibration data. Specifically, the peak vibration displacement data in the initial environmental-vibration data is extracted, and the vibration frequency distribution is obtained by performing frequency domain analysis on the peak vibration displacement data through Fourier transform. The correlation coefficient between the vibration frequency distribution and each type of environmental data in the initial environmental-vibration data is calculated using the Pearson correlation analysis method. The vibration frequency distribution data and environmental data with correlation coefficients greater than a preset threshold are extracted to obtain the environmental-vibration associated dataset. The initial environmental-vibration data includes marine wind load data, marine wave load data, and marine tidal current data.

[0059] In one optional embodiment, the multi-source environmental monitoring network of the offshore converter station can collect real-time data on wind load, wave load, and tidal current effects through sensors deployed at key locations within the station. Wind load monitoring devices can be installed at the top of the station to record wind speed and direction, wave load sensors can be positioned at the bottom of the station to measure wave height and period, and tidal current sensors can monitor water flow velocity and direction.

[0060] In an optional embodiment, assume that a monitoring point records wind speed of 15 m / s, wave height of 2.5 m, and tidal current speed of 0.8 m / s, which constitute a real-time external environment dynamic data set, providing a basis for subsequent analysis. Such multi-source monitoring can comprehensively capture the impact of the environment on the station body structure and improve data accuracy. A high-precision vibration sensor array synchronously collects vibration displacement peak value and dynamic load response data at the welded joint. For example, the sensor is arranged at the key welding point of the converter station, collects 1000 vibration data per second, records a displacement peak value of 0.02 mm, and a dynamic load response of 500 N. These data form an initial vibration fatigue and environment correlation data set, reflecting the real-time response of the structure under the action of the environment. Synchronous collection ensures data time consistency, laying a foundation for correlation analysis.

[0061] For the initial data set, a wavelet transform algorithm is used for denoising processing. Wavelet transform can effectively filter out random noise by decomposing the signal into different frequency components. For example, for vibration displacement data, abnormal high-frequency fluctuations caused by wave impact can be eliminated after denoising, and the main vibration trend is retained, obtaining the denoised vibration data set. This will significantly improve data reliability and reduce subsequent analysis errors.

[0062] For the denoised dynamic load response data, Z-score standardization method is used for calibration. For example, the mean of the dynamic load response data is 450 N, and the standard deviation is 50 N. By using the Z-score method, the data is standardized to a dimensionless value, which is convenient for cross-data set comparison.

[0063] After calibration, the vibration displacement peak value and the dynamic load response characteristics are extracted to construct a structured vibration stress data set. For example, a displacement peak value of 0.02 mm and a standardized load value of 1.2 constitute a feature pair, reflecting the stress state of the welding point. This step unifies the data scale and improves analysis consistency.

[0064] In an optional embodiment, if the vibration displacement peak value exceeds the preset threshold value of 0.015 mm, a Fourier transform algorithm is used for frequency domain analysis. For example, a displacement peak value of 0.02 mm triggers analysis, obtaining vibration frequency distribution data, showing that the main frequency is 2 Hz. This reveals that the vibration is mainly driven by specific environmental factors, such as periodic wave action. Frequency domain analysis helps to identify the source of vibration and improve the pertinence of structure safety evaluation.

[0065] For the correlation between vibration frequency distribution data and environmental data, Pearson correlation coefficient is used. For example, the calculation result shows that the correlation coefficient of wind load and vibration frequency is 0.85, the correlation coefficient of wave load is 0.65, and the correlation coefficient of tidal flow is 0.45, indicating that the wind load has the greatest impact on vibration. The environmental and vibration correlation data set is formed to determine the contribution of each environmental factor and provide a basis for optimization design. For example, based on the environmental and vibration correlation data set, a multi-dimensional feature matrix is constructed, including wind speed, wave height, tidal flow speed and vibration frequency dimensions.

[0066] Optionally, principal component analysis algorithm is used to extract main vibration stress features. For example, analysis shows that wind speed and vibration frequency contribute to 80% of the variance, and the final vibration stress feature data set is generated. Simplify the data dimension, highlight the key influencing factors, facilitate the prediction of structure fatigue life, and improve the safety and maintenance efficiency of offshore converter station.

[0067] S102, obtaining local plastic deformation data of the welding node, and performing welding node environmental coupling analysis according to the environmental-vibration correlation data set and the local plastic deformation data to obtain a fatigue stress feature vector;

[0068] In this embodiment, the environmental-vibration correlation data set and the local plastic deformation data are input into a preset linear regression model to obtain a corresponding environmental load coefficient set, and the correlation coefficient between the environmental load coefficient set and a pre-acquired vibration feature set is calculated by a Pearson correlation analysis method to generate a coupling coefficient set. The vibration feature set includes vibration frequency interval distribution and stress amplitude data extracted from the environmental-vibration initial data. According to a preset coupling threshold, significant coefficients are extracted from the coupling coefficient set to obtain a significant coefficient set, and the change trend of the significant coefficient set is analyzed by a preset time series decomposition method to obtain a change feature vector. The vibration feature set and the change feature vector are fused to obtain a fatigue stress feature vector.

[0069] In a specific embodiment, the vibration frequency interval of the welding node under various load conditions is obtained from the post-cleaning data set, and the time series signal is decomposed by a short-time Fourier transform method to obtain the frequency interval distribution. The frequency interval distribution and stress amplitude data are obtained by a vibration sensor array, and a data fusion method is used to generate a vibration feature set.

[0070] For the vibration feature set, a linear regression model is constructed in combination with the local plastic deformation data, wherein the linear regression model refers to the linear relationship Y=β0+β1X1+β2X2+...+βnXn between the independent variable X (vibration feature set and local plastic deformation data) and the dependent variable Y (environmental load coefficient).n X n +ε, β0 is the intercept, β0 to β n are regression coefficients, and ε is an error term, to obtain a trained linear regression model.

[0071] Through the trained linear regression model, the vibration feature set and the local plastic deformation data are input, the environmental load coefficient is calculated, and a preliminary environmental load coefficient set is obtained. For the preliminary environmental load coefficient set, if there is an abnormal value, the preset threshold T (T is determined based on the statistical distribution of historical data) is used for screening, if the coefficient value exceeds T, the abnormal value is removed, and a cleaned environmental load coefficient set is obtained.

[0072] The cleaned environmental load coefficient set is used, combined with the vibration feature set, and the Pearson correlation coefficient between the vibration feature and the environmental load coefficient is calculated by correlation analysis method. The Pearson correlation coefficient is r = ∑[(X i -μx)(Y i -μY)] / √[∑(Xi-μx) 2Σ (Y i -μY)2], where X i is the vibration feature value, Y i is the environmental load coefficient, μx and μY are the mean values of X and Y, respectively, to obtain a correlation coefficient matrix.

[0073] Through the correlation coefficient matrix, the corresponding relationship between the vibration feature and the environmental load coefficient with the correlation coefficient absolute value higher than the preset threshold R (R is determined based on the business accuracy requirement) is extracted, and a coupling coefficient set is generated. The coupling coefficient set is used, combined with the frequency interval distribution, to determine the coupling law of the vibration feature and the environmental load, and a final coupling coefficient set is obtained.

[0074] The significant coefficients are extracted from the coupling coefficient set, the time series decomposition method is used to analyze the change trend of the coefficients over time, and the change trend of the coefficients is obtained. According to the change trend of the coefficients, the vibration feature set is fused, and the vector mapping method is used to generate the fatigue stress feature vector of the welding node, and the fatigue stress feature vector is obtained.

[0075] Specifically, significant coefficients are extracted from the coupling coefficient set, and a time series decomposition method is used to analyze the trend of the coefficients over time to obtain the trend of the coefficients. Specifically, significant coefficients are extracted from the coupling coefficient set, and a time series decomposition method is used to analyze the trend of the coefficients. Time series decomposition divides the coefficients into trend, seasonal and residual components. For example, analysis may show that the coupling coefficient shows an upward trend when the wind speed is consistently high, reflecting the cumulative effect of long-term load. Preferably, the moving average method can be used to smooth the trend to enhance the stability of the analysis. According to the trend of the coefficients, the vibration feature set is fused, and a vector mapping method is used to generate a fatigue stress feature vector. Vector mapping integrates multi-dimensional features into a single vector, representing the fatigue state of the welded joint. For example, the generated feature vector may contain comprehensive information about frequency, amplitude and coupling coefficient, reflecting the fatigue characteristics of the joint under certain conditions. It should be noted that vector mapping facilitates subsequent classification or prediction, improving monitoring efficiency.

[0076] S103, according to the fatigue stress feature vector, predicting the crack propagation trend, obtaining the crack propagation feature vector, and calculating the comprehensive damage assessment parameter according to the crack propagation feature vector;

[0077] In this embodiment, according to the fatigue stress feature vector, the crack propagation trend is predicted to obtain the crack propagation feature vector. Specifically, the stress concentration feature set is obtained by dividing the fatigue stress feature vector into subsets according to the stress amplitude and spatial position through the pre-trained K-means clustering algorithm; the time series features of stress fluctuation are extracted from the stress concentration feature set through Fourier transform to obtain the stress fluctuation sequence; the stress fluctuation sequence is input into the pre-trained support vector regression model to predict the crack propagation rate through the support vector regression model to obtain the crack propagation rate set; the trend of crack propagation over time is analyzed according to the crack propagation rate set to obtain crack propagation trend description data, and the crack propagation trend description data is fused with the stress concentration feature set to generate a crack propagation feature vector.

[0078] In an optional embodiment, when obtaining stress concentration area identification and dynamic load response data from the fatigue stress feature description vector of the welded joint, high stress areas can be identified by analyzing the spatial characteristics of stress distribution. Stress concentration areas usually occur at geometric discontinuities of the welded joint, such as weld edge or joint corner. Stress distribution map can be generated by finite element analysis, combined with stress gradient threshold to mark the stress concentration area. For example, assuming that a certain welded joint has a stress peak of 400 MPa under a load of 100 kN, which is more than twice the average stress, the area is identified as a stress concentration area. Dynamic load response data is collected by vibration sensors to record the vibration amplitude and frequency of the joint under different loads to generate a time series data set. These data provide a basis for subsequent analysis.

[0079] In an optional embodiment, when separating the feature subsets corresponding to the stress concentration regions using the data segmentation method, a clustering algorithm such as K-means clustering can be used to group the feature vectors according to the stress amplitude and spatial location. For example, for a certain contact data set, the clustering results may separate the feature subsets of the stress concentration regions into a high amplitude group and a low amplitude group, with the high amplitude group corresponding to the high stress points on the weld edge. The separated feature subsets contain frequency, amplitude, and location information, facilitating further analysis of the stress concentration characteristics. For example, when extracting the time series features of stress fluctuations using the Fast Fourier Transform method, the vibration signal can be converted to frequency domain data to identify the main vibration frequencies. For example, the vibration signal of a certain contact shows that the main frequencies are concentrated at 10 Hz and 50 Hz, reflecting the vibration pattern under specific loads. The obtained stress fluctuation sequence contains both periodic and aperiodic fluctuation information. If the fluctuation amplitude exceeds a pre-set threshold T1, for example, T1 is set to 1.5 times the average amplitude, then high-frequency noise is removed by low-pass filtering to obtain the cleaned stress fluctuation sequence. This cleaning process preserves the key low-frequency fluctuation trend, improving the accuracy of subsequent modeling.

[0080] In an optional embodiment, a support vector regression model is used, with the stress fluctuation sequence and dynamic load response data as input to train the model, the model form is f(x) = w^Tφ(x) + b, where w is the weight vector, φ(x) is the feature mapping function, and b is the bias term, obtaining the trained support vector regression model. Further, when predicting the crack propagation rate using the support vector regression model, the cleaned stress fluctuation sequence and dynamic load response data can be used as input features. For example, the input features include vibration frequency 10 Hz, amplitude 0.5 mm, and load change rate 2 kN / s, and the model outputs the crack propagation rate in mm / cycle. The trained model can capture the nonlinear relationship between stress fluctuation and crack propagation. For example, when analyzing the trend of crack propagation rate using time series decomposition method, the rate sequence can be decomposed into trend, seasonal, and residual components. For example, the decomposition results may show that the crack propagation rate shows an upward trend during the high load period, reflecting the cumulative effect of long-term load on crack propagation. The extracted trend features, such as the linear growth slope of the rate, can be used to quantify the stability of crack propagation.

[0081] In an optional embodiment, when generating the comprehensive feature vector of crack propagation, the trend features and stress concentration features can be fused through vector mapping methods. For example, the fused feature vector may contain the trend slope of the crack propagation rate, the peak stress of the stress concentration region, and the main frequency component of the vibration frequency. This comprehensive feature vector can fully represent the fatigue damage state of the contact, facilitating subsequent crack propagation prediction and fatigue life assessment.

[0082] In the embodiment, the comprehensive damage evaluation parameter is calculated according to the crack propagation characteristic vector, specifically: crack propagation path prediction is performed according to the crack propagation characteristic vector to obtain a crack propagation path characteristic set; a crack path deflection angle is calculated according to the crack propagation path characteristic set to generate a deflection angle distribution set, and an angle change trend is extracted according to the deflection angle distribution set by a preset time sequence analysis method to obtain an angle change trend set; a local material degradation index is calculated according to the angle change trend set and a pre-acquired fatigue damage ratio to generate a material degradation distribution set; and the comprehensive damage evaluation parameter is calculated according to the crack propagation rate set and the material degradation distribution set to obtain a comprehensive damage parameter set.

[0083] In the embodiment, crack propagation path prediction is performed according to the crack propagation characteristic vector to obtain a crack propagation path characteristic set, specifically: a stress intensity factor of a crack tip is calculated according to the crack propagation characteristic vector by linear elastic fracture mechanics theory, and a crack tip opening displacement is calculated according to the crack propagation characteristic vector by elastic-plastic fracture mechanics theory; crack propagation path prediction is performed according to the stress intensity factor of the crack tip and the crack tip opening displacement to obtain the crack propagation path characteristic set.

[0084] In a specific embodiment, crack propagation path prediction is performed according to the crack propagation characteristic vector to obtain a crack propagation path characteristic set, specifically: a stress concentration region identifier and a dynamic load response characteristic are extracted from the crack propagation trend vector. An interpolation or data fusion method is used to map discrete sensor stress data to a geometric model of a node to generate a parameterized local stress field of a crack tip region.

[0085] According to the parameterized local stress field, a linear elastic fracture mechanics (LEFM, Linear Elastic Fracture Mechanics) method is used to calculate a stress intensity factor (SIF, Stress Intensity Factor) of the current crack tip, including an opening component, a sliding component and a tearing component. For the case of considering plastic deformation, an elastic-plastic fracture mechanics (EPFM, Elastic-Plastic Fracture Mechanics) method can be used to calculate a J integral or a crack tip opening displacement (CTOD, Crack Tip Opening Displacement) as a driving parameter to obtain a crack driving force parameter set.

[0086] Starting from the initial crack shape and location detected or assumed. Using maximum circumferential stress criterion, maximum energy release rate criterion or strain energy density factor method, etc. According to the crack driving force parameter set, the crack propagation direction angle of the next step is calculated. From the crack propagation trend vector, the current crack propagation rate set is obtained, combined with the number of load cycles (such as the predicted cycle in the future period of time), the next step of propagation increment is calculated.

[0087] According to the calculated propagation direction angle and propagation increment, the crack is extended in the geometric model to generate the updated crack front. Repeat the calculation process of the propagation direction angle and the propagation increment to make multi-step prediction until the crack reaches the preset critical length or completes the preset number of iterations, and finally obtain the predicted crack path coordinate set

[0088] In a specific embodiment, if the coefficient in the pre-acquired environmental influence classification set exceeds the preset threshold, the crack path deflection angle is calculated by a geometric analysis algorithm to generate a deflection angle distribution set. According to the deflection angle distribution set, a time series analysis method is used to extract the angle change trend to obtain an angle change trend set. Through the angle change trend set and the fatigue damage ratio, the local material degradation index is calculated to generate a material degradation distribution set. Through the propagation rate distribution set and the material degradation distribution set, the comprehensive damage assessment parameter is calculated to obtain a comprehensive damage parameter set.

[0089] Specifically, for example, in the field of fatigue analysis of welded joints in offshore converter stations, obtaining a comprehensive environmental influence coefficient set is an important step in analyzing crack propagation trends. The comprehensive environmental influence coefficient set usually includes the comprehensive influence of temperature, humidity, wind speed, wave load and other environmental factors on crack propagation.

[0090] Illustratively, environmental data such as temperature range -10℃ to 40℃, humidity 50% to 90%, wind speed 5m / s to 20m / s can be collected by sensors, combined with fatigue test data to generate an environmental influence coefficient set. It should be noted that the data collected should ensure real-time and representativeness to reflect the variability of the actual marine environment.

[0091] In an optional embodiment, the comprehensive environmental coefficient is classified using a clustering analysis method. The K-means clustering algorithm can be used to divide the coefficient into three categories: high impact, medium impact and low impact. For example, an environment with wind speed higher than 15m / s and humidity exceeding 80% is classified as high impact, generating an environmental influence classification set. It should be noted that the number of categories should be adjusted according to the working conditions to ensure the accuracy and practicality of the classification.

[0092] In an optional embodiment, if the coefficient in the classification set exceeds the preset threshold, such as the wind speed influence coefficient being greater than 0.8, the crack path deflection angle is calculated by a geometric analysis algorithm. For example, under the action of high wind speed, the crack path may be deflected due to uneven stress distribution, and the stress field of the weld area can be analyzed by finite element simulation to calculate the deflection angle of 5° to 15°, and a deflection angle distribution set is generated. It should be noted that the algorithm needs to consider the weld geometry, such as the weld root curvature, to improve the accuracy of angle calculation. For example, based on the deflection angle distribution set, a time series analysis method is used to extract the angle change trend, and an autoregressive model can be used to analyze the change of the angle over time. For example, the data shows that under the action of continuous high wind speed, the deflection angle increases by 2° every 1000 cycles, and a set of angle change trend is generated. It should be noted that the analysis needs to be combined with the load frequency to ensure that the trend reflects the real crack behavior.

[0093] In an optional embodiment, a local material degradation index is calculated based on the angle change trend set and the fatigue damage ratio. For example, assuming that the fatigue damage ratio is 0.4, combined with the angle change rate, the material degradation index is calculated to be 0.6, and a material degradation distribution set is generated. It should be noted that the calculation needs to consider the fatigue limit of the material to avoid misjudgment caused by too high index.

[0094] In an optional embodiment, a comprehensive damage assessment parameter is calculated based on the propagation rate distribution set and the material degradation distribution set. For example, combined with the rate of 0.01 mm / cycle and the degradation index of 0.6, the comprehensive damage parameter is calculated to be 0.7, and a comprehensive damage parameter set is generated. It should be noted that the parameter calculation needs to integrate multi-dimensional data to ensure the comprehensiveness and reliability of the evaluation.

[0095] S104, according to the comprehensive damage assessment parameter, identifying the stress concentration area of the welded joint, and obtaining the crack propagation rate and stress quantification data of the stress concentration area, to generate the durability evaluation result of each welded joint of the offshore converter station according to the crack propagation rate and stress quantification data.

[0096] In this embodiment, the durability evaluation result of each welded joint of the offshore converter station is generated according to the crack propagation rate and stress quantification data, specifically: from the crack propagation rate set, the crack propagation rate data corresponding to the stress concentration area is extracted to obtain a key crack propagation rate set; from the stress concentration feature set, the stress concentration feature data corresponding to the stress concentration area is extracted to obtain a key stress concentration feature set; the key crack propagation rate set and the key stress concentration feature set are input into a preset machine learning model, to generate the durability evaluation value of each welded joint by the machine learning model.

[0097] In a specific embodiment, in particular in the field of welded joint durability evaluation of marine converter stations, for the analysis of comprehensive damage evaluation parameters, the stress concentration area of the welded joint is automatically identified by finite element analysis software, the area with a stress concentration coefficient of 2.3 is extracted as a key monitoring point, and crack propagation rate data and stress concentration characteristic data of the key monitoring point are obtained; wherein the stress concentration characteristic data includes stress intensity factor and stress concentration coefficient;

[0098] Further, through time series analysis, combined with the dynamic change of the stress intensity factor at the crack tip, a regression model of crack propagation rate and frequency is constructed, and the correlation coefficient is 0.85, indicating that the vibration frequency has a significant effect on the crack propagation rate. Based on this, the durability quantitative algorithm is adopted, the crack propagation rate, stress intensity factor and stress concentration coefficient are input into the machine learning model, and the durability quantitative index is calculated as 0.72. This index reflects the long-term service capability of the welded joint. Through the data fusion algorithm, the durability index is compared with the historical operation data, and it is confirmed that the index below 0.8 needs to trigger a maintenance warning. The above process is processed through an automatic algorithm and model, forming a complete logical chain from stress analysis to durability evaluation, ensuring the rigor of data processing and the reliability of the results.

[0099] Preferably, if the durability quantitative index is lower than the preset threshold value, the micro damage distribution characteristics of the welded joint under the specific environmental load coupling coefficient are determined by comparing and analyzing the local plastic deformation degree and the local material degradation index, and the associated data of the stress amplitude distribution range and the crack propagation rate are fused, and a micro durability evaluation report of the welded joint of the offshore converter station is obtained. By collecting the initial data of the durability quantitative index, the collected index data is cleaned and normalized by using the data standardization processing method to obtain the standardized durability quantitative data set. According to the standardized durability quantitative data set, in combination with the local plastic deformation and the local material degradation data, the correlation analysis method is used to extract the comparative features of the two, and the local damage comparative feature set is determined. If the feature value in the local damage comparative feature set exceeds the preset threshold value, the stress amplitude distribution data and the crack propagation rate data are associated and mapped by using the data fusion technology to obtain the stress crack correlation feature set. According to the stress crack correlation feature set, the environmental load coupling data is fused, and the support vector machine algorithm is used to classify the influencing factors under the specific environmental load to determine the environmental load influence classification set. Through the environmental load influence classification set, in combination with the micro damage distribution data, the damage distribution characteristics of the welded joint are analyzed layer by layer to obtain the micro damage distribution characteristic set of the welded joint. According to the micro damage distribution characteristic set of the welded joint, the specific environmental load data of the offshore converter station is matched and analyzed to judge the micro durability influence degree, and the micro durability evaluation data set is obtained. If the key indicators in the micro durability evaluation data set are lower than the preset threshold value, the micro damage distribution characteristics of the welded joint of the offshore converter station are presented in multiple dimensions by using the data visualization tool, and the final evaluation result data is determined.

[0100] Exemplarily, in the field of offshore converter station welded joint micro-durability evaluation, when the durability quantitative index is lower than the preset threshold value 0.75, the system automatically starts the multi-dimensional analysis process. Firstly, the local plastic deformation degree of the welded joint is quantitatively calculated through the finite element simulation software, and the area with a local strain value of 0.002 is extracted as the key analysis object. Combined with the material mechanics algorithm, the influence of plastic deformation on material performance is evaluated, and the deformation influence factor is 1.5. Then, the system calls the material degradation evaluation model, calculates the local material degradation index based on historical environmental data and service time, and compares and analyzes the plastic deformation data. It is found that the degradation index is positively correlated with the deformation degree, and the correlation coefficient is 0.78. Next, the system extracts the stress amplitude distribution range data, analyzes the distribution characteristics of the stress amplitude between 5 to 15 MPa, and adopts the statistical regression algorithm to construct the correlation model of stress amplitude and crack propagation rate, and obtains the crack propagation rate of 0.008 mm / day when the stress amplitude is 10 MPa. Further, the system sets the environmental load coupling coefficient to 1.2, and combines the micro-damage distribution algorithm to calculate the micro-damage distribution characteristics of the welded joint in the salt spray corrosion and temperature alternating environment, and obtains the damage concentration area ratio of 12.5%. Finally, the system integrates the above analysis results through the data integration module, automatically generates the micro-durability evaluation report, and points out in the report that the damage distribution characteristics are highly correlated with the environmental load coupling coefficient, which provides data support for subsequent optimization design. The whole process is realized through automatic algorithm, which ensures the logical rigor and data consistency.

[0101] The embodiment of the present application can obtain key factors affecting the durability of the welded joint from multiple dimensions by capturing the real-time correlation between environmental loads (such as wind, wave, and tidal current) and structural vibration responses, avoiding evaluation deviation caused by disconnection of data, then identifying the internal relationship between environmental data and vibration data, filtering irrelevant noise, and extracting significant correlation features, thereby improving the reliability and representativeness of the data set and laying a foundation for environmental coupling analysis; the physical authenticity of fatigue evaluation is enhanced by introducing the material deformation factor, thereby accurately reflecting the material behavior of the welded joint under actual loads, then quantifying the coupling effect of environmental loads and welded joint responses to generate a fatigue stress feature vector, thereby capturing the dynamic influence of environmental factors on fatigue stress and solving the problem of neglecting the coupling effect of environmental loads in the existing method; the generation and expansion direction of the crack are predicted by using the model, thereby identifying potential damage in advance and realizing the early warning function, then integrating multiple factors such as crack propagation and material degradation to provide a unified damage quantification index, thereby simplifying the evaluation process and providing a comprehensive damage state description; the stress concentration area is identified to focus on the key area of crack initiation, thereby improving the pertinence and efficiency of the evaluation, then obtaining the crack propagation rate and stress quantification data of the stress concentration area to generate the durability evaluation result of the welded joint, and supporting long-term reliable operation of offshore facilities. Compared with the prior art, the present application can dynamically couple the influence of marine environmental loads and fatigue damage evolution on the welded joint to improve the accuracy of durability evaluation of the welded joint of the offshore converter station.

[0102] Embodiment two:

[0103] As shown in Figure 2 The present embodiment provides a durability evaluation device for a welded joint of an offshore converter station, comprising an original data acquisition module 201, a stress feature acquisition module 202, a stress damage evaluation module 203, and a durability evaluation module 204, wherein,

[0104] The original data acquisition module 201 is used for synchronously collecting multi-source environmental dynamic data and vibration data of the welded joint of the offshore converter station to obtain environmental-vibration initial data, and acquiring an environmental-vibration correlation data set by performing correlation analysis on the environmental-vibration initial data;

[0105] In the embodiment, the original data acquisition module 201 obtains the environment-vibration correlation data set by performing correlation analysis on the environment-vibration initial data, specifically: the original data acquisition module 201 extracts the vibration displacement peak value data in the environment-vibration initial data, and performs frequency domain analysis on the vibration displacement peak value data by Fourier transform to obtain the vibration frequency distribution; the correlation coefficients of the vibration frequency distribution and each type of environment data in the environment-vibration initial data are calculated respectively by the Pearson correlation analysis method, and the vibration frequency distribution data and the environment data with a correlation coefficient greater than a preset threshold are extracted to obtain the environment-vibration correlation data set; wherein the environment-vibration initial data includes offshore wind load data, offshore wave load data and offshore tidal flow data.

[0106] The stress feature acquisition module 202 is configured to obtain the local plastic deformation data of the welding joint, and perform welding joint environment coupling analysis according to the environment-vibration correlation data set and the local plastic deformation data to obtain a fatigue stress feature vector.

[0107] In the embodiment, the stress feature acquisition module 202 performs welding joint environment coupling analysis according to the environment-vibration correlation data set and the local plastic deformation data to obtain a fatigue stress feature vector, specifically: the stress feature acquisition module 202 inputs the environment-vibration correlation data set and the local plastic deformation data into a preset linear regression model to obtain a corresponding environment load coefficient set; the correlation coefficients between the environment load coefficient set and a pre-acquired vibration feature set are calculated by the Pearson correlation analysis method to generate a coupling coefficient set; wherein the vibration feature set includes vibration frequency interval distribution and stress amplitude data extracted from the environment-vibration initial data; according to a preset coupling threshold, significant coefficients are extracted from the coupling coefficient set to obtain a significant coefficient set, and the change trend of the significant coefficient set is analyzed by a preset time series decomposition method to obtain a change feature vector; the vibration feature set and the change feature vector are fused to obtain a fatigue stress feature vector.

[0108] The stress damage assessment module 203 is configured to predict a crack propagation trend according to the fatigue stress feature vector to obtain a crack propagation feature vector, and calculate a comprehensive damage assessment parameter according to the crack propagation feature vector.

[0109] In this embodiment, the stress damage assessment module 203 predicts the crack propagation trend according to the fatigue stress feature vector, and obtains a crack propagation feature vector. Specifically, the stress damage assessment module 203 performs subset division on the fatigue stress feature vector according to the stress amplitude and spatial position through a pre-trained K-means clustering algorithm, and obtains a stress concentration feature set; extracts time sequence features of stress fluctuations from the stress concentration feature set through Fourier transform, and obtains a stress fluctuation sequence; inputs the stress fluctuation sequence into a pre-trained support vector regression model, so as to perform crack propagation rate prediction through the support vector regression model, and obtain a crack propagation rate set; analyzes the change trend of crack propagation with time according to the crack propagation rate set, and obtains crack propagation trend description data; and performs feature fusion on the crack propagation trend description data and the stress concentration feature set, and generates a crack propagation feature vector.

[0110] In this embodiment, the stress damage assessment module 203 calculates a comprehensive damage assessment parameter according to the crack propagation feature vector. Specifically, the stress damage assessment module 203 performs crack propagation path prediction according to the crack propagation feature vector, and obtains a crack propagation path feature set; calculates a crack path deflection angle according to the crack propagation path feature set, so as to generate a deflection angle distribution set, and extracts an angle change trend according to the deflection angle distribution set through a pre-set time sequence analysis method, and obtains an angle change trend set; calculates a local material degradation index according to the angle change trend set and a pre-acquired fatigue damage ratio, and generates a material degradation distribution set; calculates a comprehensive damage assessment parameter according to the crack propagation rate set and the material degradation distribution set, and obtains a comprehensive damage parameter set.

[0111] The durability assessment module 204 is configured to identify a stress concentration area of a welded joint according to the comprehensive damage assessment parameter, and acquire crack propagation rate and stress intensity data of the stress concentration area, so as to generate a durability assessment result of each welded joint of the offshore converter station according to the crack propagation rate and stress intensity data.

[0112] In this embodiment, the durability assessment module 204 generates a durability assessment result of each welded joint of the offshore converter station according to the crack propagation rate and stress intensity data. Specifically, the durability assessment module 204 extracts crack propagation rate data corresponding to the stress concentration area from the crack propagation rate set, and obtains a key crack propagation rate set; extracts stress concentration feature data corresponding to the stress concentration area from the stress concentration feature set, and obtains a key stress concentration feature set; inputs the key crack propagation rate set and the key stress concentration feature set into a pre-set machine learning model, so as to generate a durability assessment value of each welded joint through the machine learning model.

[0113] The working principle and step flow of the embodiment can be seen in the relevant description of embodiment one, but are not limited thereto.

[0114] The embodiment of the present application captures the real-time correlation between environmental loads (such as wind, wave, and tidal current) and structural vibration response through the original data acquisition module 201, obtains key factors affecting the durability of the welded joint from multiple dimensions, avoids evaluation deviation caused by disconnection of data, then filters irrelevant noise by identifying the internal relationship between environmental data and vibration data, extracts significant relevant features, thereby improving the reliability and representativeness of the data set, and laying a foundation for environmental coupling analysis; through the stress feature acquisition module 202, the material deformation factor is introduced to enhance the physical authenticity of fatigue evaluation, thereby accurately reflecting the material behavior of the welded joint under actual load, then quantifying the coupling effect of environmental load and welded joint response to generate a fatigue stress feature vector, thereby capturing the dynamic influence of environmental factors on fatigue stress, solving the problem of neglecting the coupling effect of environmental load in the existing method; through the durability evaluation module 203, the model is used to predict the generation and expansion direction of the crack, thereby identifying potential damage in advance to realize the early warning function, then integrating multiple factors such as crack propagation and material degradation to provide a unified damage quantification index, thereby simplifying the evaluation process and providing a comprehensive damage state description; through the durability evaluation module 204, the stress concentration area is identified, and the key area where the crack starts is focused to improve the pertinence and efficiency of the evaluation, then the crack propagation rate and stress quantification data of the stress concentration area are obtained to generate the durability evaluation result of the welded joint, supporting long-term reliable operation of offshore facilities.

[0115] Embodiment three

[0116] The embodiment provides a terminal device, which comprises a processor, a memory, a communication interface and a communication bus, and the processor, the memory and the communication interface complete mutual communication through the communication bus.

[0117] The memory is used for storing at least one executable instruction, and the executable instruction makes the processor execute the operation of the offshore converter station welded joint durability evaluation method according to any one of the above.

[0118] Embodiment four

[0119] The embodiment of the present application provides a computer readable storage medium, which comprises a stored computer program, wherein when the computer program runs, the computer readable storage medium controls the device or apparatus where the computer readable storage medium is located to execute the offshore converter station welded joint durability evaluation method according to any one of the above.

[0120] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing relevant hardware, and the program can be stored in a computer-readable storage medium. When the program is executed, the program can include the processes of the above-mentioned embodiment methods. The storage medium can be a magnetic disc, an optical disc, a read-only memory (ROM), a random access memory (RAM), or the like.

[0121] The above-mentioned specific embodiments further illustrate the purposes, technical solutions, and beneficial effects of the present application. It should be understood that the above-mentioned embodiments are only specific embodiments of the present application and are not used to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, and the like made by those skilled in the art within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for evaluating the durability of welded joints in offshore converter stations, characterized in that, include: Simultaneously collect multi-source environmental dynamic data and vibration data of welding nodes of the offshore converter station to obtain initial environmental-vibration data, and obtain environmental-vibration correlation dataset by performing correlation analysis on the initial environmental-vibration data; The local plastic deformation data of the welded node is obtained, and the environmental coupling analysis of the welded node is performed based on the environmental-vibration correlation dataset and the local plastic deformation data to obtain the fatigue stress feature vector. Based on the fatigue stress characteristic vector, the crack propagation trend is predicted to obtain the crack propagation characteristic vector, and the comprehensive damage assessment parameters are calculated based on the crack propagation characteristic vector. Based on the comprehensive damage assessment parameters, stress concentration areas of welded nodes are identified, and crack propagation rate and stress quantification data of the stress concentration areas are obtained. Based on the crack propagation rate and stress quantification data, durability assessment results of each welded node of the offshore converter station are generated.

2. The durability assessment method for welded joints in offshore converter stations as described in claim 1, characterized in that, The process of obtaining an environment-vibration correlation dataset by performing correlation analysis on the initial environment-vibration data specifically involves: The peak vibration displacement data is extracted from the initial environmental vibration data, and the vibration frequency distribution is obtained by performing frequency domain analysis on the peak vibration displacement data through Fourier transform. Using the Pearson correlation analysis method, the correlation coefficients between the vibration frequency distribution and each type of environmental data in the initial environmental-vibration data are calculated respectively. Vibration frequency distribution data and environmental data with correlation coefficients greater than a preset threshold are extracted to obtain an environmental-vibration associated dataset. The initial environmental-vibration data includes marine wind load data, marine wave load data, and marine tidal current data.

3. The durability assessment method for welded joints in offshore converter stations as described in claim 1, characterized in that, Based on the aforementioned environment-vibration correlation dataset and local plastic deformation data, an environmental coupling analysis of the welded joint is performed to obtain the fatigue stress feature vector, specifically: The environmental-vibration correlation dataset and local plastic deformation data are input into a preset linear regression model to obtain the corresponding set of environmental load coefficients; The correlation coefficient between the environmental load coefficient set and the pre-acquired vibration feature set is calculated using the Pearson correlation analysis method to generate a coupling coefficient set; wherein, the vibration feature set includes the vibration frequency range distribution and stress amplitude data extracted from the environmental-vibration initial data; Based on a preset coupling threshold, significant coefficients are extracted from the coupling coefficient set to obtain a significant coefficient set. The changing trend of the significant coefficient set is analyzed using a preset time series decomposition method to obtain a change feature vector. The vibration feature set and the change feature vector are fused to obtain the fatigue stress feature vector.

4. The durability assessment method for welded joints in offshore converter stations as described in claim 1, characterized in that, Based on the fatigue stress characteristic vector, the crack propagation trend is predicted, and the crack propagation characteristic vector is obtained, specifically: By using a pre-trained K-means clustering algorithm, the fatigue stress feature vector is partitioned into subsets based on stress amplitude and spatial location to obtain a stress concentration feature set. By using Fourier transform, the time series features of stress fluctuations are extracted from the stress concentration feature set to obtain the stress fluctuation sequence; The stress fluctuation sequence is input into a pre-trained support vector regression model to predict the crack propagation rate and obtain a crack propagation rate set. Based on the crack propagation rate set, the crack propagation trend over time is analyzed to obtain crack propagation trend description data. The crack propagation trend description data is then fused with the stress concentration feature set to generate a crack propagation feature vector.

5. The durability assessment method for welded joints in offshore converter stations as described in claim 4, characterized in that, Based on the crack propagation feature vector, the comprehensive damage assessment parameters are calculated as follows: Based on the crack propagation feature vector, crack propagation path prediction is performed to obtain a crack propagation path feature set. Based on the crack propagation path feature set, the crack path deflection angle is calculated to generate a deflection angle distribution set. Then, using a preset time series analysis method, the angle change trend is extracted from the deflection angle distribution set to obtain an angle change trend set. Based on the set of angle change trends and the pre-acquired fatigue damage ratio, calculate the local material deterioration index and generate a material deterioration distribution set; Based on the crack propagation rate set and the material degradation distribution set, comprehensive damage assessment parameters are calculated to obtain a comprehensive damage parameter set.

6. The durability assessment method for welded joints in an offshore converter station as described in claim 5, characterized in that, Based on the crack propagation feature vector, crack propagation path prediction is performed to obtain a crack propagation path feature set, specifically: Using linear elastic fracture mechanics theory and based on the crack propagation characteristic vector, the stress intensity factor at the crack tip is calculated, and using elastoplastic fracture mechanics theory and based on the crack propagation characteristic vector, the crack tip opening displacement is calculated. Based on the stress intensity factor at the crack tip and the crack tip opening displacement, the crack propagation path is predicted, and a crack propagation path feature set is obtained.

7. The durability assessment method for welded joints in an offshore converter station as described in claim 5, characterized in that, Based on the crack propagation rate and stress quantification data, durability assessment results for each welded node of the offshore converter station are generated, specifically as follows: From the crack propagation rate set, extract the crack propagation rate data corresponding to the stress concentration region to obtain the key crack propagation rate set; From the stress concentration feature set, extract the stress concentration feature data corresponding to the stress concentration region to obtain the key stress concentration feature set; The key crack propagation rate set and key stress concentration feature set are input into a preset machine learning model to generate durability evaluation values ​​for each welded node through the machine learning model.

8. A durability assessment device for welded joints in offshore converter stations, characterized in that, It includes a raw data acquisition module, a stress characteristic acquisition module, a stress damage assessment module, and a durability assessment module. The raw data acquisition module is used to simultaneously collect multi-source environmental dynamic data and vibration data of welding nodes of the offshore converter station to obtain initial environmental-vibration data, and to obtain environmental-vibration associated dataset by performing correlation analysis on the initial environmental-vibration data. The stress feature acquisition module is used to acquire the local plastic deformation data of the welded node, and perform environmental coupling analysis of the welded node based on the environmental-vibration correlation dataset and the local plastic deformation data to obtain the fatigue stress feature vector. The stress damage assessment module is used to predict crack propagation trend based on the fatigue stress feature vector, obtain crack propagation feature vector, and calculate comprehensive damage assessment parameters based on the crack propagation feature vector. The durability assessment module is used to identify stress concentration areas of welded nodes based on the comprehensive damage assessment parameters, and to obtain crack propagation rate and stress quantification data of the stress concentration areas, so as to generate durability assessment results for each welded node of the offshore converter station based on the crack propagation rate and stress quantification data.

9. A terminal device, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the operation of the durability assessment method for welded nodes of offshore converter stations as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device or apparatus containing the computer-readable storage medium to perform the durability assessment method for welded joints of offshore converter stations as described in any one of claims 1 to 7.

Images