Artificial intelligence-based offshore booster station steel structure safety monitoring method and system

Abstract

According to the offshore booster station steel structure safety monitoring method and system based on artificial intelligence, the stress, vibration, corrosion, crack and environmental data of the steel structure are collected through the multi-modal sensor arranged at the offshore booster station steel structure part, and time alignment is performed on the multi-modal data through the time synchronization protocol; the multi-mode sensor comprises a stress sensor, a vibration sensor, a corrosion sensor, an ultrasonic sensor and an environment sensor. On the basis of finite element analysis and real-time sensor data, a dynamically-updated digital twinborn model is constructed, the limitation of traditional static monitoring is broken through, the stress distribution and damage state of the steel structure can be reflected in real time, a scientific basis is provided for simulating structural response under an extreme load, and the application prospect is wide. And the accuracy and predictability of steel structure safety evaluation in the environment are obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore engineering structure health monitoring, and in particular to an offshore booster station steel structure safety monitoring method and system based on artificial intelligence. Background Art

[0002] The offshore booster station is an important part of the offshore wind farm. Its main function is to collect the electric energy output by the wind turbine and boost it to transmit it to the land power grid. In the marine environment, the steel structure of the offshore booster station needs to withstand long-term wind loads, waves, salt spray corrosion and periodic fatigue loads. External factors can cause stress concentration, crack expansion and corrosion damage to the steel structure, posing a threat to the long-term safety and service life of the structure.

[0003] Traditional manual inspection: Professionals regularly inspect the appearance of steel structures and evaluate the cracks and corrosion conditions of steel structures by combining ultrasonic and radiographic detection. Relying on manual experience, the monitoring results are easily affected by human factors, the efficiency is low, and real-time monitoring cannot be achieved.

[0004] Sensor monitoring: Install stress sensors, vibration sensors or corrosion sensors at key locations of steel structures, and use sensor data to analyze the health status of steel structures. However, this method is usually based on a single type of sensor data, which cannot fully reflect the actual operating status of the steel structure, and sensor data is easily affected by noise, which affects monitoring accuracy.

[0005] Finite element analysis: Finite element analysis is used to simulate the stress state and crack propagation of steel structures. However, finite element analysis requires offline calculations and cannot dynamically update the model in combination with real-time monitoring data, and thus cannot adapt to the changing marine environment.

[0006] Lag of early warning and maintenance mechanisms: Existing monitoring systems mostly use fixed threshold early warning or rely on manual experience to formulate maintenance plans. They are unable to combine real-time monitoring data for scientific evaluation and optimization, which will lead to excessive maintenance or delayed repairs, and increase operation and maintenance costs. Summary of the invention

[0007] In view of the deficiencies in the prior art, the present invention provides an offshore booster station steel structure safety monitoring method and system based on artificial intelligence to solve the problems raised in the above background technology.

[0008] In order to achieve the above object, the present invention adopts the following technical solution: The artificial intelligence-based offshore booster station steel structure safety monitoring method includes the following steps: Step 1: Data collection: The stress, vibration, corrosion, crack and environmental data of the steel structure are collected by using multimodal sensors arranged on the steel structure of the offshore booster station, and the multimodal data are time-aligned by using a time synchronization protocol; Step 2: Data preprocessing: perform signal denoising and normalization on the collected multimodal data, and filter the data by setting abnormal thresholds to mark abnormal data; Step 3: Edge computing: Perform feature extraction and preliminary anomaly detection on the processed multimodal data on the edge computing device, and upload the screened abnormal data to the cloud analysis system in real time; Step 4: Artificial intelligence analysis: The uploaded abnormal data and related multimodal data are integrated and analyzed through the cloud analysis system, and an abnormality detection module and trend prediction module based on the deep learning model are constructed to predict the crack propagation and corrosion development trend of the steel structure; Step 5: Digital twin model construction: Based on the results of artificial intelligence analysis and the initial finite element model of the steel structure, a digital twin model is established and dynamically updated to simulate the health status of the steel structure under extreme load conditions; Step 6: Intelligent early warning and maintenance: Based on the digital twin model and artificial intelligence analysis results, calculate the health score, trigger graded early warnings, and generate optimized predictive maintenance plans.

[0009] A further improvement of the present invention is that the acquisition frequency of the multimodal sensor is dynamically adjusted, using a first sampling frequency during normal operation and increasing to a second sampling frequency under extreme weather conditions; The calculation formula of the first sampling frequency is as follows: , in, represents the sensor sampling frequency under normal working conditions, Represents the average value of the stress signal collected by the sensor within the time. Indicates the maximum stress value that the sensor can detect, , Represents the scale factor and offset of the sampling frequency; The calculation formula of the second sampling frequency is as follows: , in, represents the sensor sampling frequency under extreme weather conditions, Indicates the real-time monitored wind speed. Indicates the set wind speed threshold. , Represents the scale factor and offset for extreme weather sampling frequency.

[0010] A further improvement of the present invention is that in step 2, wavelet transform is used to perform signal denoising, and the normalization adopts the following formula: The wavelet transform: , in, is the signal value at time t, is the number of decomposition levels, is the time step of the signal, is the coefficient of the jth layer and wavelet decomposition position k, is the basis function of the jth layer and wavelet decomposition position k; The normalization formula: , in, is the normalized signal value, and its value range is (0, 1); is the original value of the current signal, is the minimum value of the signal, is the maximum value of the signal; Average based on historical data and standard deviation The data is filtered and the data that meets the conditions is marked as abnormal. The conditions are as follows: , in, is the original value of the current signal, is the historical average value of the signal, is the standard deviation of the signal, is the threshold for anomaly detection.

[0011] A further improvement of the present invention is that the edge computing device in step 3 extracts features from the marked abnormal data and transmits them to the cloud in real time, and stores normal data locally.

[0012] A further improvement of the present invention is that the artificial intelligence analysis in step 4 performs anomaly detection on the time series data through a long short-term memory network, and the constructed trend prediction module predicts crack growth and corrosion rate based on a gradient enhanced decision tree; The gradient boosting decision tree is an integrated learning algorithm based on the additive model, forward step-by-step optimization and decision tree, and its basic formula is as follows: The additive model says: , in, is the output of the ensemble model, is the total number of decision trees, Represents the prediction result of the mth tree; Objective function: , in, is the regularization term, L is the optimization objective function of the gradient boosting decision tree, is the true value of the i-th sample, is the input feature of the i-th sample, No. Wheel model pair sample The predicted value of is the mth decision tree for the sample The incremental predicted value of represents the loss function; Loss function expansion (gradient update part): , in, is the first-order gradient, is the second-order gradient, is the true value of the i-th sample, is the input feature of the i-th sample, No. Wheel model pair sample The predicted value of is the mth decision tree for the sample The incremental predicted value of is the loss function; Regularization term: , in, is the number of leaf nodes, is the leaf node weight, and is the regularization coefficient, is the regularization term; Node splitting gain: , Among them, I is the current node sample set, and represents the left and right node sample set after the split, represents the gain of node splitting, is the first-order gradient sum of the current node sample, is the second-order gradient sum of the current node sample, and is the regularization coefficient.

[0013] A further improvement of the present invention is that the multimodal data fusion in step 4 adopts the following deep neural network formula: , in, For multimodal data input, is the weight matrix, is the bias term, is the activation function.

[0014] A further improvement of the present invention is that the digital twin model established in step 5 is based on the finite element analysis of the steel structure, updates the geometric parameters and physical state of the model in real time, and dynamically simulates the structural response under extreme load conditions; the stress distribution of key parts is calculated based on the finite element method: , in, is the stress at the key part, is the force, is the cross-sectional area of ​​the stress-bearing part; The dynamic stress update formula, combined with the real-time data collected by the sensor, updates the stress state of the steel structure under dynamic load: , in, is the real-time dynamic stress, is the initial static stress, is the dynamic stress change; The crack growth prediction formula is based on the crack growth theory and predicts the crack growth rate of steel structures: , in, is the crack growth rate caused by each cyclic load; , is the material characteristic constant, is the stress intensity factor range, and the calculation formula is: , in, is the crack geometry factor, is the crack length, is the dynamic stress change, is the dynamic stress change.

[0015] A further improvement of the present invention is that the health score calculation formula in step 6 is: , in, Score your health. is the current degree of damage, is the critical damage value; Trigger low, medium or high level warnings based on the health score and generate corresponding maintenance plans.

[0016] The artificial intelligence-based offshore booster station steel structure safety monitoring system includes: The data acquisition module collects stress, vibration, corrosion, crack and environmental data of the steel structure through multi-modal sensors arranged on the steel structure of the offshore booster station, and performs time alignment on the multi-modal data through a time synchronization protocol; The data preprocessing module performs signal denoising and normalization on the collected multimodal data, and filters the data by setting abnormal thresholds and marking abnormal data; The edge computing module performs feature extraction and preliminary anomaly detection on the processed multimodal data on the edge computing device, and uploads the screened abnormal data to the cloud analysis system in real time; The artificial intelligence analysis module uses the cloud analysis system to integrate and analyze the uploaded abnormal data and related multimodal data, build an abnormality detection module and trend prediction module based on the deep learning model, and predict the crack expansion and corrosion development trend of steel structures; The digital twin model building module combines the results of artificial intelligence analysis and builds and dynamically updates the digital twin model based on the initial finite element model of the steel structure to simulate the health status of the steel structure under extreme load conditions; The intelligent early warning and maintenance module calculates health scores, triggers graded early warnings, and generates optimized predictive maintenance plans based on the digital twin model and artificial intelligence analysis results.

[0017] A further improvement of the present invention is that, in the data acquisition module, the acquisition frequency of the multimodal sensor is dynamically adjusted, using a first sampling frequency during normal operation and increasing to a second sampling frequency under extreme weather conditions; The calculation formula of the first sampling frequency is as follows: , in, represents the sensor sampling frequency under normal working conditions, Represents the average value of the stress signal collected by the sensor within the time. Indicates the maximum stress value that the sensor can detect, , Represents the scale factor and offset of the sampling frequency; The calculation formula of the second sampling frequency is as follows: , in, represents the sensor sampling frequency under extreme weather conditions, Indicates the wind speed monitored in real time. Indicates the set wind speed threshold. , Represents the scale factor and offset for extreme weather sampling frequency.

[0018] Compared with the prior art, the present invention has at least the following beneficial technical effects: Based on finite element analysis and real-time sensor data, the present invention constructs a dynamically updated digital twin model, breaking through the limitations of traditional static monitoring. It can reflect the stress distribution and damage status of steel structures in real time, and provide a scientific basis for simulating structural responses under extreme loads, significantly improving the accuracy and predictability of safety assessment of steel structures in different environments.

[0019] The present invention realizes intelligent fusion analysis of multimodal data through deep learning algorithms, can accurately detect anomalies and predict crack propagation and corrosion rates, and the artificial intelligence model can extract implicit rules in the data, showing significant advantages in multimodal data fusion and dynamic trend prediction, making steel structure safety monitoring move towards an intelligent stage.

[0020] The present invention designs a graded warning mechanism based on health scores, which comprehensively considers multi-dimensional information such as stress distribution, crack extension and corrosion degree. The quantitative analysis of health scores can accurately trigger low-level, medium-level or high-level warnings, significantly reducing the delay in identifying safety hazards. It has advantages in changeable marine environments, can buy response time for managers, and effectively reduce the risk of accidents.

[0021] The present invention combines artificial intelligence models to predict crack expansion and corrosion trends, and realizes maintenance optimization based on trend prediction. Predictive maintenance can plan maintenance time and scope in advance, avoid resource waste and delayed repair problems, improve the pertinence and accuracy of maintenance, significantly reduce overall maintenance costs, and extend the service life of steel structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 The present invention is a flow chart of the artificial intelligence-based offshore booster station steel structure safety monitoring method.

[0024] Figure 2 The present invention is a structural block diagram of the artificial intelligence-based offshore substation steel structure safety monitoring system. DETAILED DESCRIPTION

[0025] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0026] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0027] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0028] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0029] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0030] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0031] Example 1 Please refer to the attached Figure 1 The artificial intelligence-based offshore booster station steel structure safety monitoring method provided by the present invention comprises the following steps: Step 1, data collection: The stress, vibration, corrosion, crack and environmental data of the steel structure are collected through multimodal sensors arranged at the steel structure of the offshore booster station, and the multimodal data are time-aligned through the time synchronization protocol. The multimodal sensors include stress sensors, vibration sensors, corrosion sensors, ultrasonic sensors and environmental sensors, and the sensors are arranged at the welds, nodes and load-bearing components of the steel structure; Step 2: Data preprocessing: perform signal denoising and normalization on the multimodal data collected in step 1, and filter the data by setting an abnormal threshold to mark abnormal data; Step 3: Edge computing: Perform feature extraction and preliminary anomaly detection on the multimodal data processed in step 2 on the edge computing device, and upload the screened abnormal data to the cloud analysis system in real time; Step 4: Artificial intelligence analysis: The abnormal data uploaded in step 3 and related multimodal data are integrated and analyzed through the cloud analysis system to build an abnormality detection module and trend prediction module based on the deep learning model to predict the crack expansion and corrosion development trend of the steel structure; Step 5: Digital twin model construction: Based on the results of the artificial intelligence analysis in step 4 and the initial finite element model of the steel structure, a digital twin model is established and dynamically updated to simulate the health status of the steel structure under extreme load conditions; Step 6: Intelligent early warning and maintenance: Based on the digital twin model and artificial intelligence analysis results in step 5, calculate the health score, trigger graded early warnings, and generate an optimized predictive maintenance plan.

[0032] Step 1: By deploying multimodal sensors at key locations of the offshore booster station steel structure, stress, vibration, corrosion, crack and environmental data can be collected, covering multi-dimensional information on the health status of the steel structure. The application of the time synchronization protocol ensures the time consistency of data from different sensors, laying the foundation for subsequent data fusion analysis. Through the comprehensive collection of multimodal data, the defects of the traditional single data monitoring method can be effectively compensated, significantly improving the comprehensiveness and accuracy of monitoring; Step 2: De-noise and normalize the collected multimodal data, and filter out abnormal data by setting abnormal thresholds to mark potential problems. Step 2 uses denoising technology to eliminate the impact of environmental interference on sensor data, normalization to solve the differences between data of different dimensions, and abnormal screening to detect structural problems in advance. Data preprocessing improves data quality and provides high-quality input for subsequent feature extraction and artificial intelligence analysis; Step 3: The edge computing device performs feature extraction and preliminary anomaly detection on the preprocessed multimodal data, and uploads the screened anomaly data to the cloud analysis system in real time. This link extracts key information through feature extraction, and uses edge computing devices to achieve preliminary local anomaly detection, reducing the computing load and data transmission pressure of the cloud system. Real-time anomaly upload enables the system to respond quickly, improving the efficiency and sensitivity of steel structure safety monitoring; Step 4: In the cloud analysis system, the abnormal data and multimodal data uploaded by edge computing are integrated and analyzed through deep learning models to achieve anomaly detection and trend prediction. The artificial intelligence model can extract implicit rules in multimodal data, accurately detect abnormal conditions, and predict the crack expansion and corrosion development trends of steel structures. Compared with traditional monitoring methods, artificial intelligence analysis significantly improves the sensitivity of anomaly detection and the accuracy of trend prediction, providing strong technical support for the dynamic assessment of structural health status; Step 5. Combine the AI ​​analysis results with the initial finite element model to establish and dynamically update the digital twin model to simulate the health status of the steel structure under extreme load conditions. The digital twin model integrates real-time sensor data to intuitively display the stress distribution, crack propagation, and corrosion degree of the steel structure, while simulating the structural response under extreme wind loads or wave impact conditions. This model breaks through the limitations of traditional static monitoring and provides a more accurate and scientific basis for structural health assessment in the environment; Step 6: Based on the digital twin model and artificial intelligence analysis results, calculate the health score, trigger graded warnings, and generate an optimized predictive maintenance plan. Through the health score, the system can quantify the health status of the steel structure into intuitive indicators and trigger low-level, medium-level or high-level warnings, significantly reducing the delay in identifying safety hazards. Predictive maintenance can extend the service life of the steel structure and improve operation and maintenance efficiency by planning the maintenance time and scope in advance, optimizing resource allocation, avoiding unnecessary maintenance costs and delayed repair problems.

[0033] In summary, each step of the present invention is a key link in the overall safety monitoring process, from the comprehensive collection of multimodal data, to the accurate prediction of artificial intelligence analysis, to the dynamic simulation and intelligent early warning of the digital twin model. Each step is aimed at solving the defects of traditional monitoring methods and improving the intelligence level and economy of the whole life cycle management of steel structures.

[0034] The acquisition frequency of the multimodal sensor is dynamically adjusted, using the first sampling frequency during normal operation and increasing to the second sampling frequency under extreme weather conditions; The calculation formula of the first sampling frequency is as follows: , in, represents the sensor sampling frequency under normal working conditions, Represents the average value of the stress signal collected by the sensor within the time. Indicates the maximum stress value that the sensor can detect, , Represents the scale factor and offset of the sampling frequency; The calculation formula for the second sampling frequency is as follows: , in, represents the sensor sampling frequency under extreme weather conditions, Indicates the real-time monitored wind speed. Indicates the set wind speed threshold. , Represents the scale factor and offset for extreme weather sampling frequency.

[0035] Dynamically adjust the acquisition frequency of the multimodal sensor to optimize the utilization efficiency and response speed of monitoring resources according to changes in the operating environment. Under normal working conditions, the first sampling frequency is used, and the sampling frequency is dynamically adjusted by the ratio of the average value of the sensor stress signal to the maximum detection value to ensure the accuracy of monitoring and avoid the waste of resources caused by excessively high sampling frequency; In extreme weather conditions, the second sampling frequency is calculated based on the real-time wind speed and wind speed threshold to ensure that detailed data is captured in high-risk environments and potential problems are quickly identified. This mechanism effectively balances monitoring accuracy and system resource utilization, improving adaptability and monitoring efficiency under working conditions.

[0036] In step 2, wavelet transform is used to denoise the signal, and the normalization adopts the following formula: Wavelet Transform: , in, is the signal value at time t, is the number of decomposition levels, is the time step of the signal, is the coefficient of the jth layer and wavelet decomposition position k, is the basis function of the jth layer and wavelet decomposition position k; Normalization formula: , in, is the normalized signal value, and its value range is (0, 1); is the original value of the current signal, is the minimum value of the signal, is the maximum value of the signal; Average based on historical data and standard deviation The data is filtered and the data that meets the conditions is marked as abnormal. The conditions are as follows: , in, is the original value of the current signal, is the historical average value of the signal, is the standard deviation of the signal, is the threshold for anomaly detection; The edge computing device in step 3 extracts features from the abnormal data marked in step 2 and transmits them to the cloud in real time, and stores normal data locally.

[0037] Signal denoising by wavelet transform in step 2 can effectively remove environmental noise in the sensor acquisition signal while retaining key feature information, significantly improving data quality; Normalization processing makes data of different dimensions unified in the same range (0, 1), which is convenient for subsequent analysis and model fusion; The anomaly detection mechanism based on historical mean and standard deviation can quickly identify data points that deviate significantly from the normal range, providing an accurate basis for anomaly identification.

[0038] Combined with the edge computing in step 3, abnormal data is extracted and transmitted to the cloud analysis system in real time to achieve rapid response, while normal data is stored locally to reduce the computing and transmission burden on the cloud, further optimizing the system's resource utilization and operating efficiency, improving the real-time and accuracy of data processing, and enhancing the monitoring sensitivity to sudden anomalies, providing high-quality data support for subsequent analysis and decision-making.

[0039] The artificial intelligence analysis in step 4 detects anomalies in time series data through long short-term memory networks, and the constructed trend prediction module predicts crack growth and corrosion rates based on gradient enhanced decision trees; The gradient boosting decision tree is an integrated learning algorithm based on the additive model, forward step-by-step optimization and decision tree. Its basic formula is as follows: The additive model says: , in, is the output of the ensemble model, is the total number of decision trees, Represents the prediction result of the mth tree; Objective function: , in, is the regularization term, L is the optimization objective function of the gradient boosting decision tree, is the true value of the i-th sample, is the input feature of the i-th sample, No. Wheel model pair sample The predicted value of is the mth decision tree for the sample The incremental predicted value of represents the loss function; Loss function expansion (gradient update part): , in, is the first-order gradient, is the second-order gradient, is the true value of the i-th sample, is the input feature of the i-th sample, No. Wheel model pair sample The predicted value of is the mth decision tree for the sample The incremental predicted value of is the loss function; Regularization term: , in, is the number of leaf nodes, is the leaf node weight, and is the regularization coefficient, is the regularization term; Node splitting gain: , Among them, I is the current node sample set, and represents the left and right node sample set after the split, represents the gain of node splitting, is the first-order gradient sum of the current node sample, is the second-order gradient sum of the current node sample, and is the regularization coefficient; In step 4, multimodal data fusion uses the following deep neural network formula: , in, For multimodal data input, is the weight matrix, is the bias term, is the activation function.

[0040] In step 4, multimodal data is deeply processed through artificial intelligence analysis to realize intelligent anomaly detection and trend prediction. Long short-term memory network detects anomalies in time series data, which can capture the long-term dependency and dynamic changes of the operating status of steel structures, and effectively improve the accuracy and sensitivity of anomaly detection; The trend prediction module built based on the gradient boosting decision tree accurately predicts the crack growth and corrosion rate through additive model, forward step optimization and regularization term, showing excellent generalization performance and modeling ability for nonlinear data; The deep fusion analysis of multimodal data uses a neural network model to extract potential connections between different data sources and improve the comprehensiveness and efficiency of the overall analysis.

[0041] This step provides strong technical support for the accurate assessment of the health status of steel structures and prediction of future trends by integrating the advantages of deep learning and gradient enhancement decision trees, laying a solid foundation for subsequent early warning and maintenance decisions.

[0042] The digital twin model established in step 5 is based on the finite element analysis of the steel structure, updates the geometric parameters and physical state of the model in real time, and dynamically simulates the structural response under extreme load conditions; Based on the finite element method, calculate the stress distribution of key parts: , in, is the stress at the key part, is the force, is the cross-sectional area of ​​the stress-bearing part; The dynamic stress update formula combines the real-time data collected by the sensor to update the stress state of the steel structure under dynamic loads: , in, is the real-time dynamic stress, is the initial static stress, is the dynamic stress change; The crack growth prediction formula is based on the crack growth theory to predict the crack growth rate of steel structures: , in, is the crack growth rate caused by each cyclic load; , is the material characteristic constant, is the stress intensity factor range, and the calculation formula is: , in, is the crack geometry factor, is the crack length, is the dynamic stress change, is the dynamic stress change.

[0043] The digital twin model established in step 5 is based on finite element analysis and real-time sensor data to achieve dynamic simulation and prediction of the health status of the steel structure. The stress distribution of key parts is calculated by the finite element method, and the geometric parameters and physical state are updated in real time in combination with sensor data, which can dynamically reflect the stress changes of the steel structure under different load conditions; The crack propagation prediction formula combines crack propagation theory and stress intensity factor to accurately simulate the crack propagation rate and provide a forward-looking assessment of the damage development of the structure. Especially under extreme load conditions, the digital twin model can dynamically simulate the stress response and crack propagation trend of steel structures in real time, providing a scientific basis for managers and improving the accuracy of safety assessments.

[0044] This model breaks through the limitations of traditional static analysis and significantly enhances the ability of health assessment and risk prediction of steel structures in marine environments.

[0045] The health score calculation formula in step 6 is: , in, Score your health. is the current degree of damage, is the critical damage value; Trigger low, medium or high level warnings based on the health score and generate corresponding maintenance plans.

[0046] The health score calculation formula in step 6 provides a scientific basis for early warning and maintenance by quantifying the health status of the steel structure.

[0047] The health score H comprehensively considers the current damage level D and the critical damage value , which can convert the structural health status into intuitive numerical indicators and trigger low-level, medium-level or high-level warnings.

[0048] Low-level warnings indicate minor damage risks, intermediate-level warnings indicate potential hidden dangers that require special attention, and high-level warnings warn of structural failure risks, giving managers more reaction time.

[0049] Through health score-driven graded early warning and maintenance plans, targeted maintenance can be achieved, unnecessary waste of resources can be avoided, operation and maintenance costs can be reduced, and the safety and reliability of steel structure operation can be significantly improved.

[0050] Example 2 like Figure 2 As shown, the artificial intelligence-based offshore booster station steel structure safety monitoring system provided by the present invention includes: The data acquisition module collects stress, vibration, corrosion, crack and environmental data of the steel structure through multi-modal sensors arranged on the steel structure of the offshore booster station, and performs time alignment on the multi-modal data through a time synchronization protocol; The data preprocessing module performs signal denoising and normalization on the collected multimodal data, and filters the data by setting abnormal thresholds and marking abnormal data; The edge computing module performs feature extraction and preliminary anomaly detection on the processed multimodal data on the edge computing device, and uploads the screened abnormal data to the cloud analysis system in real time; The artificial intelligence analysis module uses the cloud analysis system to integrate and analyze the uploaded abnormal data and related multimodal data, build an abnormality detection module and trend prediction module based on the deep learning model, and predict the crack expansion and corrosion development trend of steel structures; The digital twin model building module combines the results of artificial intelligence analysis and builds and dynamically updates the digital twin model based on the initial finite element model of the steel structure to simulate the health status of the steel structure under extreme load conditions; The intelligent early warning and maintenance module calculates health scores, triggers graded early warnings, and generates optimized predictive maintenance plans based on the digital twin model and artificial intelligence analysis results.

[0051] In this embodiment, in the data acquisition module, the acquisition frequency of the multimodal sensor is dynamically adjusted, using a first sampling frequency during normal operation and increasing to a second sampling frequency under extreme weather conditions; The calculation formula of the first sampling frequency is as follows: , in, represents the sensor sampling frequency under normal working conditions, Represents the average value of the stress signal collected by the sensor within the time. Indicates the maximum stress value that the sensor can detect, , Represents the scale factor and offset of the sampling frequency; The calculation formula of the second sampling frequency is as follows: , in, represents the sensor sampling frequency under extreme weather conditions, Indicates the real-time monitored wind speed. Indicates the set wind speed threshold. , Represents the scale factor and offset for extreme weather sampling frequency.

[0052] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

[0053] In addition, it should be understood that although this specification is described in accordance with the implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation modes that can be understood by those skilled in the art. The above content is only to illustrate the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. The artificial intelligence-based offshore booster station steel structure safety monitoring method is characterized by: The steps include: Step 1: Data collection: The stress, vibration, corrosion, crack and environmental data of the steel structure are collected by using multimodal sensors arranged on the steel structure of the offshore booster station, and the multimodal data are time-aligned by using a time synchronization protocol; Step 2: Data preprocessing: perform signal denoising and normalization on the collected multimodal data, and filter the data by setting abnormal thresholds to mark abnormal data; Step 3: Edge computing: Perform feature extraction and preliminary anomaly detection on the processed multimodal data on the edge computing device, and upload the screened abnormal data to the cloud analysis system in real time; Step 4: Artificial intelligence analysis: The uploaded abnormal data and related multimodal data are integrated and analyzed through the cloud analysis system, and an abnormality detection module and trend prediction module based on the deep learning model are constructed to predict the crack propagation and corrosion development trend of the steel structure; Step 5: Digital twin model construction: Based on the results of artificial intelligence analysis and the initial finite element model of the steel structure, a digital twin model is established and dynamically updated to simulate the health status of the steel structure under extreme load conditions; Step 6: Intelligent early warning and maintenance: Based on the digital twin model and artificial intelligence analysis results, calculate the health score, trigger graded early warnings, and generate optimized predictive maintenance plans.

2. The artificial intelligence-based offshore booster station steel structure safety monitoring method according to claim 1 is characterized in that: The acquisition frequency of the multimodal sensor is dynamically adjusted, using a first sampling frequency during normal operation and increasing to a second sampling frequency under extreme weather conditions; The calculation formula of the first sampling frequency is as follows: ， in, represents the sensor sampling frequency under normal working conditions, Represents the average value of the stress signal collected by the sensor within the time. Indicates the maximum stress value that the sensor can detect, , Represents the scale factor and offset of the sampling frequency; The calculation formula of the second sampling frequency is as follows: ， in, represents the sensor sampling frequency under extreme weather conditions, Indicates the wind speed monitored in real time. Indicates the set wind speed threshold. , Represents the scale factor and offset for extreme weather sampling frequency.

3. The artificial intelligence-based offshore booster station steel structure safety monitoring method according to claim 1 is characterized in that: In step 2, wavelet transform is used to perform signal denoising, and the normalization adopts the following formula: The wavelet transform: , in, is the signal value at time t, is the number of decomposition levels, is the time step of the signal, is the coefficient of the jth layer and wavelet decomposition position k, is the basis function of the jth layer and wavelet decomposition position k; The normalization formula: , in, is the normalized signal value, and its value range is (0, 1); is the original value of the current signal, is the minimum value of the signal, is the maximum value of the signal; Average based on historical data and standard deviation The data is filtered and the data that meets the conditions is marked as abnormal. The conditions are as follows: ， in, is the original value of the current signal, is the historical average value of the signal, is the standard deviation of the signal, is the threshold for anomaly detection.

4. The artificial intelligence-based offshore booster station steel structure safety monitoring method according to claim 3 is characterized in that: The edge computing device in step 3 extracts features from the marked abnormal data and transmits them to the cloud in real time, and stores normal data locally.

5. The artificial intelligence-based offshore booster station steel structure safety monitoring method according to claim 1 is characterized in that: The artificial intelligence analysis in step 4 performs anomaly detection on time series data through a long short-term memory network, and the constructed trend prediction module predicts crack propagation and corrosion rate based on a gradient enhanced decision tree; The gradient boosting decision tree is an integrated learning algorithm based on the additive model, forward step-by-step optimization and decision tree, and its basic formula is as follows: The additive model says: , in, is the output of the ensemble model, is the total number of decision trees, Represents the prediction result of the mth tree; Objective function: , in, is the regularization term, L is the optimization objective function of the gradient boosting decision tree, is the true value of the i-th sample, is the input feature of the i-th sample, No. Wheel model pair sample The predicted value of is the mth decision tree for the sample The incremental predicted value of represents the loss function; Loss function expansion (gradient update part): ， in, is the first-order gradient, is the second-order gradient, is the true value of the i-th sample, is the input feature of the i-th sample, No. Wheel model pair sample The predicted value of is the mth decision tree for the sample The incremental predicted value of is the loss function; Regularization term: , in, is the number of leaf nodes, is the leaf node weight, and is the regularization coefficient, is the regularization term; Node splitting gain: , Among them, I is the current node sample set, and represents the left and right node sample set after splitting, represents the gain of node splitting, is the first-order gradient sum of the current node sample, is the second-order gradient sum of the current node sample, and is the regularization coefficient.

6. The artificial intelligence-based offshore booster station steel structure safety monitoring method according to claim 5 is characterized in that: The multimodal data fusion in step 4 adopts the following deep neural network formula: ， in, For multimodal data input, is the weight matrix, is the bias term, is the activation function.

7. The artificial intelligence-based offshore booster station steel structure safety monitoring method according to claim 1 is characterized in that: The digital twin model established in step 5 is based on the finite element analysis of the steel structure, updates the geometric parameters and physical state of the model in real time, and dynamically simulates the structural response under extreme load conditions; the stress distribution of key parts is calculated based on the finite element method: ， in, is the stress at the key part, is the force, is the cross-sectional area of ​​the stress-bearing part; The dynamic stress update formula, combined with the real-time data collected by the sensor, updates the stress state of the steel structure under dynamic load: ， in, is the real-time dynamic stress, is the initial static stress, is the dynamic stress change; The crack growth prediction formula is based on the crack growth theory and predicts the crack growth rate of steel structures: ， in, is the crack growth rate caused by each cyclic load; , is the material characteristic constant, is the stress intensity factor range, and the calculation formula is: ， in, is the crack geometry factor, is the crack length, is the dynamic stress change, is the dynamic stress change.

8. The artificial intelligence-based offshore booster station steel structure safety monitoring method according to claim 1 is characterized in that: The health score calculation formula in step 6 is: ， in, Score your health. is the current degree of damage, is the critical damage value; Trigger low, medium or high level warnings based on the health score and generate corresponding maintenance plans.

9. The offshore booster station steel structure safety monitoring system based on artificial intelligence is characterized by: include: The data acquisition module collects stress, vibration, corrosion, crack and environmental data of the steel structure through multi-modal sensors arranged on the steel structure of the offshore booster station, and performs time alignment on the multi-modal data through a time synchronization protocol; The data preprocessing module performs signal denoising and normalization on the collected multimodal data, and filters the data by setting abnormal thresholds and marking abnormal data; The edge computing module performs feature extraction and preliminary anomaly detection on the processed multimodal data on the edge computing device, and uploads the screened abnormal data to the cloud analysis system in real time; The artificial intelligence analysis module uses the cloud analysis system to integrate and analyze the uploaded abnormal data and related multimodal data, build an abnormality detection module and trend prediction module based on the deep learning model, and predict the crack expansion and corrosion development trend of steel structures; The digital twin model building module combines the results of artificial intelligence analysis and builds and dynamically updates the digital twin model based on the initial finite element model of the steel structure to simulate the health status of the steel structure under extreme load conditions; The intelligent early warning and maintenance module calculates health scores, triggers graded early warnings, and generates optimized predictive maintenance plans based on the digital twin model and artificial intelligence analysis results.

10. The artificial intelligence-based offshore booster station steel structure safety monitoring system according to claim 9 is characterized in that: In the data acquisition module, the acquisition frequency of the multimodal sensor is dynamically adjusted, using a first sampling frequency during normal operation and increasing to a second sampling frequency under extreme weather conditions; The calculation formula of the first sampling frequency is as follows: ， in, represents the sensor sampling frequency under normal working conditions, Represents the average value of the stress signal collected by the sensor within the time. Indicates the maximum stress value that the sensor can detect, , Represents the scale factor and offset of the sampling frequency; The calculation formula of the second sampling frequency is as follows: ， in, represents the sensor sampling frequency under extreme weather conditions, Indicates the wind speed monitored in real time. Indicates the set wind speed threshold. , Represents the scale factor and offset for extreme weather sampling frequency.

Images