Sensor-based salt-fog corrosion early warning method and system for offshore platform generator set

Abstract

The application belongs to the technical field of corrosion prevention and control and intelligent monitoring of offshore platform equipment, and discloses a salt mist corrosion early warning method and system for offshore platform generator sets based on sensors, which solves the core problems of passive treatment as the main method of salt mist corrosion protection, insufficient dynamic sensing capability, and low precision of erosion trend prediction in the prior art. The method collects environmental data such as salt mist concentration, temperature and humidity through a sensor network, constructs a salt mist spatial distribution map corresponding to the generator set component area, determines the dynamic distribution law of the salt mist, calculates the salt mist erosion risk and determines the corrosion threat level, fits the erosion intensity correlation mode to determine the corrosion sensitivity, and finally predicts the erosion intensity change trend and generates an early warning signal and a protection scheme. Through the application, the pre-active early warning and precise control of salt mist corrosion are realized, and the safety and long-term stability of the operation of the offshore platform generator set are greatly improved.

Description

Technical Field

[0001] This application belongs to the field of corrosion prevention and intelligent monitoring technology for marine platform equipment, specifically a sensor-based method and system for early warning of salt spray corrosion of offshore platform generator sets. Background Technology

[0002] Offshore oil and gas platforms are core facilities for marine energy development. As the core of the platform's power supply system, the safe and stable operation of generator sets directly determines the platform's production continuity and operational safety. In the extreme marine environment of high salinity, high humidity, and strong winds, salt spray corrosion is the primary cause of corrosion of generator set metal components and electrical system failures, seriously threatening the platform's power supply reliability and representing a critical risk that urgently needs to be addressed in the field of marine energy development.

[0003] Currently, the protection against salt spray corrosion of offshore platform generator sets mostly adopts passive measures such as coating protection and regular maintenance. These measures lack the ability to perceive the dynamic changes in the salt spray environment in real time and cannot capture the spatial distribution differences and temporal evolution characteristics of salt spray around the generator set. This results in a disconnect between protective measures and actual corrosion risks, leading to high operation and maintenance costs and difficulty in avoiding sudden corrosion failures.

[0004] Existing technologies cannot grasp the dynamic distribution of salt spray in various parts of the generator set in real time, nor can they accurately quantify the salt spray erosion intensity and future trend in different areas. They cannot achieve differentiated and proactive corrosion risk warning, which has become a core technical bottleneck restricting the long-term safe and stable operation of offshore platform generator sets. Summary of the Invention

[0005] To address the above problems, this application provides a sensor-based method and system for early warning of salt spray corrosion of offshore platform generator sets, in order to solve the core problems of existing salt spray corrosion protection, which is mainly based on passive treatment, lacks dynamic sensing capabilities, and has low accuracy in predicting corrosion trends.

[0006] To achieve the above objectives, the technical solution adopted in this application is as follows: In a first aspect, this application provides a sensor-based method for early warning of salt spray corrosion of offshore platform generator sets, including: An environmental dataset is obtained by collecting data on salt spray particle concentration, wind speed, humidity, and temperature through a sensor network deployed around the generator set on the offshore platform. Based on the environmental dataset, a spatial distribution map of salt spray in the corresponding area of ​​the generator set component is constructed. Based on the aforementioned salt spray spatial distribution map, the spatial gradient evolution parameters, particle deposition location, and windward accumulation of salt spray on the surface of each component of the generator set are extracted to determine the dynamic distribution law of salt spray. When the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds the preset evolution threshold, the wind direction and humidity correlation data are integrated to calculate the salt spray corrosion risk on the surface of each component of the generator set and determine the degree of potential corrosion threat. To determine the degree of potential corrosion threat, historical corrosion data and material properties of the corresponding metal components are retrieved, and corrosion intensity correlation patterns are fitted in conjunction with salt spray exposure parameters to determine the corrosion sensitivity of the corresponding metal components. Based on the corrosion sensitivity, potential corrosion threat level and salt spray dynamic distribution pattern, the trend of erosion intensity change in the corresponding area of ​​the generator set component is predicted; Based on the trend of erosion intensity change, salt spray corrosion early warning signals and protection schemes are generated for the corresponding parts of the generator set.

[0007] Secondly, this application also provides a sensor-based salt spray corrosion early warning system for offshore platform generator sets, comprising: The data acquisition and mapping module is used to collect environmental datasets such as salt spray particle concentration, wind speed, humidity and temperature through a sensor network deployed around the generator set on the offshore platform, and to construct a spatial distribution map of salt spray in the corresponding area of ​​the generator set component based on the environmental dataset. The distribution pattern determination module is used to extract the spatial gradient evolution parameters, particle deposition location and windward accumulation of salt spray on the surface of each component of the generator set based on the salt spray spatial distribution map, and to determine the dynamic distribution pattern of salt spray. The corrosion risk assessment module is used to calculate the salt spray corrosion risk on the surface of each component of the generator set and determine the degree of potential corrosion threat when the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds the preset evolution threshold by integrating wind direction and humidity correlation data. The corrosion sensitivity determination module is used to retrieve the corrosion history data and material properties of the corresponding metal parts based on the degree of potential corrosion threat, combine them with salt spray exposure parameters, fit the corrosion intensity correlation pattern, and determine the corrosion sensitivity of the corresponding metal parts. The corrosion trend prediction module is used to predict the corrosion intensity change trend of the corresponding area of ​​the generator set components by combining the corrosion sensitivity, the degree of potential corrosion threat and the dynamic distribution law of salt spray. The early warning scheme generation module is used to generate salt spray corrosion early warning signals and protection schemes for corresponding parts of the generator set based on the trend of erosion intensity changes.

[0008] Compared with the prior art, this application has the following beneficial effects: (1) By collecting multi-dimensional real-time data of salt spray environment through distributed sensor network, a salt spray spatial distribution map that accurately matches the corresponding area of ​​generator set components is constructed to determine the dynamic distribution law of salt spray. This solves the pain point that traditional technology cannot accurately capture the spatial distribution difference and dynamic evolution characteristics of salt spray, realizes the targeted positioning of high-risk areas of salt spray corrosion, and provides accurate data support for corrosion risk assessment.

[0009] (2) Based on the dynamic distribution law of salt spray, the risk of salt spray corrosion is quantitatively calculated. By combining the material properties of metal components, corrosion history data and salt spray exposure parameters, the corrosion intensity correlation pattern is fitted to accurately determine the corrosion sensitivity of metal materials. This enables the accurate classification of potential corrosion threats of different components of the generator set, and greatly improves the accuracy and pertinence of salt spray corrosion risk assessment.

[0010] (3) Combining corrosion sensitivity, potential corrosion threat level and salt spray dynamic distribution law, accurately predict the erosion intensity change trend of corresponding areas of generator set components, and simultaneously generate graded early warning signals and adapted protection schemes to realize the transformation of salt spray corrosion from post-event passive protection to pre-event active early warning mode, and solve the industry problem of lagging traditional protection measures and poor prevention and control effect from the source.

[0011] (4) It forms a closed-loop control logic for the entire process of data collection, feature extraction, risk assessment, trend prediction, and early warning and prevention. It can adapt to the complex and ever-changing meteorological conditions at sea and can complete intelligent monitoring and early warning of salt spray corrosion without human intervention. This not only significantly improves the safety and stability of the generator set operation on the offshore platform, but also greatly reduces the cost of equipment corrosion prevention and maintenance. Attached Figure Description

[0012] Figure 1 This is a schematic flowchart illustrating a sensor-based early warning method for salt spray corrosion of offshore platform generator sets, provided in an embodiment of this application.

[0013] Figure 2 This is a schematic diagram of a sensor-based salt spray corrosion early warning system for offshore platform generator sets, provided as an embodiment of this application. Detailed Implementation

[0014] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0015] Reference Figure 1 This application provides a sensor-based method for early warning of salt spray corrosion of offshore platform generator sets, including: S11: By deploying a sensor network around the generator set on the offshore platform, collect data on salt spray particle concentration, wind speed, humidity and temperature to obtain an environmental dataset, and construct a spatial distribution map of salt spray in the corresponding area of ​​the generator set component based on the environmental dataset; S12: Based on the salt spray spatial distribution map, extract the spatial gradient evolution parameters, particle deposition location and windward accumulation of salt spray on the surface of each component of the generator set, and determine the dynamic distribution law of salt spray. S13: When the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds the preset evolution threshold, the wind direction and humidity correlation data are integrated to calculate the salt spray corrosion risk on the surface of each component of the generator set and determine the degree of potential corrosion threat. S14: Based on the degree of potential corrosion threat, retrieve the corrosion history data and material properties of the corresponding metal parts, combine them with salt spray exposure parameters, fit the corrosion intensity correlation pattern, and determine the corrosion sensitivity of the corresponding metal parts. S15: Based on the corrosion sensitivity, potential corrosion threat level and salt spray dynamic distribution law, predict the trend of erosion intensity change in the corresponding area of ​​generator set components; S16: Based on the trend of erosion intensity change, generate salt spray corrosion early warning signals and protection schemes for the corresponding parts of the generator set.

[0016] This embodiment discloses a sensor-based early warning method for salt spray corrosion of offshore platform generator units. It is applied to a deep-water semi-submersible oil and gas production platform in the South my country Sea. The platform is 180 kilometers from land, located in a tropical marine monsoon climate zone with an average annual temperature of 26.8℃ and an average annual relative humidity of 85%. The platform is affected by typhoons and strong monsoons for six months of the year. The prevailing wind direction is southeast, with an average annual wind speed of 7.2 m / s and an extreme maximum wind speed reaching 55 m / s. The average annual concentration of salt spray in the sea atmosphere is 3.5 mg / m³. 3 During a typhoon's passage, salt spray concentrations can reach as high as 12.8 mg / m³. 3 This area is one of the most severely affected by salt spray corrosion in my country's marine environment. The platform is equipped with three 12MW gas turbine generator sets, which supply power to all of the platform's production, living, and emergency systems. The generator sets are the platform's core power equipment; any failure due to salt spray corrosion will directly lead to production interruptions and even safety accidents. Historical operational data from the platform's first five years of operation shows that the generator sets experienced 17 electrical short circuits and metal component corrosion failures due to salt spray corrosion, resulting in a total of 328 hours of unplanned downtime and direct economic losses exceeding 120 million yuan. Salt spray corrosion has become a core risk factor restricting the long-term safe and stable operation of the platform's generator sets.

[0017] The method in this embodiment relies on the platform's industrial IoT intelligent monitoring and corrosion control system. The system comprises six core units working collaboratively: a distributed intelligent sensor network, an edge computing unit, a marine meteorological monitoring array, an automated monitoring system, corrosion control execution terminals, and a central control visualization platform. The system adopts a distributed control architecture, with intelligent terminals at each monitoring point performing local data acquisition, preprocessing, and edge computing, eliminating the need for centralized forwarding by a host computer. Data transmission latency is ≤10ms, and it features a single-point fault redundancy switching mechanism, avoiding the single-point failure risks of centralized control. It is suitable for 24 / 7 uninterrupted operation under extreme marine weather conditions. The system's distributed intelligent sensor network consists of intelligent sensors deployed around and at key locations of the generator set, including salt spray concentration sensors, integrated temperature and humidity sensors, wind speed and direction sensors, metal corrosion rate sensors, coating condition sensors, and stress monitoring sensors. All sensors are IP68-rated waterproof and corrosion-resistant encapsulated, adapting to the harsh operating environment of high salt spray and high humidity at sea. The sampling accuracy and response speed meet the real-time monitoring requirements of this method.

[0018] This embodiment provides a complete, detailed, and reproducible exposition of all the technical features of the method. All operating parameters, algorithm models, and judgment thresholds are based on GB / T 10125-2021 "Artificial Atmosphere Corrosion Test - Salt Spray Test", GB / T19292.1-2018 "Corrosion Classification of Metals and Alloys - Atmospheric Corrosion", SY / T 10008-2010 "Recommended Practices for Planning, Design and Construction of Offshore Fixed Platforms - Load and Resistance Coefficient Design Method", "Code for Corrosion Protection Design of Electrical Equipment in Offshore Petroleum Engineering", and the platform's continuous 36-month historical operating data and corrosion failure big data statistical calibration. All intermediate calculation steps and value bases are disclosed throughout the entire process of data acquisition, feature extraction, risk calculation, trend prediction, and early warning output. Custom technical names are explained simultaneously upon their first appearance. There are no skips, generalizations, or ambiguous descriptions throughout the entire process, ensuring full disclosure and complete disclosure of all implementation details of the method in this application, ensuring that those skilled in the art can repeatedly implement this application.

[0019] In step S11, an environmental dataset is obtained by collecting salt spray particle concentration, wind speed, humidity and temperature through a sensor network deployed around the generator set on the offshore platform, and a spatial distribution map of salt spray in the corresponding area of ​​the generator set component is constructed based on the environmental dataset.

[0020] In one implementation, this step is the data foundation of the entire method. Its core purpose is to acquire multi-dimensional environmental data around the generator set through a distributed sensor network, complete data standardization preprocessing and spatial location registration, and finally construct a salt spray spatial distribution map that accurately matches the corresponding area of ​​the generator set components, providing a standardized spatial data benchmark for subsequent salt spray distribution feature extraction and corrosion risk assessment.

[0021] The specific implementation process is as follows: First, taking three generator sets as the core, each generator set is divided into six core component areas: compressor room, combustion chamber, turbine end, intake system, exhaust system, and electrical control cabinet, totaling 18 monitoring sub-areas. Each monitoring sub-area is equipped with corresponding intelligent sensor points to achieve full-range, blind-spot-free monitoring of the generator sets. Through a sensor network deployed at 12 key monitoring points around the generator sets and 18 built-in monitoring points in the corresponding areas of each component, real-time environmental data such as salt spray particle concentration, wind speed, humidity, and temperature within the target area are continuously and synchronously collected to obtain the raw environmental dataset. The basic acquisition frequency of the sensor network is 1 time / minute. For high-risk warning areas identified subsequently, the acquisition frequency can be dynamically adjusted based on the warning signal, balancing monitoring accuracy and equipment lifespan.

[0022] In this step, intelligent salt spray concentration sensors are deployed at four locations on the generator set: the windward side, the leeward side, the top, and the bottom, with three points at each location, forming a three-dimensional monitoring network. This network can accurately capture differences in salt spray concentration at different spatial locations. The sensors use the laser scattering method for measurement, with a measurement range of 0-20 mg / m³. 3 Measurement accuracy ±0.05 mg / m 3 The response time is ≤1s. Integrated temperature and humidity intelligent sensors are deployed in the internal cavities and external shells of each component, synchronously collecting ambient temperature and humidity data as well as equipment surface temperature and humidity data. Temperature measurement range: -40℃~85℃, accuracy ±0.2℃; relative humidity measurement range: 0-100%RH, accuracy ±2%RH. Wind speed and direction intelligent sensors are deployed on the windward side and top of the generator platform, collecting real-time wind speed and direction data to provide basic data for salt spray propagation path analysis. Wind speed measurement range: 0-60m / s, accuracy ±0.1m / s; wind direction measurement range: 0-360°, accuracy ±1°. All sensors achieve microsecond-level synchronous acquisition through the IEEE 1588 precision time synchronization protocol. The protocol controls the sampling time deviation of different sensors within ±1μs through real-time measurement and compensation of network transmission delay, ensuring complete temporal alignment of environmental data from different locations and types, avoiding spatial distribution calculation deviations caused by time asynchrony. In this embodiment, the sensor network of a single generator set can generate 30 sets of raw environmental data per minute, and the three generator sets can generate a total of 90 sets of raw data per minute. Each set of data includes six core parameters: salt spray concentration, wind speed, wind direction, ambient temperature, ambient humidity, and equipment surface temperature, which fully characterize the salt spray environment status of the monitored area.

[0023] After the original environmental dataset is collected, it undergoes filtering preprocessing to remove invalid and abnormal data generated during the collection process, resulting in a standardized environmental dataset. The specific data processing procedure is divided into three levels: The first stage is gross error removal, which uses the Raida criterion (3σ criterion) to screen outliers in the raw data. First, the arithmetic mean and standard deviation of 20 consecutive sampling points of a single sensor are calculated. If the absolute value of the difference between the measured value of a certain sampling point and the mean exceeds 3 times the standard deviation, the data is determined to be an outlier and is removed. This step can effectively remove abnormal data with large jumps caused by sea splash, instantaneous power failure of the sensor, and electromagnetic interference.

[0024] The second level is rationality verification. After gross error elimination, invalid data that exceeds the sensor's measurement range or the reasonable range for the marine environment is further eliminated, including salt spray concentrations exceeding 20 mg / m³. 3 Data with temperatures exceeding -20℃ to 60℃ or relative humidity exceeding 0% to 100% are considered invalid and removed to ensure the physical validity of the data source.

[0025] The third stage is filtering and smoothing. A sliding window midpoint filtering algorithm is used to smooth the valid data after outlier removal. The window length is set to 10 consecutive sampling points. The algorithm executes as follows: the measurements of the 10 consecutive sampling points centered on the current sampling point are sorted in ascending order of numerical value. The arithmetic mean of the 5th and 6th digits after sorting is taken as the filtered output value of the current sampling point, replacing the original measurement value. This algorithm can effectively suppress the influence of impulse interference and random noise on the data while preserving the data trend characteristics, avoiding deviations in subsequent spatial distribution calculations, and ultimately obtaining a standardized environmental dataset.

[0026] Subsequently, combining the spatial layout information of each component of the generator set, the standardized environmental dataset is spatially registered to establish a one-to-one correspondence between the standardized environmental data and the corresponding regions of each component of the generator set, resulting in a calibrated environmental dataset with completed spatial registration. The specific process is as follows: First, a high-precision 3D model of the generator set is imported, and the 3D spatial coordinates of the corresponding regions of each core component are extracted. A spatial boundary model of the component region is established, clarifying the 3D coordinate range of each component's corresponding region. The latitude, longitude, altitude, and relative position information of each smart sensor's installation point are converted into spatial coordinates in a coordinate system unified with the generator set's 3D model, establishing a mapping relationship between sensor points and corresponding component regions, clarifying the component region to which each sensor point belongs. After completing the coordinate mapping, an inverse distance weighted spatial interpolation algorithm (IDW) is used to perform spatial interpolation calculations on the discrete sensor point data, obtaining the continuous salt spray environmental parameter distribution field within the entire spatial range of the generator set. The core calculation formula of the interpolation algorithm is:

[0027] In the formula, Z(x0) is the predicted environmental parameter value of the interpolation point x0, and Z(xi) is the value of the i-th known sensor point x0. i Standardized measurement value, d i From the interpolation point x0 to the known sensor point x i The Euclidean distance is calculated, where p is the distance weight exponent. In this embodiment, p is set to 2, a value determined through cross-validation using historical platform data, achieving an optimal balance between interpolation accuracy and computational efficiency. This algorithm performs interpolation calculations on each of the 0.1m×0.1m×0.1m grid nodes within the generator unit's three-dimensional space, obtaining a continuous distribution field of salt spray concentration, temperature, humidity, and wind speed parameters throughout the entire generator unit space, accurately reconstructing the continuous distribution of salt spray within the generator unit space. Through spatial interpolation and coordinate registration, standardized environmental data is accurately mapped to the three-dimensional spatial coordinates of the corresponding areas of each component of the generator unit, establishing a one-to-one correspondence between the corresponding area of ​​each component and the environmental data, resulting in a calibrated environmental dataset with complete spatial registration. This completely solves the problem of "data disconnection from equipment location and inability to accurately locate salt spray erosion areas" in traditional monitoring methods.

[0028] Finally, based on the calibration environment dataset, a salt spray spatial distribution map is constructed that corresponds one-to-one with the corresponding areas of each component of the generator set. This salt spray spatial distribution map uses the generator set's three-dimensional spatial model as the visualization base, salt spray concentration as the core characterization parameter, and combines wind speed, temperature, and humidity data. It uses color thermal gradients to represent the salt spray concentration and aggregation degree at different spatial locations, creating a three-dimensional visualization distribution map. The distribution map can intuitively display the spatial distribution differences, aggregation state, and propagation path of salt spray in the corresponding areas of each component of the generator set, achieving accurate spatial matching between salt spray distribution data and equipment component areas. This provides a precise spatial benchmark for the subsequent extraction of dynamic salt spray distribution patterns. In this embodiment, the salt spray spatial distribution map is updated every minute, reflecting the dynamic changes in the salt spray environment around the generator set in real time. It also supports functions such as historical data backtracking, spatial gradient calculation, and highlighting of abnormal areas. It can be directly connected to the platform's central control visualization platform, providing intuitive monitoring data support for maintenance personnel.

[0029] In step S12, based on the salt spray spatial distribution map, the spatial gradient evolution parameters, particle deposition location and windward accumulation of salt spray on the surface of each component of the generator set are extracted to determine the dynamic distribution law of salt spray.

[0030] In one embodiment, the core objective of this step is to accurately extract the spatial distribution and dynamic change characteristics of salt spray on the surface of various components of the generator set based on the salt spray spatial distribution map through image segmentation, spatial feature extraction, and temporal feature analysis. Ultimately, the dynamic distribution law of salt spray is determined, and the accumulation area, evolution trend, and key corrosion area of ​​salt spray on the equipment surface are identified, providing a core basis for subsequent corrosion risk assessment.

[0031] The specific implementation process is as follows: First, retrieve the real-time updated spatial distribution map of salt spray. Using a preset image segmentation algorithm, divide the salt spray monitoring sub-regions into areas that match the corresponding regions of each component of the generator set, thus obtaining the component-level region segmentation results. The preset image segmentation algorithm used in this step is the watershed image segmentation algorithm. This algorithm is an image region segmentation method based on topological topography theory. The core principle is to regard the grayscale image as a topological topographic map. The grayscale value of each pixel in the image corresponds to the altitude of that point. The higher the grayscale value, the higher the altitude, and the lower the grayscale value, the lower the altitude. The local grayscale minimum value region in the image is regarded as a catchment basin, and the local grayscale maximum value region is regarded as a watershed line. By simulating the water immersion process, water is gradually immersed starting from the catchment basin. When the water flows from different catchment basins meet, the watershed boundary is constructed, ultimately achieving accurate segmentation of different regions in the image. The core function of this algorithm in this application is to accurately divide the salt spray spatial distribution map into independent monitoring sub-regions corresponding to each component of the generator set based on the component boundaries of the 3D model of the generator set. This avoids cross-interference of salt spray data from different component regions and ensures the regional specificity of subsequent feature extraction. The detailed execution process is as follows: First, the 3D salt spray spatial distribution map is sliced ​​along the axial direction of the generator set to obtain 6 two-dimensional grayscale heat maps corresponding to the core components of the generator set. The grayscale value is linearly positively correlated with the salt spray concentration value; the higher the salt spray concentration, the larger the grayscale value of the corresponding pixel. Gaussian filtering is then applied to the two-dimensional grayscale heat maps with a kernel size of 5×5 and a standard deviation σ=1 to eliminate the influence of image noise on the segmentation results. A So... The bel operator calculates the gradient of the smoothed grayscale image to obtain the gradient magnitude map. The gradient magnitude corresponds to the spatial rate of change of salt spray concentration; the larger the gradient magnitude, the more significant the spatial difference in salt spray concentration. Based on the component boundary information of the unit's 3D model, the gradient magnitude map is marked with foreground and background. The foreground is marked as the core region corresponding to each component, and the background is marked as the gap region between components and the external region of the unit. Based on the marking results, a watershed transformation is performed, treating the gradient magnitude map as a topological topographic map, with grayscale values ​​corresponding to altitude. Image segmentation is achieved by simulating the immersion process, ultimately obtaining salt spray monitoring sub-regions that match the corresponding regions of each generator unit component, completing the component-level region division. In this embodiment, the image segmentation algorithm accurately divides 18 salt spray monitoring sub-regions that match the corresponding regions of the generator unit components, defining precise spatial boundaries for subsequent feature extraction. The segmentation accuracy reaches 98.6%, with no region overlap or omission issues.

[0032] For each salt spray monitoring sub-region in the component-level region division results, the spatial gradient evolution parameters of salt spray on the surface of each component of the generator set, particle deposition location, and windward accumulation amount are extracted to obtain the basic parameters of salt spray spatial distribution. The detailed calculation process of each parameter is as follows: Calculation of spatial gradient evolution parameters: For each monitoring sub-region, based on the interpolation results of the salt spray spatial distribution map, the first-order spatial partial derivative of the salt spray concentration distribution function is calculated to obtain the gradient components of the salt spray concentration in the three spatial dimensions of x, y, and z. The gradient magnitude is the spatial gradient evolution parameter, and the calculation formula is as follows: In the formula, The spatial gradient evolution parameter is expressed in mg / (m²). 3 m), where C is the salt spray concentration distribution function. , , These are the partial derivatives of the salt spray concentration in the three spatial dimensions. This parameter characterizes the spatial accumulation trend of salt spray on the surface of a component. The larger the value, the higher the degree of local accumulation of salt spray in that area, and the greater the risk of corrosion.

[0033] Particle deposition location: For each monitoring sub-region, the local maximum value of salt spray concentration is located by using a salt spray concentration peak detection algorithm. Combined with the structural characteristics of the components, the main deposition locations of salt spray particles are determined. Key areas prone to deposition, such as equipment gaps, flange connections, air inlets, and structural corners, are marked. At the same time, the peak salt spray concentration and area of ​​each deposition location are recorded.

[0034] Calculation of Salt Fog Accumulation on the Windward Side: For each monitoring sub-area, based on real-time wind direction data, the windward side area of ​​the component is divided. The salt spray deposition flux within the windward side area is calculated cumulatively over time to obtain the cumulative salt spray adhesion per unit area, i.e., the windward accumulation. The specific calculation process is as follows: First, a fixed calculation time window of 30 minutes is set, and the time window is divided into several consecutive calculation periods according to the sensor acquisition step size. Within each calculation period, the real-time salt spray concentration, real-time wind speed, and the angle between the real-time wind direction and the normal of the windward side of the component are collected synchronously. Combined with the salt spray particle deposition efficiency coefficient on the corresponding metal surface, which has been calibrated in advance through on-site hanging plate tests, the salt spray deposition flux within that period is calculated. The salt spray deposition flux of all periods within the entire time window is accumulated and summed to finally obtain the cumulative salt spray adhesion per unit area of ​​the metal surface, i.e., the windward accumulation. The smaller the angle between the wind direction and the normal to the windward surface, the greater the effective settling amount of salt spray particles, with the effective settling amount reaching its peak when the incident air is perpendicular. The settling efficiency coefficient is calibrated differently depending on the metal material of the component and the state of the surface coating. In this embodiment, the settling efficiency coefficient for the 316L stainless steel surface is calibrated to 0.0028. Through the above cumulative calculations, the cumulative adhesion level of salt spray on the windward surface can be accurately quantified, providing basic data for subsequent corrosion risk assessment.

[0035] In this embodiment, the calculated evolution parameter of the salt spray spatial gradient on the windward side of the compressor compartment is 0.22 mg / (m²). 3 The combustion chamber concentration is 0.08 mg / (m²). 3 The concentration in the turbine end region is 0.05 mg / (m). 3 The concentration of the electrical control cabinet area is 0.10 mg / (m). 3 The study clarified the spatial distribution differences of salt spray on different component surfaces. The particle deposition locations were mainly concentrated at the compressor inlet, the flange connection on the windward side of the unit casing, and the ventilation opening of the electrical control cabinet. Statistics on the accumulation on the windward side showed that the salt spray accumulation on the southeast windward side was 4.7 times that on the leeward side, which fully demonstrated the significant influence of wind direction on salt spray deposition.

[0036] Combining the continuously acquired spatial distribution map of salt spray with the component-level regional division results, the temporal variation characteristics of salt spray parameters in each salt spray monitoring sub-region are extracted to obtain the dynamic variation characteristics of salt spray. The specific data processing procedure is as follows: using a 1-minute time step, the time series of the maximum value, average value, and spatial gradient evolution parameters of salt spray concentration in each monitoring sub-region are extracted within 30 consecutive time periods; a univariate linear regression model is used to fit the trend of the time series, with time as the independent variable and the measured salt spray parameter value as the dependent variable. The model parameters are solved using the least squares method to obtain the temporal change rate of the salt spray parameters. Positive values ​​represent an increasing trend in salt spray concentration, negative values ​​represent a decreasing trend, and the larger the absolute value, the faster the change rate. Through the above fitting calculation, the dynamic variation characteristics of salt spray on the surface of each component are obtained, fully characterizing the change law of salt spray concentration and accumulation area over time. In this embodiment, the calculated temporal change rate of salt spray concentration on the windward side of the compressor compartment is 0.08 mg / (m³). 3 The concentration of salt spray in the leeward region showed a continuous upward trend (h), while the temporal variation rate of salt spray concentration in the leeward region was only 0.01 mg / (m³). 3 h), with no obvious upward trend, clarifying the dynamic evolution differences in salt spray distribution in different regions.

[0037] Finally, by integrating the basic parameters of salt spray spatial distribution and the dynamic change characteristics of salt spray, the dynamic distribution law of salt spray on the surface of each component of the generator set was determined. This law fully characterizes the spatial distribution differences, temporal evolution characteristics, accumulation areas and propagation paths of salt spray on the surface of each component of the generator set, providing complete basic data support for subsequent corrosion risk assessment.

[0038] In step S13, when the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds the preset evolution threshold, the wind direction and humidity correlation data are fused to calculate the salt spray erosion risk on the surface of each component of the generator set and determine the degree of potential corrosion threat.

[0039] In one implementation, the core objective of this step is to accurately calculate the risk of salt spray erosion based on the dynamic distribution pattern of salt spray, through threshold comparison, multi-source data fusion, and risk matrix quantification, and to determine the degree of potential corrosion threat to the surface of the corresponding component, thus providing a clear target and data foundation for subsequent material sensitivity analysis and trend prediction.

[0040] The specific implementation process is as follows: First, the spatial gradient evolution parameters corresponding to the dynamic distribution law of the salt spray are retrieved and compared item by item with the preset evolution threshold to obtain the parameter threshold comparison results. In this embodiment, the preset evolution threshold is set to 0.15 mg / (m³). 3 The threshold (m) is calibrated using 36 months of corrosion failure data from the platform. When the spatial gradient change rate of salt spray exceeds this threshold, the corrosion rate of the corresponding metal material will exhibit a nonlinear jump, with the jump reaching 3-5 times that under normal conditions. This is the core critical value for determining abnormal salt spray accumulation. During the comparison process, for the spatial gradient evolution parameters of each monitoring sub-region, the average value of three consecutive sampling periods is used as the comparison benchmark value to avoid misjudgment caused by fluctuations in a single sampling. Simultaneously, the coefficient of variation is checked for the parameter values ​​over the three periods. If the coefficient of variation exceeds 20%, the sampling period is extended to five periods to ensure the stability of the comparison benchmark value. In this embodiment, the average value of the spatial gradient evolution parameters of the monitoring sub-region on the windward side of the compressor room is 0.22 mg / (m²). 3 The spatial gradient evolution parameters of the remaining 17 monitoring sub-regions were all below the preset evolution threshold, with no abnormalities.

[0041] Based on the parameter threshold comparison results, abnormal component regions whose spatial gradient evolution parameters exceed the preset evolution threshold are selected. Simultaneously, wind direction and humidity correlation data corresponding to these abnormal component regions are retrieved to obtain an abnormal region analysis dataset. This wind direction and humidity correlation data includes the dominant wind direction, wind speed temporal variation data, the angle between the wind direction and the component surface, ambient relative humidity, equipment surface humidity, and humidity temporal fluctuation range data for the region. All data perfectly match the time period of the salt spray dynamic distribution pattern, with consistent sampling time steps to ensure data temporal consistency. In this embodiment, the retrieved abnormal region data shows that the region faces the dominant southeast wind direction over the sea, with a real-time wind speed of 6.8 m / s. The angle between the wind direction and the windward surface of the component is 85°, nearly perpendicular to the incident direction, making it highly susceptible to salt spray particle deposition. Simultaneously, the ambient relative humidity in this region is stable at 88%, falling within the highly sensitive humidity range (60%-90%) for metal salt spray corrosion. Within this humidity range, a continuous water film easily forms on the metal surface, providing the necessary conditions for salt spray electrochemical corrosion and significantly accelerating the corrosion process.

[0042] By integrating the data from the abnormal area analysis dataset, the salt spray erosion risk of the abnormal component area is calculated, and the salt spray erosion risk classification result is obtained. In this embodiment, the salt spray erosion risk is divided into four levels: low, medium, high, and extremely high, corresponding to four warning signals: blue, yellow, orange, and red, respectively. Different levels correspond to different protection schemes and control strategies. The specific data processing process is as follows: First, the weights of the risk assessment indicators are determined by the analytic hierarchy process (AHP), and an evaluation system containing three primary indicators: salt spray anomaly degree, wind speed influence degree, and humidity influence degree is constructed. By constructing a judgment matrix and performing consistency verification, the weights of the three indicators are determined to be 0.5, 0.25, and 0.25, respectively, and the consistency ratio CR = 0.03 < 0.1, which meets the consistency requirements. Then, each indicator is normalized and quantified in the [0,1] interval, where the quantified value of the salt spray anomaly degree is the spatial gradient evolution parameter of the abnormal area and the preset evolution parameter. The ratio of the variable threshold is truncated to 1 for any portion exceeding 1. In this embodiment, the value is approximately 1.47 (0.22 / 0.15), which becomes 0.85 after normalization. The quantification value of the wind speed impact is based on a comprehensive calibration of the angle between wind speed and wind direction; in this embodiment, the value is 0.92. The quantification value of the humidity impact is based on a calibration of the matching degree between relative humidity and the corrosion-sensitive range; in this embodiment, the value is 0.96. Finally, the final risk value is calculated by weighted summation. The normalized quantification values ​​of each indicator are multiplied by their corresponding weight coefficients and then summed to obtain the final salt spray corrosion risk value. In this embodiment, the calculated final risk value is 0.895, corresponding to a high-risk level.

[0043] Finally, based on the salt spray corrosion risk classification results, the potential corrosion threat level of the surfaces of each component of the generator set was determined. In this embodiment, the potential corrosion threat level of the metal surface of the compressor chamber was determined to be high, thus locking in the target area for subsequent corrosion sensitivity analysis.

[0044] In step S14, based on the potential corrosion threat level, the corrosion history data and material properties of the corresponding metal parts are retrieved, and combined with salt spray exposure parameters, an erosion intensity correlation pattern is fitted to determine the corrosion sensitivity of the corresponding metal parts.

[0045] In one implementation, the core objective of this step is to construct an erosion intensity correlation model for areas with corrosion threats by combining metal material properties, historical corrosion data, and salt spray exposure parameters through multiple linear regression fitting. This model quantifies the sensitivity of metal materials to salt spray corrosion, clarifies the corrosion vulnerability of different materials in the current environment, and provides a material-level correction basis for subsequent erosion intensity trend prediction.

[0046] The specific implementation process is as follows: First, the corresponding area of ​​the generator set component corresponding to the potential corrosion threat level is located. Corrosion history data and material properties of the metal components within the corresponding area are retrieved to construct a basic corrosion dataset. In this embodiment, the salt spray corrosion sensitivity of metal materials is divided into four levels: low sensitivity (S < 0.3), medium sensitivity (0.3 ≤ S < 0.6), high sensitivity (0.6 ≤ S < 0.8), and extremely high sensitivity (S ≥ 0.8). The materials used for the core components of the generator set include 316L austenitic stainless steel, Q355ND low-carbon alloy steel, 6061-T6 aluminum alloy, and epoxy zinc-rich coated carbon steel. The basic corrosion resistance of different materials is calibrated through the GB / T10125-2021 standard neutral salt spray test and on-site plate test. The components in the target area of ​​this step... The material is 316L austenitic stainless steel with a chromium content of 16%-18.5%, a nickel content of 10%-14%, and a molybdenum content of 2%-3%. It has excellent resistance to chloride ion corrosion. The corrosion history data includes the historical corrosion rate, corrosion failure cases, and coating aging data of this material under the same operating conditions on the platform. In this embodiment, the average annual corrosion rate of this material under the platform operating conditions is 0.08 mm / a. No serious corrosion failure has occurred in the history of operation, but pitting pits with a maximum depth of 0.12 mm have appeared under high salt spray conditions. All of the above data are included in the basic corrosion dataset.

[0047] The salt spray exposure parameters of the corresponding areas of the generator set components are retrieved synchronously and added to the basic corrosion dataset to obtain a complete corrosion analysis dataset. The salt spray exposure parameters here include the annual cumulative salt spray exposure duration, historical cumulative salt spray concentration, and salt spray exposure intensity under current operating conditions for the area. In this embodiment, the annual cumulative salt spray exposure duration for this area is 8200 hours, accounting for 93.6% of the total annual operating time, and the historical cumulative salt spray concentration is 28760 mg / L. h / m 3 The current salt spray exposure intensity is 3.2 mg / m³. 3 This falls under high-exposure conditions. Furthermore, Z-score standardization was performed on all continuous numerical variables in the complete corrosion analysis dataset to eliminate dimensional differences between parameters, providing standardized data input for subsequent model fitting.

[0048] Based on the complete corrosion analysis dataset, a correlation model was obtained by fitting the corrosion intensity relationship between salt spray corrosion intensity and environmental parameters and metal material properties. The specific fitting process was as follows: First, the dependent and independent variables of the model were determined. The salt spray corrosion rate (i.e., corrosion intensity) of the metal material was used as the dependent variable, and salt spray concentration, relative humidity, ambient temperature, and material corrosion resistance coefficient were used as independent variables to construct a multiple linear regression model. Then, 36 months of historical corrosion data and environmental monitoring data from the platform were used as the training set, totaling 1280 valid samples. The regression coefficients of the model were solved using the ordinary least squares method to obtain the initial fitting model. The initial model was subjected to significance and goodness-of-fit tests. The F-test showed a significance p-value < 0.001, indicating that the overall linear relationship of the model was significant. The adjusted coefficient of determination R0 was also significant. adj 2 =0.94, indicating that the model can explain 94% of the corrosion rate variation and has excellent fitting effect; t-tests were performed on the independent variables, and the p-values ​​of all independent variables were <0.05, which are all significant influencing factors. Finally, the optimized corrosion intensity correlation fitting model was obtained, which can accurately quantify the influence weight of each parameter on the salt spray corrosion intensity and clarify the quantitative correlation between environmental parameters, material properties and corrosion rate.

[0049] Based on the corrosion intensity correlation fitting model, the corrosion sensitivity of the metal components in the corresponding area of ​​the generator set components is determined. The corrosion sensitivity quantification formula is as follows: In the formula: S represents the salt spray corrosion sensitivity of the metallic material, ranging from 0 to 1. A higher value indicates a stronger sensitivity of the material to salt spray corrosion; α, β, and γ are dimensional weight coefficients, taking values ​​of 0.5, 0.3, and 0.2 respectively, calibrated according to the marine environment metallic material corrosion failure weight matrix. This matrix is ​​obtained through statistical analysis of historical corrosion failure data from the platform, with the material's inherent corrosion resistance having the highest weight on corrosion sensitivity; C is the normalized value of the material's corrosion resistance, calculated based on the annual corrosion rate of the material in the standard salt spray test. The higher the corrosion rate, the larger the normalized value. In this embodiment, the value for 316L stainless steel is 0.52; T is the normalized value of the effect of temperature fluctuation, calculated by combining the equipment operating temperature fluctuation range and the corrosion rate temperature coefficient. The larger the temperature fluctuation amplitude, the larger the normalized value. In this embodiment, the value is 0.68; H is the normalized value of the effect of humidity and salt spray exposure duration, comprehensively calibrated by combining environmental humidity and salt spray exposure duration. The higher the humidity and the longer the exposure duration, the larger the normalized value. In this embodiment, the value is 0.65.

[0050] Substituting the data from this embodiment, the salt spray corrosion sensitivity of the 316L stainless steel compressor chamber is calculated as S = 0.5 × 0.52 + 0.3 × 0.68 + 0.2 × 0.65 = 0.58, which is classified as a medium corrosion-sensitive material. This clarifies the susceptibility of this material to salt spray corrosion under the current operating conditions.

[0051] In step S15, the corrosion sensitivity, potential corrosion threat level and salt spray dynamic distribution pattern are combined to predict the erosion intensity change trend of the corresponding area of ​​the generator set component.

[0052] In one implementation, the core objective of this step is to combine corrosion sensitivity, the degree of potential corrosion threat, and the dynamic distribution pattern of salt spray, and use an Auto-Regressive Integrated Moving Average (ARIMA) time series analysis model to predict the salt spray concentration trend. This is then corrected by incorporating material sensitivity to accurately predict the erosion intensity change trend of local areas of the equipment over a future period, thereby achieving early warning of salt spray corrosion and fundamentally solving the core pain point of traditional technologies: "post-event handling and delayed early warning."

[0053] The specific implementation process is as follows: First, based on the corrosion sensitivity, potential corrosion threat level, and dynamic distribution pattern of salt spray, a basic dataset for trend analysis is integrated. This dataset contains 72 consecutive hours of time-series data on salt spray concentration, wind speed and direction, temperature and humidity, as well as quantitative values ​​for corrosion sensitivity, risk level, and salt spray spatial gradient evolution parameters. All data are aligned according to time series, with a sampling interval of 1 minute, totaling 4320 sets of valid data, providing a complete time-series data foundation for trend fitting. Simultaneously, the time-series data is preprocessed, using linear interpolation to fill in missing values, with the missing value rate controlled within 0.5%, ensuring the continuity of the time-series data; stationarity preprocessing is also performed on the time-series data to lay the foundation for subsequent time-series analysis.

[0054] Using a pre-defined time series analysis technique, feature extraction and trend fitting are performed on the time-series variation data of salt spray concentration and environmental parameter fluctuation data in the trend analysis dataset to obtain the initial erosion intensity evolution trend. In this embodiment, the pre-defined time series analysis technique is an Auto-Regressive Integrated Moving Average (ARIMA) model, where p is the autoregressive order, d is the differencing order, and q is the moving average order. The detailed model construction, feature extraction, and trend fitting process is divided into five core steps, and the calculation logic and implementation details are fully disclosed: Stationarity Test and Difference Processing of Time Series Data: The stationarity of time series data is a prerequisite for fitting the ARIMA model. This step uses the ADF unit root test to test the stationarity of the salt spray concentration time series data. The null hypothesis is "the time series data has a unit root and is a non-stationary series," and the alternative hypothesis is "the time series data does not have a unit root and is a stationary series." In this embodiment, the ADF test statistic of the original salt spray concentration time series data is -1.28, and the corresponding p-value is 0.32 > 0.05. Therefore, the null hypothesis cannot be rejected, and the original time series data is determined to be a non-stationary series.

[0055] The original data were first-order differencing, and the ADF unit root test was performed again on the differencing sequence. The test statistic was -8.62, and the corresponding P value was <0.01. At the 99% confidence level, the null hypothesis was rejected, and the differencing sequence was determined to be a stationary sequence. Therefore, the difference order of the model was determined to be d=1.

[0056] Model order determination and parameter optimization: The core of model order determination is to determine the autoregression order p and the moving average order q. This step adopts a combined order determination method of "graphical qualitative method + information criterion quantitative method": First, the autocorrelation function (ACF) and partial autocorrelation function (PACF) plots of the stationary series after differencing are plotted. The range of values ​​for p and q is initially determined by the tailing and truncation characteristics. The ACF plot shows that the series decays rapidly after lag of 1, exhibiting tailing characteristics. The PACF plot shows that the series is truncated after lag of 2. The range of values ​​for p and q is initially determined to be 0-3.

[0057] Subsequently, the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were used for quantitative order determination. All 16 order combinations with p=0-3 and q=0-3 were iterated, and ARIMA models were fitted for each combination, with the corresponding AIC and BIC values ​​calculated. Smaller AIC and BIC values ​​indicate higher model fitting accuracy and stronger generalization ability. Through iterative calculation, when p=2 and q=1, both the AIC and BIC values ​​of the model reached their minimum values. Therefore, the optimal order of the model was finally determined to be ARIMA(2,1,1).

[0058] Model parameter estimation and significance test: The ARIMA(2,1,1) model was estimated using the conditional least squares method to obtain the estimated values ​​of the autoregressive coefficients and moving average coefficients. The specific parameter estimation results are as follows: the autoregressive term AR(1) coefficient is 0.72 (P value < 0.01), the AR(2) coefficient is -0.28 (P value < 0.01), and the moving average term MA(1) coefficient is 0.45 (P value < 0.01). The P values ​​of all parameters are less than 0.01, and the parameter estimation results are significantly effective at the 99% confidence level.

[0059] White noise test of model residuals: The model residual test is the core step in verifying the effectiveness of the model. If the model residuals are a white noise sequence, it means that the model has extracted all the effective information from the time series data, and there are no remaining interpretable patterns. If the residuals are not a white noise sequence, it means that the model still has room for optimization. In this step, the Ljung-Box white noise test is used to test the model residual sequence. The lag order is set to 12. The test results show that the chi-square statistic is 10.26, and the corresponding p-value is 0.59 > 0.05. The null hypothesis that "the residual sequence is white noise" cannot be rejected. Therefore, the model residuals are determined to be a white noise sequence, the model fit is effective, and it can be used for subsequent trend prediction.

[0060] Trend Prediction and Initial Erosion Intensity Conversion: Using the validated ARIMA(2,1,1) model, rolling predictions of salt spray concentration for the next 24 hours are performed, outputting a sequence of predicted salt spray concentration values ​​and a 95% confidence interval. The prediction step size is 1 minute, with a total of 1440 prediction points, fully covering the time window of the next 24 hours. Combining the erosion intensity correlation fitting model constructed in step S14, the predicted salt spray concentration values ​​and environmental parameter prediction values ​​are substituted into the model to calculate the corrosion rate prediction sequence for the next 24 hours point by point, i.e., the evolution trend of the initial erosion intensity, characterizing the change law of salt spray erosion intensity without material sensitivity correction.

[0061] Based on the corrosion sensitivity, the evolution trend of the initial erosion intensity is corrected and calibrated to obtain the calibrated time-series evolution result of salt spray erosion intensity. The correction formula is as follows: In the formula: I t Let I0 be the predicted corrosion intensity for the next t time period; S be the corrosion sensitivity of the metal material; and k be the salt spray concentration trend change rate, calculated from the salt spray concentration change value predicted by the time series model. The calculation logic is the difference between the predicted salt spray concentration at time t and the measured value at the current time, divided by the measured value at the current time, to obtain the salt spray concentration trend change rate. In this embodiment, the calculated trend change rate of salt spray concentration for the next 24 hours is k=0.72, the corrosion sensitivity is S=0.58, and the initial corrosion intensity benchmark value I0 is 0.06 mm / a. Substituting these values ​​into the formula, the predicted corrosion intensity for the next 24 hours is It=0.06×(1+0.58×0.72)≈0.085 mm / a, which is 42% higher than the current corrosion intensity.

[0062] Based on the calibrated time-series evolution of salt spray erosion intensity, the future trend of erosion intensity in corresponding areas of each component of the generator set is predicted. The specific prediction logic is as follows: First, based on the calibrated time-series evolution of erosion intensity, the peak, trough, average rate of change, and cumulative magnitude of erosion intensity within the next 24 hours are determined. Second, combined with the salt spray erosion risk grading standard, the erosion intensity is divided into four levels: low, medium, high, and extremely high, corresponding to the erosion risk level at different future time periods. Finally, the direction of change, magnitude of change, and risk level transition nodes of the erosion intensity in the target area are clarified, completing the accurate prediction of the future trend of erosion intensity. In this embodiment, the salt spray erosion intensity in the compressor room area will continue to rise within the next 24 hours, exceeding the high-risk level threshold in approximately 12 hours, and remaining in the high-erosion-risk range for the next 24 hours, providing a core basis for subsequent early warning signal generation and protection scheme matching.

[0063] In step S16, based on the trend of erosion intensity change, a salt spray corrosion early warning signal and protection scheme for the corresponding part of the generator set are generated.

[0064] In one implementation, the core objective of this step is to select suitable protection schemes based on the trend of erosion intensity changes, combined with the actual operating status and structural characteristics of the equipment, through a multi-dimensional matching algorithm, generate graded early warning signals, realize closed-loop management of "early warning-prevention", and eliminate the risk of salt spray corrosion in its infancy.

[0065] The specific implementation process is as follows: First, retrieve the future trend of corrosion intensity in the corresponding areas of each component of the generator set, combine the surface coating status and local stress concentration data of the corresponding component's metal parts, match the preset protection level requirements, and obtain the protection adaptation parameters of the target component. The surface coating status is collected by a coating status intelligent sensor. The sensor uses electrochemical impedance spectroscopy to measure the coating impedance value and assess the coating integrity, aging degree, and damage condition. In this embodiment, the epoxy zinc-rich coating integrity rate in the compressor chamber area is 82%, with no large-area peeling, but there is slight local aging, and the coating impedance value has decreased by 18% compared to the initial value. Local stress concentration data is collected by a stress monitoring intelligent sensor. Strain gauge sensors are used to measure the static and dynamic stress of the components. In this embodiment, the flange connection of the compressor chamber is a stress concentration area with a stress value of 72% of the design allowable stress. The risk of corrosion failure in this area is much higher than in other parts. The preset protection level requirements are divided into four levels according to the functional importance of the component, the scope of the shutdown impact, and the severity of the failure consequences. The compressor chamber is a core component of the generator set. Once corrosion failure occurs, it will lead to unplanned shutdown of the unit, corresponding to the level two protection requirement. The above data are integrated to obtain the protection adaptation parameters for the target component.

[0066] Based on the target component's protection adaptation parameters, a protection scheme matching the corrosion intensity level and metal material properties is extracted from a pre-set corrosion control parameter library. In this embodiment, the system's built-in corrosion control parameter library includes 28 standardized protection execution schemes across six categories: ventilation and dehumidification, coating repair, cathodic protection, sealing protection, shutdown protection, and emergency isolation. Each scheme is labeled with the appropriate corrosion intensity level, metal material type, component functional attributes, and protection level requirements, and can be directly deployed to the corrosion control execution terminal for execution without manual secondary configuration. The matching of protection schemes employs a multi-dimensional weighted matching algorithm. The specific process is as follows: three matching dimensions are constructed: corrosion intensity matching degree, material compatibility, and protection level matching degree, with weights of 0.4, 0.3, and 0.3, respectively; the matching degree of each dimension is quantitatively scored within the [0,1] interval, with higher scores indicating better compatibility; the comprehensive matching score of each candidate scheme is calculated, and the scheme with the highest comprehensive score is selected as the optimal adaptation scheme. In this embodiment, the standardized protection implementation plan obtained is "30% increase in ventilation and dehumidification system power + special inspection of windward sealing structure + contingency plan for local coating repair". The plan clarifies the implementing entity, implementation steps, acceptance standards, and emergency response procedures, and can be directly issued to the corrosion prevention and control implementation terminal for execution.

[0067] Based on the classification results of the future trend of corrosion intensity and the protection adaptation parameters of the target component, a graded salt spray corrosion early warning signal for the corresponding component of the generator set is generated. In this embodiment, the early warning signal is divided into four levels: blue, yellow, orange, and red, corresponding to low, medium, high, and extremely high corrosion intensity levels, respectively. Different levels of early warning signals correspond to different push ranges, handling time limits, and control requirements. In this embodiment, the future trend of corrosion intensity is at a high-risk level, corresponding to an orange level-two early warning signal. The early warning signal includes the risk area, risk level, predicted corrosion intensity, suggested handling measures, and handling time limit, and is simultaneously pushed to the platform control system, on-site operation and maintenance terminals, and equipment management system to achieve multi-terminal synchronous early warning. The handling time limit is set to 24 hours.

[0068] Step S16 is followed by step S17: using the salt spray corrosion early warning signal, the frequency of the sensor network collecting data on the correlation between the amount of salt spray accumulation on the windward side and humidity fluctuation is adjusted in real time, and the dynamic distribution law of salt spray is continuously updated. In one embodiment, the core objective of this step is to dynamically adjust the acquisition frequency of the sensor network by driving the early warning signal, complete the real-time processing and distribution pattern update of high-frequency monitoring data, and feed the updated distribution pattern back to the corrosion threat determination stage, thereby realizing closed-loop optimization and dynamic adaptive control of the entire method and adapting to the changing working conditions of the marine environment.

[0069] The specific implementation process is as follows: First, based on the classification results of the salt spray corrosion early warning signal, a sampling frequency adjustment command is sent to the sensor network to match and determine the corresponding sampling frequency range of the sensor network. In this embodiment, a mapping rule between early warning level and sampling frequency is pre-established: blue early warning (low risk) corresponds to a sampling frequency of 1 time / 30 seconds, yellow early warning (medium risk) corresponds to 1 time / 15 seconds, orange early warning (high risk) corresponds to 1 time / 10 seconds, and red early warning (extremely high risk) corresponds to 1 time / 2 seconds. In this embodiment, the generated orange level 2 early warning corresponds to a high-priority sampling mode, so the matched and determined target sampling frequency range is 1 time / 10 seconds, replacing the original basic sampling frequency of 1 time / minute, to achieve high-frequency encrypted monitoring of high-risk areas. At the same time, a dynamic adjustment rule for the sampling frequency is set: if the early warning level is upgraded to red level 1 early warning, the sampling frequency will be further increased to 1 time / 2 seconds; if the early warning level is reduced to yellow level 1 early warning, the sampling frequency will be adjusted to 1 time / 30 seconds; if the early warning is lifted, it will return to the basic sampling frequency, taking into account both monitoring accuracy and equipment lifespan.

[0070] According to the aforementioned acquisition frequency range, the acquisition frequency of the sensor network for the correlation between the accumulation amount and humidity fluctuation of the corresponding components of the generator set on the windward side is adjusted in real time. Simultaneously, filtering preprocessing and standardization processing of the acquired data are completed to obtain a high-time-density salt spray monitoring dataset. In this embodiment, under high-frequency acquisition mode, the intelligent salt spray concentration sensor and the integrated temperature and humidity sensor in the windward side area of ​​the compressor room acquire data once every 10 seconds. A single unit can generate 600 sets of monitoring data per minute, which is 20 times higher than the basic acquisition mode. This allows for accurate capture of instantaneous fluctuations and subtle changes in salt spray concentration, avoiding data omissions in the basic acquisition mode. Simultaneously, the raw data acquired at high frequency undergoes a three-level preprocessing process consistent with step S11, sequentially completing outlier removal, filtering smoothing, and standardization processing, ultimately yielding a high-time-density, high-quality salt spray monitoring dataset.

[0071] Based on the high temporal density salt spray monitoring dataset, and combined with the spatial layout information of each component of the generator set, the spatial distribution and temporal variation characteristics of salt spray concentration are iteratively updated to obtain the real-time corrected dynamic distribution law of salt spray. The update process is completely consistent with step S12, sequentially completing spatial interpolation calculation, salt spray spatial distribution map update, image segmentation region division, and spatial parameter and temporal feature extraction. The entire process of iterative correction of the dynamic distribution law of salt spray is completed every 10 minutes. Compared with the update frequency of once per hour in the basic mode, the real-time performance and accuracy of the distribution law are greatly improved, and it can promptly reflect sudden changes in the salt spray environment.

[0072] The real-time corrected dynamic distribution pattern of salt spray is synchronously fed back to the potential corrosion threat level assessment stage, completing the closed-loop control of salt spray monitoring and dynamic optimization. The execution logic after feedback is as follows: based on the real-time corrected dynamic distribution pattern of salt spray, threshold comparison, risk calculation, and threat level assessment are re-executed. The corrosion risk level is reassessed every 2 hours. The warning level and the implementation intensity of the protection plan are dynamically adjusted according to the real-time changes in salt spray distribution. If the salt spray concentration continues to rise and exceeds the predicted value by 10%, the protection measures will be upgraded and emergency sealing protection will be initiated. If the salt spray concentration drops to a safe range and is lower than the predicted value by 20%, the protection level will be gradually reduced, realizing the dynamic adaptive adjustment of the corrosion protection strategy and completing the closed-loop optimization of the entire method.

[0073] refer to Figure 2 The second embodiment provides a sensor-based salt spray corrosion early warning system for offshore platform generator sets, including: The data acquisition and mapping module is used to collect environmental datasets such as salt spray particle concentration, wind speed, humidity and temperature through a sensor network deployed around the generator set on the offshore platform, and to construct a spatial distribution map of salt spray in the corresponding area of ​​the generator set component based on the environmental dataset. The distribution pattern determination module is used to extract the spatial gradient evolution parameters, particle deposition location and windward accumulation of salt spray on the surface of each component of the generator set based on the salt spray spatial distribution map, and to determine the dynamic distribution pattern of salt spray. The corrosion risk assessment module is used to calculate the salt spray corrosion risk on the surface of each component of the generator set and determine the degree of potential corrosion threat when the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds the preset evolution threshold by integrating wind direction and humidity correlation data. The corrosion sensitivity determination module is used to retrieve the corrosion history data and material properties of the corresponding metal parts based on the degree of potential corrosion threat, combine them with salt spray exposure parameters, fit the corrosion intensity correlation pattern, and determine the corrosion sensitivity of the corresponding metal parts. The corrosion trend prediction module is used to predict the corrosion intensity change trend of the corresponding area of ​​the generator set components by combining the corrosion sensitivity, the degree of potential corrosion threat and the dynamic distribution law of salt spray. The early warning scheme generation module is used to generate salt spray corrosion early warning signals and protection schemes for corresponding parts of the generator set based on the trend of erosion intensity changes.

[0074] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A sensor-based method for early warning of salt spray corrosion of offshore platform generator sets, characterized in that, include: An environmental dataset is obtained by collecting data on salt spray particle concentration, wind speed, humidity, and temperature through a sensor network deployed around the generator set on the offshore platform. Based on the environmental dataset, a spatial distribution map of salt spray in the corresponding area of ​​the generator set component is constructed. Based on the aforementioned salt spray spatial distribution map, the spatial gradient evolution parameters, particle deposition location, and windward accumulation of salt spray on the surface of each component of the generator set are extracted to determine the dynamic distribution law of salt spray. When the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds the preset evolution threshold, the wind direction and humidity correlation data are integrated to calculate the salt spray corrosion risk on the surface of each component of the generator set and determine the degree of potential corrosion threat. To determine the degree of potential corrosion threat, historical corrosion data and material properties of the corresponding metal components are retrieved, and corrosion intensity correlation patterns are fitted in conjunction with salt spray exposure parameters to determine the corrosion sensitivity of the corresponding metal components. Based on the corrosion sensitivity, potential corrosion threat level and salt spray dynamic distribution pattern, the trend of erosion intensity change in the corresponding area of ​​the generator set component is predicted; Based on the trend of erosion intensity change, salt spray corrosion early warning signals and protection schemes are generated for the corresponding parts of the generator set.

2. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 1, characterized in that, The method involves using a sensor network deployed around the generator set on the offshore platform to collect environmental datasets, including salt spray particle concentration, wind speed, humidity, and temperature. Based on these datasets, a spatial distribution map of salt spray in the corresponding area of ​​the generator set components is constructed. By deploying a sensor network at various monitoring points around the generator set on the offshore platform, real-time environmental data such as salt spray particle concentration, wind speed, humidity and temperature are continuously and synchronously collected to obtain an environmental dataset. The environmental dataset is preprocessed by filtering to remove invalid and abnormal data generated during the collection process, resulting in a standardized environmental dataset. By combining the spatial layout information of each component of the generator set, the standardized environmental dataset is spatially registered, and a one-to-one correspondence is established between the standardized environmental data and the corresponding regions of each component of the generator set, thus obtaining a calibration environmental dataset with completed spatial registration. Based on the calibration environment dataset, a spatial distribution map of salt spray corresponding to the regions of each component of the generator set is constructed.

3. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 1, characterized in that, Based on the salt spray spatial distribution map, the spatial gradient evolution parameters, particle deposition locations, and windward accumulation of salt spray on the surfaces of various components of the generator set are extracted to determine the dynamic distribution law of salt spray, including: Based on the salt spray spatial distribution map, a preset image segmentation algorithm is used to divide the salt spray monitoring sub-regions that match the corresponding regions of each component of the generator set, and the component-level region division results are obtained. For each salt spray monitoring sub-region in the component-level region division results, the spatial gradient evolution parameters of salt spray on the surface of each component of the generator set, particle deposition location and windward accumulation amount are extracted to obtain the basic parameters of salt spray spatial distribution. By combining the continuously acquired salt spray spatial distribution map with the component-level region division results, the temporal variation characteristics of salt spray parameters in each salt spray monitoring sub-region are extracted to obtain the dynamic variation characteristics of salt spray. By integrating the basic parameters of salt spray spatial distribution and the dynamic change characteristics of salt spray, the dynamic distribution law of salt spray on the surface of each component of the generator set is determined.

4. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 1, characterized in that, When the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds a preset evolution threshold, the system integrates wind direction and humidity correlation data to calculate the salt spray corrosion risk on the surface of each component of the generator set and determines the degree of potential corrosion threat, including: The spatial gradient evolution parameters corresponding to the dynamic distribution law of salt spray are retrieved and compared with the preset evolution threshold item by item to obtain the parameter threshold comparison result. Based on the parameter threshold comparison results, abnormal component regions whose spatial gradient evolution parameters exceed the preset evolution threshold are selected, and wind direction and humidity correlation data corresponding to the abnormal component regions are retrieved simultaneously to obtain an abnormal region analysis dataset. By integrating the data from the abnormal region analysis dataset, the salt spray corrosion risk of the abnormal component region is calculated, and the salt spray corrosion risk classification result is obtained. Based on the salt spray corrosion risk classification results, the degree of potential corrosion threat to the surface of each component of the generator set is determined.

5. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 1, characterized in that, The process involves retrieving historical corrosion data and material properties of the corresponding metal components based on the level of potential corrosion threat, combining this data with salt spray exposure parameters, fitting an erosion intensity correlation pattern, and determining the corrosion sensitivity of the corresponding metal components. This includes: Locate the corresponding area of ​​the generator set component that corresponds to the potential corrosion threat level, retrieve the corrosion history data and material properties of the metal components in the corresponding area of ​​the generator set component, and construct a basic corrosion dataset; Simultaneously retrieve the salt spray exposure parameters of the corresponding areas of the generator set components and supplement them into the basic corrosion dataset to obtain a complete corrosion analysis dataset; Based on the complete corrosion analysis dataset, the correlation pattern between salt spray erosion intensity and environmental parameters and metal material properties is fitted to obtain the erosion intensity correlation fitting model. Based on the corrosion intensity correlation fitting model, the corrosion sensitivity of metal components in the corresponding area of ​​the generator set components is determined.

6. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 1, characterized in that, The method of predicting the erosion intensity trend of corresponding areas of generator unit components by combining the corrosion sensitivity, potential corrosion threat level and salt spray dynamic distribution pattern includes: Based on the aforementioned corrosion sensitivity, potential corrosion threat level, and dynamic distribution pattern of salt spray, a basic dataset for trend analysis is obtained. Using a preset time series analysis technique, feature extraction and trend fitting are performed on the time series variation data of salt spray concentration and the fluctuation data of environmental parameters in the trend analysis basic dataset to obtain the evolution trend of initial erosion intensity. Based on the corrosion sensitivity, the evolution trend of the initial erosion intensity is corrected and calibrated to obtain the calibrated time-series evolution result of salt spray erosion intensity; Based on the time-series evolution of calibrated salt spray erosion intensity, the future trend of erosion intensity in corresponding areas of each component of the generator set is predicted.

7. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 1, characterized in that, The step of generating salt spray corrosion early warning signals and protection schemes for corresponding parts of the generator set based on the trend of corrosion intensity changes includes: The future trend of corrosion intensity in the corresponding areas of each component of the generator set is retrieved. Combined with the surface coating status and local stress concentration data of the corresponding component's metal parts, the preset protection level requirements are matched to obtain the protection adaptation parameters of the target component. Based on the target component protection adaptation parameters, a protection scheme that matches the corrosion intensity level and metal material properties is extracted from the preset corrosion prevention and control parameter library. Based on the classification results of the future trend of the erosion intensity and the protection adaptation parameters of the target component, a graded salt spray corrosion early warning signal is generated for the corresponding component of the generator set.

8. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 1, characterized in that, After generating the salt spray corrosion early warning signal and protection plan for the corresponding parts of the generator set, it also includes: The salt spray corrosion early warning signal is used to adjust the frequency of data collection on the correlation between salt spray accumulation and humidity fluctuation on the windward side in real time, and continuously update the dynamic distribution pattern of salt spray.

9. The sensor-based early warning method for salt spray corrosion of offshore platform generator sets according to claim 8, characterized in that, The process of adjusting the frequency of data collection on the correlation between salt spray corrosion early warning signal and humidity fluctuation on the windward side using the sensor network to continuously update the dynamic distribution pattern of salt spray includes: Based on the classification results of the salt spray corrosion early warning signal, a sampling frequency adjustment command is sent to the sensor network to match and determine the corresponding sampling frequency range of the sensor network; According to the aforementioned collection frequency range, the collection frequency of the sensor network for the correlation data of the accumulation amount and humidity fluctuation on the windward side of the corresponding component of the generator set is adjusted in real time, and the filtering preprocessing and standardization processing of the collected data are completed simultaneously to obtain a high time-density salt spray monitoring dataset. Based on the high temporal density salt spray monitoring dataset, combined with the spatial layout information of each component of the generator set, the spatial distribution and temporal variation characteristics of salt spray concentration are iteratively updated to obtain the real-time corrected dynamic distribution law of salt spray. The real-time corrected dynamic distribution pattern of salt spray is synchronously fed back to determine the degree of potential corrosion threat.

10. A sensor-based early warning system for salt spray corrosion of offshore platform generator sets, characterized in that, include: The data acquisition and mapping module is used to collect environmental datasets such as salt spray particle concentration, wind speed, humidity and temperature through a sensor network deployed around the generator set on the offshore platform, and to construct a spatial distribution map of salt spray in the corresponding area of ​​the generator set component based on the environmental dataset. The distribution pattern determination module is used to extract the spatial gradient evolution parameters, particle deposition location and windward accumulation of salt spray on the surface of each component of the generator set based on the salt spray spatial distribution map, and to determine the dynamic distribution pattern of salt spray. The corrosion risk assessment module is used to calculate the salt spray corrosion risk on the surface of each component of the generator set and determine the degree of potential corrosion threat when the spatial gradient evolution parameter of the dynamic distribution law of salt spray exceeds the preset evolution threshold by integrating wind direction and humidity correlation data. The corrosion sensitivity determination module is used to retrieve the corrosion history data and material properties of the corresponding metal parts based on the degree of potential corrosion threat, combine them with salt spray exposure parameters, fit the corrosion intensity correlation pattern, and determine the corrosion sensitivity of the corresponding metal parts. The corrosion trend prediction module is used to predict the corrosion intensity change trend of the corresponding area of ​​the generator set components by combining the corrosion sensitivity, the degree of potential corrosion threat and the dynamic distribution law of salt spray. The early warning scheme generation module is used to generate salt spray corrosion early warning signals and protection schemes for corresponding parts of the generator set based on the trend of corrosion intensity changes.

Images