A steel structure building anticorrosive coating life prediction method and system

Abstract

The application relates to the technical field of building engineering and discloses a steel structure building anticorrosive coating life prediction method and system, which comprises the following steps: acquiring multi-source environment data of different parts of a steel structure building, extracting environment fluctuation characteristics and grouping similar areas by using time series decomposition and clustering analysis; fitting the relationship between environment factors and coating aging rate by using a regression model, determining the aging acceleration coefficient and rate estimation value of a specific area; generating a preliminary life distribution map according to the matching of historical performance records; generating multi-scenario environment evolution paths and coating degradation trajectories by using Monte Carlo simulation, determining a comprehensive life prediction interval; dynamically adjusting prediction parameters based on the comprehensive life prediction interval to obtain an optimized life distribution description; generating a visual mapping, identifying and sorting high-risk parts, and outputting a maintenance decision report. The application solves the problem of insufficient coating life prediction precision under complex environments and realizes accurate and dynamic life prediction and risk positioning.

Description

Technical Field

[0001] This application relates to the field of building engineering technology, and in particular to a method and system for predicting the lifespan of anti-corrosion coatings for steel structure buildings. Background Technology

[0002] Steel structures, due to their high strength and excellent construction performance, are widely used in large industrial plants, stadiums, transportation hubs, and super high-rise buildings. Anti-corrosion coatings are the first and most critical line of defense in protecting steel structures from environmental corrosion and extending their service life. Once the coating fails, it will directly lead to steel corrosion, significantly increasing maintenance costs and seriously threatening structural safety. Therefore, accurately predicting the remaining lifespan of anti-corrosion coatings is crucial for developing scientific and economical maintenance strategies and ensuring structural safety, especially given the ever-expanding scale of buildings and the increasingly diverse usage environments. This has become a core technical challenge that urgently needs to be solved in this field.

[0003] Currently, traditional methods for predicting coating life mostly rely on empirical formulas derived from accelerated aging test data in laboratories, or static models based on limited single environmental parameters (such as average annual temperature and annual precipitation). These methods have significant limitations: First, they are difficult to reflect the complex interactions and synergistic effects of multiple environmental factors such as temperature, humidity, and ultraviolet radiation in real service environments; second, steel structure buildings are massive in scale, and the microclimate environments faced by different parts (such as roofs, sunny sides, and shaded sides) vary greatly, while existing methods generally ignore this spatial variability and adopt a "one-size-fits-all" prediction model, resulting in prediction results that are seriously inconsistent with the actual aging conditions of specific parts, and exhibiting large errors when facing complex and dynamic environmental conditions.

[0004] Although advancements in technology have led to the increasing application of prediction methods based on multi-source environmental data, and advanced data processing techniques such as Monte Carlo simulation and Bayesian updates have provided new approaches to addressing uncertainties in prediction, existing solutions still face significant challenges in practical applications. The core issues are: first, the lack of effective means to efficiently integrate and extract features from multi-dimensional, time-varying environmental data from different parts of a building; second, the difficulty in quantifying the dynamic fluctuations of environmental parameters and their impact on the spatial non-uniformity of coating aging; and finally, the inability to construct a dynamically updated, highly robust lifespan prediction model.

[0005] Therefore, there is an urgent need in this field for an innovative technical solution to overcome the above-mentioned shortcomings. This solution must be able to: 1) precisely perceive and integrate real-time multi-source environmental data from different parts of the steel structure building; 2) quantitatively analyze the dynamic fluctuation characteristics of environmental parameters and their impact on the spatial non-uniformity of coating aging; and 3) construct a dynamically updated, highly robust life prediction model, thereby providing a reliable basis for achieving accurate predictive maintenance. Summary of the Invention

[0006] This application provides a method and system for predicting the lifespan of anti-corrosion coatings for steel structures, aiming to solve the problems of large errors and low accuracy in the lifespan prediction of anti-corrosion coatings for steel structures based on single environmental factors in existing technologies. By employing multi-source environmental data and advanced statistical models, this invention can provide more accurate and dynamic coating lifespan predictions, and adjust prediction parameters in real time according to actual environmental conditions, thereby improving the reliability and practicality of the predictions.

[0007] In a first aspect, this application provides a method for predicting the service life of anti-corrosion coatings for steel structure buildings, the method comprising: Step S1: Obtain real-time environmental data from different parts of the steel structure building and construct a multi-source environmental dataset; Step S2: Based on the multi-source environmental dataset, analyze the periodic fluctuation pattern of environmental parameters and extract the fluctuation feature vector representing the pattern. Step S3: Based on the fluctuation feature vector, perform cluster analysis on each part of the steel structure building to obtain the regional classification results; Step S4: For each region in the region classification results, fit the relationship between environmental factors and coating aging rate in that region, and then determine the estimated value of coating aging rate in that region. Step S5: Based on the estimated coating aging rate, query and match historical coating performance records, and generate a preliminary lifetime distribution map based on the matching results; Step S6: Generate multi-scenario environmental evolution paths based on the preliminary lifetime distribution map, and calculate the coating degradation trajectory under each path, thereby determining the comprehensive prediction range of coating lifetime. Step S7: Perform statistical analysis on the comprehensive prediction interval to determine whether there are any abnormal deviations. If so, adjust the prediction parameters to obtain the optimized lifetime distribution description. Step S8: Based on the optimized lifespan distribution description, generate a visual mapping covering all aspects of the steel structure building, identify high-risk areas with abnormal lifespan, prioritize them, and output a maintenance decision report containing the distribution and priority ranking of high-risk areas.

[0008] Secondly, this application provides a life prediction system for anti-corrosion coatings on steel structures, the system comprising: The data acquisition module is used to obtain real-time environmental data from different parts of the steel structure building and construct a multi-source environmental dataset; The feature extraction module is used to analyze the periodic fluctuation patterns of environmental parameters based on multi-source environmental datasets and extract fluctuation feature vectors that characterize the patterns. The region classification module is used to perform cluster analysis on various parts of the steel structure building based on the fluctuation feature vector to obtain the region classification results; The aging estimation module is used to fit the relationship between environmental factors and coating aging rate in each region for the region classification results, and then determine the estimated value of coating aging rate in that region. The lifetime mapping module is used to query and match historical coating performance records based on the estimated coating aging rate, and generate a preliminary lifetime distribution map based on the matching results. The interval prediction module is used to generate multi-scenario environmental evolution paths based on the preliminary lifetime distribution map, and calculate the coating degradation trajectory under each path, thereby determining the comprehensive prediction interval of the coating lifetime. The parameter adjustment module is used to perform statistical analysis on the comprehensive prediction interval to determine whether there are any abnormal deviations. If so, the prediction parameters are adjusted to obtain an optimized lifetime distribution description. The risk visualization module is used to generate a visual map covering all aspects of the steel structure building based on the optimized life distribution description, identify high-risk parts with abnormal lifespan, prioritize them, and output a maintenance decision report containing the distribution and priority ranking of high-risk parts.

[0009] Compared with the prior art, the beneficial effects of the technical solution of this application are at least as follows: 1. By deploying a distributed sensor array to construct a multi-source, real-time environmental dataset, and combining temporal decomposition and cluster analysis, the microclimate differences and periodic fluctuations of different parts of steel structure buildings can be accurately captured, ensuring the quality of model input from the data source. Employing regionalized aging rate estimation and multi-scenario Monte Carlo simulations, the spatiotemporal dynamics and uncertainties of environmental factors are fully considered, making the prediction results closer to actual aging conditions. This significantly reduces the errors of traditional prediction methods and significantly improves the accuracy and precision of predicting the lifespan of anti-corrosion coatings for steel structure buildings.

[0010] 2. A Bayesian update mechanism is introduced. When anomalies are detected in the prediction interval, the latest collected environmental data is dynamically incorporated, and the aging rate and lifetime distribution parameters are adaptively adjusted. This closed-loop feedback mechanism enables the model to continuously learn and optimize, effectively responding to sudden environmental changes and long-term climate trends, and greatly enhancing the adaptability and robustness of the prediction model.

[0011] 3. Through optimized lifespan distribution description and visualization mapping, high-risk areas of abnormal lifespan in steel structure buildings can be accurately identified and prioritized according to risk level, providing a scientific basis for structural maintenance. The final comprehensive maintenance decision report not only includes lifespan prediction data but also displays the distribution and priority ranking of high-risk areas, thus effectively guiding the maintenance of building structures.

[0012] 4. Transforming abstract predictive data into intuitive and actionable engineering guidelines helps managers accurately pinpoint the most vulnerable areas and rationally allocate maintenance resources and plan maintenance schedules based on risk levels, promoting a shift from "passive repair" to "proactive prediction and precise maintenance." By optimizing maintenance strategies, the overall lifespan of buildings can be effectively extended, and the total lifespan maintenance cost can be reduced.

[0013] 5. The multi-source data-driven prediction framework and methodology constructed in this invention are not limited to specific types of steel structure buildings. Its core processes and technologies, after adaptive adjustments, can be further extended to other steel structure facilities such as bridges and storage tanks, and even to the prediction of the lifespan of building coatings made of different materials, demonstrating broad application prospects. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart illustrating a method for predicting the lifespan of anti-corrosion coatings for steel structures, as described in this application. Figure 2 This is a structural schematic diagram of a steel structure building anti-corrosion coating life prediction system according to this application. Detailed Implementation

[0016] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] For ease of understanding, the specific process of the embodiments of this application is described below. Figure 1 The diagram shows a flowchart of a method for predicting the lifespan of anti-corrosion coatings for steel structures provided by this invention. The flowchart specifically includes the following steps: Step S1: Obtain real-time environmental data from different parts of the steel structure building and construct a multi-source environmental dataset.

[0018] In one specific embodiment, the process of performing step S1 may specifically include the following steps: Real-time environmental data is collected by a distributed sensor array deployed at key exposed parts of steel structure buildings. The real-time environmental data includes temperature and humidity values ​​and ultraviolet intensity indicators. Real-time environmental data is preprocessed, including data cleaning and format standardization. The preprocessed data is grouped according to the collection location to form an independent environmental parameter sequence for each location, and an initial multi-source environmental dataset is generated based on the environmental parameter sequence. The initial multi-source environmental dataset is timestamped to generate a unified time series multi-source environmental dataset.

[0019] Specifically, sensor arrays are deployed at key exposed areas of steel-structured buildings (such as roofs, walls, and foundations) to acquire environmental data including temperature, humidity, and UV intensity. This environmental data, as a key factor influencing the aging of building coatings, reflects the environmental stress experienced by different parts of the building. The environmental data collection covers the exposed areas of the building, ensuring representative data is obtained from multiple perspectives. The data acquisition devices transmit the collected data to a central processing unit via a wireless network, forming an initial data stream containing timestamps and monitoring values.

[0020] Due to environmental interference and equipment differences, the raw data stream contains noise and missing data. Therefore, preprocessing of the collected real-time environmental data is necessary to ensure data quality. Data cleaning steps include identifying and removing missing or outliers. For example, if a sensor malfunctions, missing data or measurement anomalies will affect the accuracy of the entire dataset; therefore, outlier detection and imputation algorithms are needed to handle missing data. In this case, methods such as median imputation are used to supplement temperature and humidity data to ensure data continuity. The cleaned data will be converted to a standardized format, such as CSV. This standardization process ensures seamless integration of data from different sensors and locations.

[0021] After data cleaning and format standardization, the environmental data collected by each sensor is grouped by location, generating an environmental parameter sequence for each location. Each location's sequence includes time-series temperature and humidity values ​​and ultraviolet (UV) intensity indicators; these data sequences represent the environmental characteristics of each location over time. This step effectively integrates environmental data from different locations to construct an initial multi-source environmental dataset. For example, the temperature, humidity, and UV data for the bridge pier location are organized into three time series, forming the parameter set for that location; the parameter sets from all locations together constitute the initial multi-source environmental dataset.

[0022] The initial multi-source environmental dataset reflects the spatial heterogeneity of the structure, but the sequences suffer from asynchronous acquisition times and inconsistent sampling frequencies. Directly using it for time series analysis would introduce bias. Therefore, the initial dataset needs to be timestamped to generate a unified time series. The alignment process first checks the timestamps of each sequence to identify time deviations caused by transmission delays or differences in triggering mechanisms. For sequences with deviations on the order of seconds, linear interpolation is used for alignment. For cases with inconsistent sampling frequencies, such as an ultraviolet sensor acquiring data once per minute while a temperature and humidity sensor acquires data once every five minutes, the ultraviolet sequence is first downsampled to a five-minute interval, and then a dynamic time warping algorithm is applied to calculate the similarity between sequences and align the time axes. After unifying the time series, all the data sequences form a consistent multi-source environmental dataset.

[0023] Step S2: Based on the multi-source environmental dataset, analyze the periodic fluctuation pattern of environmental parameters and extract the fluctuation feature vector that represents the pattern.

[0024] In one specific embodiment, the process of performing step S2 may specifically include the following steps: Based on a multi-source environmental dataset, a time-series decomposition algorithm is used to process the data sequences, decomposing each data sequence into trend components, seasonal components, and residual components. Based on the time series decomposition results, periodic components are extracted from the data sequence to obtain periodic fluctuation patterns; Based on the periodic fluctuation pattern, feature parameters for quantifying the pattern are extracted. The feature parameters include fluctuation period and fluctuation amplitude. The feature parameters are used as vector elements to construct the fluctuation feature vector.

[0025] Specifically, time-series decomposition algorithms include trend decomposition and seasonal decomposition, which can decompose each data sequence into three components: a trend component, a seasonal component, and a residual component. The trend component captures the long-term trend of change in the data; the seasonal component reflects the periodic fluctuations in the data; and the residual component is the noise and anomalous fluctuations remaining after removing the periodic fluctuations, used to accommodate random noise and irregular variations in the sequence. Through this process, periodic fluctuation patterns in the data can be effectively identified and separated. Taking the annual temperature data of the roof of a large steel structure factory as an example, after this decomposition, its trend component may reflect the slow rise of the annual average temperature, while the seasonal component clearly shows the daily temperature fluctuation pattern with a 24-hour cycle, thus separating the periodic patterns implicit in environmental parameters from long-term trends and random disturbances.

[0026] After completing the time-series decomposition, the seasonal component is extracted from the decomposition results as the periodic fluctuation pattern of the environmental parameter. This component directly reflects the repetitive variation patterns of environmental parameters such as temperature and humidity on daily and seasonal time scales. For example, in the roof area of ​​a steel structure building, the temperature may experience significant fluctuations within a 24-hour day, while ultraviolet radiation intensity may exhibit annual periodic fluctuations. By extracting the periodic component from the data sequence, these patterns can be accurately grasped, thus providing a quantitative basis for the impact of environmental changes on the coating aging rate.

[0027] After extracting the periodic fluctuation pattern, its intensity and variation are further quantified using characteristic parameters. These parameters include the fluctuation period and amplitude. The fluctuation period is determined by performing spectral analysis on the periodic component, identifying the period length corresponding to the frequency with the strongest energy. For example, spectral analysis of the daily fluctuation component of roof temperature determines that its dominant period is 24 hours. The fluctuation amplitude is characterized by calculating the average difference between the peak and trough values ​​of the periodic component within a complete cycle; for example, the calculated daily fluctuation amplitude of the roof temperature is 12 degrees Celsius. These characteristic parameters allow the periodic fluctuation pattern to not only be visually presented but also accurately reflect its potential impact on the lifespan of the anti-corrosion coating.

[0028] Feature parameters extracted from different environmental parameters at the same location are combined as elements to construct a multidimensional fluctuation feature vector. For example, the 24-hour temperature cycle, temperature range of 12 degrees Celsius, 24-hour humidity cycle, humidity range of 15% for the aforementioned roof location, and the 365-day annual UV intensity cycle and UV intensity of 40 units are combined into a vector [24, 12, 24, 15, 365, 40]. This vector, in a compact mathematical form, comprehensively represents the complex periodic load of multiple environmental factors faced by this location, helping to more comprehensively reflect the environmental fluctuation characteristics of different parts of the building. In the context of multi-source environmental datasets, the construction of the fluctuation feature vector can integrate the periodic variation characteristics of multiple environmental parameters such as temperature, humidity, and UV intensity.

[0029] This invention not only enables in-depth analysis of environmental fluctuations in various parts of steel structure buildings but also effectively improves the accuracy of coating aging prediction. The extraction of periodic fluctuation patterns and the construction of feature vectors ensure the multi-dimensionality and comprehensiveness of environmental parameter changes, thus providing strong support for subsequent fitting of aging acceleration coefficients and generation of lifetime distribution maps. This process solves the problem of capturing complex environmental change patterns, enhances the adaptability and accuracy of the prediction model, and effectively improves the reliability of coating lifetime prediction.

[0030] Step S3: Based on the fluctuation feature vector, perform cluster analysis on each part of the steel structure building to obtain the regional classification results.

[0031] In one specific embodiment, the process of performing step S3 may specifically include the following steps: Clustering algorithms are used to cluster the wave feature vectors, and parts with similar feature vectors are grouped into the same cluster, thereby grouping different parts of the steel structure building into several similar environmental regions; The degree of microclimate difference between different similar environmental regions is obtained by calculating the Euclidean distance between the cluster center vectors of different clusters. Regional classification results are generated based on the degree of microclimate differences, and the regional classification results characterize the environmental similarity of each location.

[0032] Specifically, the fluctuation feature vectors generated in the previous step, representing the periodic fluctuation patterns of the environment in each part, are used as input. These vectors integrate the periodic and amplitude information of parameters such as temperature, humidity, and ultraviolet intensity. To eliminate the influence of different parameter dimensions on distance calculations, all fluctuation feature vectors are first normalized to ensure that the values ​​of each dimension are on the same scale. The processed vectors are used as input data for the clustering algorithm. The K-means clustering algorithm is used to group this vector set. This algorithm determines multiple cluster centers through iterative calculations and assigns the feature vector of each part to the cluster center with the closest Euclidean distance, thereby dividing the entire steel structure building into several clusters in the environmental feature space. Each cluster corresponds to a similar environmental area, and the areas within the cluster have highly similar environmental periodic fluctuation patterns. For example, when analyzing a large steel structure factory building, the clustering results may show that the feature vectors of the roof and the upper exterior wall are grouped into the same cluster, while the lower supporting structure and the interior area form separate independent clusters. This intuitively reflects the differences in environmental loads caused by different building heights and orientations.

[0033] After clustering, it is necessary to quantify the degree of microclimate difference between different similar environmental regions. This degree of difference is obtained by calculating the Euclidean distance between the cluster center vectors of each cluster. The cluster center vector represents the mean of the environmental characteristics of all parts within a cluster, and the Euclidean distance between them directly measures the overall difference in environmental fluctuation patterns among different regions. Taking this factory building as an example, calculating the center vector distance between the roof area cluster and the lower supporting structure cluster may yield a large value, indicating a significant difference in their microclimate environments; while calculating the center vector distance between two adjacent exterior wall clusters may yield a smaller value, indicating that their environmental conditions are relatively similar.

[0034] The final regional classification result is generated based on the calculated degree of microclimate differences. This result reflects the environmental similarity of each region, including not only the cluster label of each region but also quantitative data on inter-cluster differences. If the Euclidean distance between two clusters is lower than a preset similarity threshold (e.g., 0.5), the two regions are considered to have high similarity in environmental fluctuations, and they are managed as a larger unified region in the maintenance strategy. Conversely, if the distance is significantly higher than the threshold, it is identified as a high-variability region. A high Euclidean distance indicates that the environmental fluctuation pattern of this region differs significantly from other regions, meaning that it is subject to greater changes in environmental factors. The microclimate changes in these regions may be more drastic or unstable, potentially having a greater impact on the coating aging process; therefore, they are identified as high-variability regions. For example, one region might exhibit vector characteristics of "high temperature, strong ultraviolet radiation, and low humidity," while another region might have characteristics of "low temperature, high humidity, and weak ultraviolet radiation." These two drastically different environmental combinations will trigger different dominant aging mechanisms (the former is dominated by photoaging, while the latter is dominated by damp heat aging), resulting in significant differences in their aging rates. The system identifies these regions as two highly variable areas. In the next step, environmental data for each region will be acquired separately to establish two different aging rate prediction models. For regions with similar environments, merging and sharing models reduces the number of unnecessary models, improves computational efficiency, avoids overfitting, and ensures prediction accuracy. For regions with vastly different environments, independent modeling ensures model specificity and accuracy, resolving the bias problem caused by "one-size-fits-all" predictions. This method can effectively identify areas with high environmental variability and potential risks in steel structure buildings, ensuring optimal allocation of maintenance resources and improving the long-term reliability and durability of building structures.

[0035] Cluster analysis, by accurately quantifying microclimate differences, can effectively identify and classify the environmental characteristics of different parts of steel structure buildings, providing a scientific basis for subsequent coating aging prediction. In practical applications, this technology can accurately identify the patterns of environmental fluctuations between regions, improve the accuracy of lifespan prediction, thereby optimizing maintenance plans, reducing unnecessary resource waste, and ensuring that high-risk areas are prioritized. This technique effectively addresses the problem of traditional methods failing to capture complex environmental differences, significantly improving the accuracy of anti-corrosion coating lifespan prediction and maintenance efficiency for steel structure buildings.

[0036] Step S4: For each region in the region classification results, fit the relationship between environmental factors and coating aging rate in that region, and then determine the estimated value of coating aging rate in that region.

[0037] In one specific embodiment, the process of performing step S4 may specifically include the following steps: For each similar environmental region in the regional classification results, obtain the environmental factor data for that region; Based on environmental factor data, a multiple linear regression model was used to fit the relationship between environmental factors and coating aging rate, and the regression coefficients corresponding to each environmental factor were obtained. The aging acceleration coefficient for this region is obtained by weighted summation of the regression coefficients. The product of the aging acceleration factor and the preset reference rate is used as the estimated value of the coating aging rate.

[0038] Specifically, for each classified similar environmental area, an independent aging rate modeling process is executed. Environmental factor data for that area is extracted. This data comes from temperature, humidity, and UV intensity values ​​collected by sensors deployed in various locations within the area, and has been preprocessed and timestamped as described in the previous steps, comprehensively reflecting the environmental characteristics of the area. Next, a multiple linear regression model is used to fit the relationship between environmental factors and the coating aging rate. The independent variables of the regression model are temperature, humidity, and UV intensity, while the dependent variable is the coating aging rate. During this process, regression coefficients for each environmental factor are calculated through regression analysis, characterizing the degree of influence of each environmental factor on the coating aging rate.

[0039] In the regression analysis, all regression coefficients were estimated using the least squares method to ensure the model's fitting accuracy. These regression coefficients quantify how different environmental factors, such as temperature, humidity, and UV intensity, affect the coating's aging process under varying climatic conditions. Next, these regression coefficients are weighted and summed to obtain a comprehensive aging acceleration coefficient. The weighted summation method assigns weights to each environmental factor based on historical data or domain knowledge. These weights reflect the relative importance of different environmental factors to coating aging, ultimately forming the aging acceleration coefficient. This aging acceleration coefficient is a comprehensive, dimensionless value that characterizes the accelerating effect of the overall environmental conditions in a specific area relative to standard conditions on coating aging. For example, a calculated aging acceleration coefficient of 1.15 for this factory area indicates that the environmental conditions in this area cause the coating aging rate to be 1.15 times that of standard conditions. This acceleration coefficient allows for a better quantification and reflection of the comprehensive impact of environmental factors on the coating's aging rate.

[0040] Finally, based on the obtained aging acceleration factor, it is multiplied by a preset baseline aging rate to obtain an estimated coating aging rate for that area. The baseline rate, as a reference value, can be derived from test data under standard laboratory conditions or long-term observation statistics. After adjustment with the aging acceleration factor, it can accurately predict the coating aging rate under actual environmental conditions.

[0041] For example, if environmental factor data (such as temperature, humidity, and UV intensity) for a certain area, after being fitted by a regression model, yields a temperature coefficient of 0.02, a humidity coefficient of 0.01, and a UV intensity coefficient of 0.005, then the aging acceleration coefficient for that area is 0.02 × weight 1 + 0.01 × weight 2 + 0.005 × weight 3. Assuming a baseline aging rate of 0.1 units per year, the coating aging rate for that area is 0.1 × acceleration coefficient, yielding a specific rate estimate. This estimate can be further used to predict the coating's lifespan, providing decision support for building maintenance and coating renewal.

[0042] This technical solution enables the aging rate of different regions to be quantified based on the environmental characteristics and actual aging of the coating in different areas, thereby improving the accuracy of coating life prediction. It provides an effective basis for the anti-corrosion management of steel structure buildings and ensures that coating maintenance can be optimized in a targeted manner according to the differentiated needs of the regional environment.

[0043] Step S5: Based on the estimated coating aging rate, query and match historical coating performance records, and generate a preliminary lifetime distribution map based on the matching results.

[0044] In one specific embodiment, the process of performing step S5 may specifically include the following steps: Using the estimated coating aging rate as the query condition, candidate historical records are retrieved from the historical coating performance database. The candidate historical records include coating aging data. Based on similarity calculation, the matching degree between the estimated coating aging rate and the candidate historical records is obtained, and historical records with a matching degree greater than the preset matching threshold are selected as valid matching records. Extract the actual coating lifetime data corresponding to all valid matching records, and generate a preliminary lifetime distribution map based on this. The preliminary lifetime distribution map characterizes the lifetime distribution of the coating under different environments. A smoothing algorithm is used to smooth the initial lifetime distribution map, generating a continuous lifetime distribution curve.

[0045] Specifically, the estimated coating aging rate is used as a query condition to retrieve relevant candidate historical records from a historical coating performance database containing a large number of historical engineering cases. These historical records include coating aging data. The historical coating performance records stored in the database reflect the aging characteristics of coatings under different environmental factors, such as temperature, humidity, and UV intensity, which are closely related to coating durability. By using the estimated coating aging rate, matching historical data can be accurately obtained.

[0046] After retrieving historical coating performance records, a similarity calculation is used to evaluate the matching degree between the estimated coating aging rate and the historical records. This matching process typically employs the cosine similarity method, calculating the cosine of the angle between the query conditions and the historical records in the feature space to quantify the similarity. Specifically, both the estimated coating aging rate and the aging data in the historical records can be converted into vector form. By calculating the cosine similarity value of these vectors, their similarity in the multidimensional environmental factor space is determined. By measuring the consistency of vector directions to evaluate the similarity of aging patterns, the limitations of comparing single numerical values ​​are effectively avoided. A matching degree threshold is set (e.g., 0.8). If the similarity exceeds the preset matching threshold, the historical record is considered to have a high matching degree with the estimated coating aging rate and is thus considered a valid matching record. This similarity-based matching method can effectively improve the matching accuracy of historical data, ensuring that the selected historical records can truly reflect the coating aging characteristics under similar environments.

[0047] Based on the selected valid matching records, the actual coating lifespan data corresponding to these records will be extracted. This lifespan data comes from the real time span of the coating from its initial use to performance failure or reaching the maintenance threshold in historical cases. Lifespan data includes the coating's durability performance under different environments, such as the actual lifespan of the coating under specific temperature, humidity, and UV intensity conditions. These discrete lifespan data points are organized with their corresponding key environmental factors as references to generate a preliminary lifespan distribution map. The horizontal axis represents environmental parameters, and the vertical axis represents coating lifespan. The graph clearly shows the lifespan variation of the coating under different environmental conditions. By aggregating these matching records and lifespan data, a preliminary lifespan distribution map can be obtained. In the map, the coating lifespan may be significantly shortened under high humidity and high UV conditions, while the coating lifespan is relatively longer under lower humidity or UV intensity conditions.

[0048] Since the initial lifetime distribution map consists of discrete data points, it needs to be smoothed to generate a continuous lifetime distribution curve in order to form a continuous and stable probability distribution for subsequent steps. A moving average algorithm is used to smooth the distribution map. This algorithm generates a continuous lifetime distribution curve by calculating the average value of the values ​​in the neighborhood of each data point. This curve eliminates the interference of individual outliers and clearly reveals the overall trend of coating lifetime changes with environmental conditions, providing a statistically consistent input probability model for subsequent Monte Carlo simulations.

[0049] In steel structure buildings, this technology allows for the acquisition of historical coating performance records based on estimated coating aging rates, enabling accurate matching and generation of preliminary lifespan distribution maps. By analyzing coating lifespan under different environments, the aging behavior of coatings under varying microclimates can be predicted, providing a precise basis for coating maintenance. Particularly in environments with high humidity and high UV intensity, it can effectively identify high-risk areas of the coating, guiding subsequent coating maintenance strategies. This method addresses current limitations in coating lifespan prediction, such as inaccurate models and insufficient consideration of environmental differences. By matching historical data and generating lifespan distribution maps, it provides a more comprehensive and scientific solution for predicting coating aging.

[0050] Step S6: Generate multi-scenario environmental evolution paths based on the preliminary lifetime distribution map, and calculate the coating degradation trajectory under each path to determine the comprehensive prediction range of coating lifetime.

[0051] In one specific embodiment, the process of performing step S6 may specifically include the following steps: Extract the probability distribution of environmental parameters from the preliminary lifetime distribution map; Based on probability distribution, Monte Carlo simulation method is used to perform multiple random samplings to generate multiple evolution paths of multi-scenario environment in future time periods; Each environmental evolution path is input into a preset aging rate model, and the cumulative degradation of the coating is calculated step by step over time to obtain the coating degradation trajectory corresponding to each environmental evolution path. For each coating degradation trajectory, determine the coating lifetime endpoint corresponding to when it reaches a preset failure threshold; Statistical analysis was performed on all coating lifetime endpoint data to fit their probability distribution and generate a probability distribution description including the lifetime mean and standard deviation. Based on the probability distribution description, its confidence interval at a specified confidence level is calculated, and this confidence interval is used as the comprehensive prediction interval for coating lifetime.

[0052] Specifically, the probability distribution characteristics of environmental parameters are extracted from the preliminary lifetime distribution map, including statistical features such as the mean and variance of variables like temperature, humidity, and UV intensity. These statistical features provide the basic data for simulating environmental changes and help construct probability distribution models for different environmental paths. Based on these probability distribution characteristics, a Monte Carlo simulation method is used for large-scale random sampling. This process generates a large number of possible future environmental sequences by repeatedly drawing random values ​​from the probability distributions of each environmental parameter. Each sampling produces a set of environmental parameter values, and multiple samplings form multiple environmental evolution paths covering the next ten years or longer. These paths together constitute a comprehensive enumeration of future environmental uncertainties. For example, some paths may simulate extreme years of high temperature and high humidity, while others simulate mild climatic conditions. These paths provide multiple environmental scenarios for calculating the coating degradation trajectory.

[0053] After generating the environmental evolution paths, each path is sequentially input into an aging rate model based on the Arrhenius equation. This physicochemical model describes the accelerating effect of environmental factors such as temperature and humidity on the aging rate of the coating material. The cumulative degradation of the coating under the specific environmental sequence is calculated step-by-step along each environmental path, resulting in a corresponding coating degradation trajectory. This trajectory quantitatively describes the decay of coating performance indicators over time. When the cumulative degradation of a certain degradation trajectory reaches a preset failure threshold, that time point is recorded as the coating lifetime endpoint for that path. The failure threshold is the critical state at which the coating becomes unusable in practical applications. These lifetime endpoints provide a variety of possible results for calculating the coating lifetime.

[0054] After simulating all paths, a lifetime dataset is compiled from all coating lifetime endpoints, and statistical analysis is performed on this dataset. The dataset is processed using a probability distribution fitting method to determine the optimal fitted distribution type and calculate its parameters, generating a probability distribution description containing the lifetime mean and standard deviation. This description fully characterizes the uncertainty features of lifetime prediction. Based on this, confidence intervals at a specified confidence level are calculated using this probability distribution description, and this statistical interval serves as the comprehensive prediction interval for coating lifetime. For example, based on the fitted normal distribution and its parameters, the upper and lower limits of lifetime at a 95% confidence level can be calculated to be eleven to sixteen years. This interval is the comprehensive lifetime prediction interval, which quantifies the possible distribution of coating lifetime.

[0055] For example, for a specific part of a steel structure building, the Monte Carlo simulation method was used to simulate the effects of different temperature and humidity combinations on the coating, generating multiple evolution paths. The degradation trajectory was calculated using an aging rate model, ultimately yielding a comprehensive prediction range for the coating's lifespan. If the prediction range is 8 to 12 years and the standard deviation is small, it indicates that the coating's lifespan is relatively stable during this period; if the standard deviation is large, it indicates that the coating's lifespan has high uncertainty, requiring further optimization of the maintenance plan.

[0056] The technical solution of this invention can effectively solve the uncertainty problem in coating life prediction. By simulating environmental changes and combining actual data, it provides more accurate life prediction, improves the robustness and accuracy of the prediction, and provides a reliable basis for maintenance decisions of steel structure buildings.

[0057] Step S7: Perform statistical analysis on the comprehensive prediction interval to determine if there are any abnormal deviations. If so, adjust the prediction parameters to obtain the optimized lifetime distribution description.

[0058] In one specific embodiment, the process of performing step S7 may specifically include the following steps: Step 71: Obtain the probability distribution description and extract the standard deviation. If the standard deviation exceeds the preset standard value, it is determined that there is an abnormal deviation. Then, the newly collected real-time environmental data is obtained. Step 72: Use the Bayesian update method to incorporate newly collected real-time environmental data and adjust the prediction parameters, including aging rate and lifetime distribution parameters. Step 73: Correct the preliminary lifetime distribution map based on the adjusted prediction parameters and regenerate the lifetime distribution description; Step 74: Perform an analysis of variance on the regenerated lifetime distribution description to evaluate its stability. If it is stable, the regenerated lifetime distribution description is used as the optimized lifetime distribution description. If it is unstable, obtain more real-time environmental data and then return to step 72 to continue iterative updates until the analysis of variance results meet the stability requirements or the maximum number of iterations is reached.

[0059] Specifically, the probability distribution description within the comprehensive prediction interval is obtained, and the standard deviation is extracted as an uncertainty quantification index. If the standard deviation exceeds a preset standard value based on engineering experience, it is determined that the initial prediction has abnormal biases caused by sudden environmental changes or model mismatch, indicating that the current lifetime prediction has significant uncertainty. At this point, to ensure the accuracy of the prediction results, an adaptive update process is triggered to acquire more newly collected real-time environmental data. This data, such as temperature, humidity, and ultraviolet radiation, is monitored in real time through distributed sensor arrays or other means to supplement the prediction results. The introduction of this new data is a key step in adjusting the prediction parameters.

[0060] A Bayesian update method is employed to integrate newly acquired environmental data into the prediction system. This method uses the existing probability distribution as the prior distribution and the likelihood function constructed based on the new data and the aging rate model as evidence. Through probabilistic inference, the posterior distribution is obtained, and the aging rate parameters and related lifetime distribution parameters are then updated. The aging rate represents the speed at which the coating performance degrades, while the lifetime distribution parameters describe the entire process from the initial state to failure. This update mechanism allows the model to incorporate the latest environmental information, ensuring high adaptability to sudden environmental changes. For example, when continuous high temperature and humidity weather is detected, the system automatically increases the weight of the damp heat aging effect, making the predicted parameters more closely reflect the actual environmental conditions.

[0061] The adjusted prediction parameters are used to revise the initial lifetime distribution map. The new lifetime distribution description will be generated based on the updated aging rate and lifetime distribution parameters, thus better reflecting the lifetime prediction of the coating under new environmental conditions. In this way, the adjusted model can quickly respond to environmental changes, improve the reliability of predictions, and avoid prediction biases caused by sudden environmental changes.

[0062] Performing analysis of variance (ANOVA) on the regenerated lifetime distribution description is a crucial step in assessing its stability. By calculating the variance of the new lifetime distribution, its stability can be measured. If the variance is below a preset stability threshold, the updated distribution is considered stable and can be used as an optimized lifetime distribution description. If the variance remains high, it indicates significant uncertainty in the current prediction. In this case, it is necessary to return to the data acquisition step for further iterative updates, continuously refining the model by introducing more field data until the ANOVA results meet the stability requirements or the maximum number of iterations is reached.

[0063] For example, in predicting the coating life of steel structure buildings, if the initial prediction shows a lifespan of 10 to 15 years, but the actual ultraviolet radiation intensity in the environment suddenly increases, and the standard deviation exceeds the preset value, it indicates that the accuracy of the original prediction has been affected. In this case, by acquiring new data and updating the aging rate using a Bayesian method, the adjusted lifespan distribution will more accurately reflect these environmental changes, ultimately narrowing the lifespan range to 11 to 14 years. Through analysis of variance, if the variance of the new distribution is less than 1.0 and its stability is improved compared to historical records, the prediction results can be confirmed as more reliable, providing a scientific basis for maintenance decisions.

[0064] This technique allows for dynamic adjustment of the prediction model, ensuring timely optimization of the lifespan distribution description in the face of environmental changes and data uncertainties. This improves prediction accuracy and avoids maintenance delays caused by prediction errors. This method is particularly suitable for equipment such as steel structures that are used for extended periods and exposed to variable environments, significantly improving building maintenance efficiency and extending their service life.

[0065] Step S8: Based on the optimized lifespan distribution description, generate a visual mapping covering all aspects of the steel structure building, identify high-risk areas with abnormal lifespan, prioritize them, and output a maintenance decision report containing the distribution and priority ranking of high-risk areas.

[0066] In one specific embodiment, the process of executing step S8 may specifically include the following steps: The optimized lifespan distribution description is converted into a heat map and overlaid onto the 3D model of the steel structure building to generate a smooth and continuous visual mapping covering all aspects of the building. Lifetime data for each part is extracted from the visualization map to generate a lifetime distribution sequence. The Z-score method is used to calculate outliers in the lifetime distribution sequence, and the building parts corresponding to the outliers are marked as high-risk parts. Calculate the risk level score for each high-risk location. The risk level score is obtained by multiplying the outlier of the high-risk location by the location's importance weight. All high-risk areas are prioritized from highest to lowest based on their risk level scores. A maintenance decision report is generated that integrates a visual mapping, a distribution list of high-risk areas, and their priority ranking, and is used as the final output to improve the accuracy of the measurement.

[0067] Specifically, based on the optimized lifespan distribution description, it is converted into a heatmap using a 3D rendering engine and overlaid onto the building's 3D model, forming a smooth and continuous visual map. This map covers all directions of the building, including east, south, west, north, as well as the top and bottom, creating a panoramic spatial view of the lifespan status. In this way, the lifespan distribution of various parts of the building can be visually displayed, facilitating subsequent identification of high-risk areas and decision analysis. In the heatmap, color gradients reflect the lifespan prediction ranges of different parts, thus revealing the health status and potential degradation trends of each part.

[0068] Based on spatialized lifetime data carried by a visualized mapping, lifetime values ​​for each structural component are extracted to construct a distribution sequence, and the Z-score method is used for outlier detection. The Z-score identifies outliers by measuring the degree of deviation of data points from the mean; for example, points with an absolute Z-score greater than a preset threshold (e.g., 3) are defined as outliers. This method accurately identifies areas where coating aging is accelerated due to environmental factors (such as UV intensity, temperature and humidity changes), and these areas are identified as high-risk areas. The building components corresponding to these outliers are marked as high-risk areas, and the causes of accelerated aging, such as UV intensity or salt spray corrosion, are recorded to form a preliminary risk list.

[0069] After identifying high-risk areas, a risk score is calculated for each area. This score is derived by multiplying the absolute value of the area's lifespan anomaly by its importance weight, thus quantifying the actual threat posed by each high-risk area. For each area in the building structure, the weight is determined based on its importance to the overall structural safety and function. This approach allows for a reasonable assessment of maintenance priorities for each area, ensuring that resources are effectively allocated to the areas requiring the most attention.

[0070] Finally, based on the risk level scores, all high-risk areas are prioritized from highest to lowest, and a maintenance decision report is generated. This report includes a distribution map of high-risk areas, priority ranking, and related environmental impact factors, helping decision-makers quickly identify areas requiring priority maintenance and guiding actual maintenance work. This process not only improves the overall accuracy of the calculations but also provides a scientific basis for the long-term maintenance of steel structure buildings, reducing the risk of unexpected failures due to coating degradation and improving the safety and durability of the buildings.

[0071] For example, in the life prediction of steel structure buildings, assuming a certain part exhibits an abnormally low lifespan in an environment with high ultraviolet radiation, after Z-score calculation, the lifespan Z-score for this part is -4, marking it as a high-risk part, with a weight value set to 0.8. The final calculated risk level score is 3.2, indicating that this part should be prioritized in the allocation of maintenance resources. Furthermore, through heat maps and risk ranking reports, high-risk concentration areas in the eastern direction can be clearly seen, facilitating timely intervention and maintenance.

[0072] This technology enables real-time response to environmental changes, dynamic optimization of coating life prediction, significant improvement in the accuracy and efficiency of maintenance plans, and ensures the long-term safety and stable operation of steel structure buildings.

[0073] The above describes a method for predicting the lifespan of anti-corrosion coatings for steel structures in the embodiments of this application. The following describes a system for predicting the lifespan of anti-corrosion coatings for steel structures in the embodiments of this application. Please refer to [link to relevant documentation]. Figure 2 This application provides a structural schematic diagram of a steel structure building anti-corrosion coating life prediction system, which includes: The data acquisition module 10 is used to acquire real-time environmental data from different parts of the steel structure building and construct a multi-source environmental dataset.

[0074] The feature extraction module 20 is used to analyze the periodic fluctuation pattern of environmental parameters based on the multi-source environmental dataset and extract the fluctuation feature vector that represents the pattern.

[0075] The region classification module 30 is used to perform cluster analysis on various parts of the steel structure building based on the fluctuation feature vector to obtain the region classification results.

[0076] The aging estimation module 40 is used to fit the relationship between environmental factors and coating aging rate in each region for the region classification results, and then determine the estimated value of coating aging rate in that region.

[0077] The lifetime mapping module 50 is used to query and match historical coating performance records based on the estimated coating aging rate, and generate a preliminary lifetime distribution map based on the matching results.

[0078] The interval prediction module 60 is used to generate multi-scenario environmental evolution paths based on the preliminary lifetime distribution map, and calculate the coating degradation trajectory under each path, thereby determining the comprehensive prediction interval of the coating lifetime.

[0079] The parameter adjustment module 70 is used to perform statistical analysis on the comprehensive prediction interval to determine whether there is any abnormal deviation. If so, the prediction parameters are adjusted to obtain an optimized life distribution description.

[0080] The risk visualization module 80 is used to generate a visual mapping covering all aspects of the steel structure building based on the optimized life distribution description, identify high-risk parts with abnormal lifespan, prioritize them, and output a maintenance decision report containing the distribution and priority ranking of high-risk parts.

[0081] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for predicting the service life of anti-corrosion coatings for steel structure buildings, characterized in that, The method includes: Step S1: Obtain real-time environmental data from different parts of the steel structure building and construct a multi-source environmental dataset, wherein the real-time environmental data includes temperature and humidity values ​​and ultraviolet intensity index. Step S2: Based on the multi-source environmental dataset, analyze the periodic fluctuation pattern of environmental parameters and extract the fluctuation feature vector representing the pattern; Step S3: Based on the fluctuation feature vector, perform cluster analysis on each part of the steel structure building to obtain the regional classification results; Step S4: For each region in the region classification results, fit the relationship between environmental factors and coating aging rate in that region, and then determine the estimated value of coating aging rate in that region. Step S5: Based on the estimated coating aging rate, query and match historical coating performance records, and generate a preliminary lifetime distribution map based on the matching results; Step S6: Generate multi-scenario environmental evolution paths based on the preliminary lifetime distribution map, and calculate the coating degradation trajectory under each path, thereby determining the comprehensive prediction range of coating lifetime. Step S7: Perform statistical analysis on the comprehensive prediction interval to determine whether there is any abnormal deviation. If so, adjust the prediction parameters to obtain an optimized lifetime distribution description. Step S8: Based on the optimized lifespan distribution description, generate a visual mapping covering all aspects of the steel structure building, identify high-risk parts with abnormal lifespan, prioritize them, and output a maintenance decision report containing the distribution and priority ranking of high-risk parts. Step S4 includes: For each similar environmental region in the regional classification results, environmental factor data of that region is obtained. The similar environmental region is a region formed by using a clustering algorithm to cluster the fluctuation feature vectors and dividing the parts of the fluctuation feature vectors that are similar into the same cluster. Based on the environmental factor data, a multiple linear regression model was used to fit the relationship between environmental factors and coating aging rate, and the regression coefficients corresponding to each environmental factor were obtained. The aging acceleration coefficient of the region is obtained by weighted summation of the regression coefficients. The product of the aging acceleration coefficient and the preset reference rate is used as the estimated value of the coating aging rate; Step S6 includes: Extract the probability distribution of environmental parameters from the preliminary lifetime distribution map; Based on the probability distribution, Monte Carlo simulation is used to perform multiple random samplings to generate multiple multi-scenario environment evolution paths in future time periods. Each environmental evolution path is input into a preset aging rate model, and the cumulative degradation of the coating is calculated step by step over time to obtain the coating degradation trajectory corresponding to each environmental evolution path. For each coating degradation trajectory, determine the coating lifetime endpoint corresponding to when it reaches a preset failure threshold; Statistical analysis was performed on all coating lifetime endpoint data to fit their probability distribution and generate a probability distribution description including the lifetime mean and standard deviation. Based on the probability distribution description, its confidence interval at a specified confidence level is calculated, and this confidence interval is used as the comprehensive prediction interval for the coating lifetime. Step S7 includes: Step 71: Obtain the probability distribution description and extract the standard deviation. If the standard deviation exceeds the preset standard value, it is determined that there is an abnormal deviation. Then, newly collected real-time environmental data is obtained. Step 72: Incorporate the newly collected real-time environmental data using the Bayesian update method, and adjust the prediction parameters, which include aging rate and lifetime distribution parameters; Step 73: Correct the preliminary lifetime distribution map according to the adjusted prediction parameters and regenerate the lifetime distribution description; Step 74: Perform an analysis of variance on the regenerated lifetime distribution description to evaluate its stability. If it is stable, use the regenerated lifetime distribution description as the optimized lifetime distribution description. If it is unstable, obtain more real-time environmental data and then return to step 72 to continue iterative updates until the analysis of variance results meet the stability requirements or the maximum number of iterations is reached.

2. The method according to claim 1, characterized in that, Step S1 includes: The real-time environmental data is collected by a distributed sensor array deployed at key exposed parts of the steel structure building; The real-time environmental data is preprocessed, including data cleaning and format standardization. The preprocessed data is grouped according to the collection location to form an independent environmental parameter sequence for each location, and an initial multi-source environmental dataset is generated based on the environmental parameter sequence. The initial multi-source environmental dataset is timestamped and the time series is unified. The data series are then integrated to form a multi-source environmental dataset.

3. The method according to claim 2, characterized in that, Step S2 includes: Based on the aforementioned multi-source environmental dataset, a time-series decomposition algorithm is used to process the data sequences, decomposing each data sequence into trend components, seasonal components, and residual components. Based on the time-series decomposition results, periodic components are extracted from the data sequence to obtain the periodic fluctuation pattern; Based on the periodic fluctuation pattern, feature parameters for quantifying the pattern are extracted, including fluctuation period and fluctuation amplitude; The fluctuation feature vector is constructed by using the feature parameters as vector elements.

4. The method according to claim 1, characterized in that, Step S3 includes: Clustering algorithms are used to cluster the wave feature vectors, grouping parts with similar feature vectors into the same cluster, thereby grouping different parts of the steel structure building into several similar environmental regions; The degree of microclimate difference between different similar environmental regions is obtained by calculating the Euclidean distance between the cluster center vectors of different clusters. The regional classification results are generated based on the degree of microclimate differences, and the regional classification results characterize the environmental similarity of each location.

5. The method according to claim 1, characterized in that, Step S5 includes: Using the estimated coating aging rate as a query condition, candidate historical records are retrieved from the historical coating performance database, which includes coating aging data. Based on similarity calculation, the matching degree between the estimated coating aging rate and the candidate historical records is obtained, and historical records with a matching degree greater than a preset matching threshold are selected as valid matching records. Extract the actual coating lifetime data corresponding to all valid matching records, and generate the preliminary lifetime distribution map accordingly. The preliminary lifetime distribution map characterizes the lifetime distribution of the coating under different environments. A smoothing algorithm is used to smooth the preliminary lifetime distribution map to generate a continuous lifetime distribution curve.

6. The method according to claim 1, characterized in that, Step S8 includes: The optimized life distribution description is converted into a heat map and overlaid onto a 3D model of the steel structure building to generate a smooth and continuous visualization map covering all aspects of the building. Lifetime data for each part is extracted from the visualization map to generate a lifetime distribution sequence. The Z-score method is used to calculate outliers in the lifetime distribution sequence, and the building parts corresponding to the outliers are marked as high-risk parts. Calculate the risk level score for each high-risk location, which is obtained by multiplying the outlier of the high-risk location by the location importance weight; Based on the aforementioned risk level scores, all high-risk areas are prioritized from highest to lowest. A maintenance decision report is generated that integrates a visual mapping, a distribution list of high-risk areas, and their priority ranking, and is used as the final output to improve the accuracy of the measurement.

7. A life prediction system for anti-corrosion coatings on steel structures, used to implement the method as described in any one of claims 1 to 6, characterized in that, The system includes: The data acquisition module is used to obtain real-time environmental data from different parts of the steel structure building and construct a multi-source environmental dataset; The feature extraction module is used to analyze the periodic fluctuation pattern of environmental parameters based on the multi-source environmental dataset, and extract the fluctuation feature vector representing the pattern. The region classification module is used to perform cluster analysis on various parts of the steel structure building based on the fluctuation feature vector to obtain the region classification results; The aging estimation module is used to fit the relationship between environmental factors and coating aging rate in each region of the region classification results, and then determine the estimated value of coating aging rate in that region. The lifetime mapping module is used to query and match historical coating performance records based on the estimated coating aging rate, and generate a preliminary lifetime distribution map based on the matching results. The interval prediction module is used to generate multi-scenario environmental evolution paths based on the preliminary lifetime distribution map, and calculate the coating degradation trajectory under each path, thereby determining the comprehensive prediction interval of the coating lifetime. The parameter adjustment module is used to perform statistical analysis on the comprehensive prediction interval to determine whether there is any abnormal deviation. If so, the prediction parameters are adjusted to obtain an optimized lifetime distribution description. The risk visualization module is used to generate a visual mapping covering all aspects of the steel structure building based on the optimized life distribution description, identify high-risk parts with abnormal lifespan, prioritize them, and output a maintenance decision report containing the distribution and priority ranking of high-risk parts.

Images