An artificial intelligence-based method for managing construction engineering quality data

Abstract

The present invention discloses an artificial intelligence-based method for managing construction project quality data, specifically related to the field of construction project quality data management, which is used to solve the problems of imbalance in spatial distribution and correlation analysis of multi-source heterogeneous quality data, monitoring blind spots, and difficulty in accurately capturing abnormal information. It dynamically integrates multi-dimensional quality data at the construction site to achieve unified processing and correlation mining of different regions and different monitoring indicators, eliminating the deviation caused by uneven data distribution and local information loss; uses time-series analysis and spatial difference analysis methods to extract inherent correlations and potential abnormal fluctuation characteristics from multi-source data, so as to obtain more accurate and comprehensive basic data under complex conditions; through data enhancement technology, it carefully reconstructs and complements abnormal or missing data, making the monitoring data continuous and complete in the time and space dimensions, and no longer affecting the overall quality judgment due to local blind spots or missing points.

Description

Technical Field

[0001] The present invention relates to the field of construction project quality data management. More specifically, the present invention relates to an artificial intelligence-based construction project quality data management method. Background Art

[0002] In the quality management of construction projects, the quality data at the construction site often comes from multiple monitoring devices and sensors. These devices may be distributed in different areas and conduct real-time monitoring for different quality indicators, such as temperature and humidity, stress, vibration, etc. With the increasing complexity and scale of modern construction projects, the sources of engineering monitoring data are increasing, and the spatial distribution of quality data has become more extensive and uneven. For example, during the concrete pouring and curing process, different temperature and humidity data may be collected at different positions and different construction stages. These data have a profound impact on the project quality but usually exist as discrete data points. In order to comprehensively evaluate the quality of concrete pouring and curing, it is necessary to effectively integrate the data from different areas through reasonable data fusion and spatial difference analysis.

[0003] Currently, the methods for processing quality data in construction projects mainly rely on the analysis of a single data source and the quality assessment of local areas. However, due to the spatial difference of data and the uneven distribution of sensors, there are often incomplete monitoring data in some areas or key positions are not covered. Traditional analysis methods are difficult to conduct effective spatial correlation analysis on multiple quality data from different positions and cannot comprehensively grasp the quality status of the entire construction area, resulting in a large deviation in quality assessment. For example, during the concrete pouring process, the data may be relatively dense in the central area of the construction but sparse in the edge area, and factors such as temperature and humidity changes and stress distribution in different areas are interrelated. These differences are not effectively integrated, leading to the inability to accurately predict some potential quality problems. The prior art lacks a means to conduct comprehensive analysis across regions and across indicators and cannot accurately reveal the quality differences between different regions and their impact on the overall project quality.

[0004] To solve the above problems, a technical solution is provided herein. Summary of the Invention

[0005] To overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for managing construction project quality data based on artificial intelligence. By dynamically integrating multi-dimensional quality data at the construction site, unified processing and associated mining of different regions and different monitoring indicators are realized, and the deviation caused by uneven data distribution and local information loss is eliminated. Using time-series analysis and spatial difference analysis methods, inherent associations and potential abnormal fluctuation characteristics are extracted from multi-source data, so as to obtain more accurate and comprehensive basic data under complex conditions. Through data enhancement technology, abnormal or missing data are carefully reconstructed and supplemented, so that the monitoring data remains continuous and complete in the time and space dimensions, and the overall quality judgment is no longer affected by local blind spots or missing points. Thus, while ensuring data coverage and accuracy, the present invention provides high-confidence data support for subsequent quality assessment, structural performance prediction, and decision-making optimization, improving the overall level of construction project quality management and the rationality of decision-making, so as to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for managing construction project quality data based on artificial intelligence, comprising the steps of:

[0008] S1, collecting multi-dimensional quality data of different regions at the construction site, and dynamically optimizing the sensor layout based on spatial data entropy and real-time data density;

[0009] S2, identifying and analyzing the collected quality data, using spatial difference analysis combined with time-series analysis to extract the correlation and abnormal fluctuations between the data of each region, and analyzing whether there is data abnormality;

[0010] S3, if there is data abnormality in a certain region, filling the quality monitoring blind area through data enhancement technology to ensure the integrity and accuracy of the data.

[0011] In a preferred embodiment, step S1 includes the following contents:

[0012] First, divide the entire construction site into several sub-regions; set multiple key monitoring points in each sub-region; according to the specific requirements of each key monitoring point, select sensors of corresponding types and precisions, and set their sampling frequencies and data transmission intervals; based on an optimization algorithm combining spatial data entropy and real-time data density, dynamically optimize the sensor layout, and the specific process is as follows:

[0013] Calculation of spatial data entropy: First, calculate the spatial data entropy of each sub-region at the construction site, and the formula is as follows: ; where represents the spatial data entropy of the th sub-region, represents at the Within the sub-region, the data distribution probability of the th quality index;

[0014] Real-time data density assessment: Subsequently, calculate the real-time data density of each sub-region, with the formula as follows: ; where represents the real-time data density of the th sub-region, is the number of data points received in the corresponding region over a past period of time, is the total time of the corresponding time period;

[0015] Optimization decision function: Combining the spatial data entropy and the real-time data density, construct the optimization decision function , which is used to decide whether to adjust the sensor layout, with the formula as follows: ;

[0016] Sensor layout adjustment: According to the value of the optimization decision function, if the optimization decision function of a certain sub-region is greater than the maximum value of the preset threshold range, then increase the number of sensors in the corresponding region; conversely, if the optimization decision function is less than the minimum value of the preset threshold range.

[0017] In a preferred embodiment, step S2 includes the following contents:

[0018] First, preprocess the collected multi-dimensional quality data, including data cleaning and normalization processing; then, use the spatial difference analysis method to compare the data in different regions and identify the significant differences in each index between regions; the specific steps are as follows:

[0019] a. Spatial data standardization: To eliminate the dimensional differences between different quality indexes, standardize each quality index and calculate the standard score of each data point;

[0020] b. Spatial difference index calculation: Calculate the spatial difference index of the standardized data for each sub-region to quantify the overall deviation degree of this region in each quality index; the calculation formula is as follows: ; where is the total number of quality indexes, is the th sub-region's th index's standard score;

[0021] c. Spatiotemporal sequence decomposition and trend extraction: Decompose the spatiotemporal sequence of the time-series data of each sub-region and extract the trend, seasonal, and residual components; use multi-scale wavelet transform for decomposition, with the formula as follows: ; among them, represents the trend component, represents the seasonal component, represents the residual component;

[0022] d. Spatiotemporal correlation matrix construction: Construct a spatiotemporal correlation matrix between sub-regions , where represents the degree of correlation between the th sub-region and the th sub-region in terms of quality indicators; The mutual information is used to measure the non-linear dependence relationship between two sub-regions. The formula is as follows: ; where is the joint probability distribution of the th and th sub-regions on a certain quality indicator, and are their marginal probability distributions respectively.

[0023] e. Abnormal fluctuation detection and comprehensive analysis of spatial differences: Use principal component analysis to reduce the dimension of the spatiotemporal correlation matrix and extract the main spatiotemporal change patterns; The steps are as follows:

[0024] Principal component extraction: Perform singular value decomposition on the spatiotemporal correlation matrix and extract the first principal components to capture the main change patterns in the data;

[0025] Projection error calculation: Calculate the projection error of each sub-region in the principal component space. The formula is as follows: ; where is the predicted value of the th sub-region on the th principal component;

[0026] f. Combine the spatial difference index and the projection error to comprehensively evaluate each sub-region to determine whether there is data abnormality; Set a double judgment criterion, specifically as follows:

[0027] Spatial difference judgment: If the spatial difference index exceeds the set spatial difference threshold, the corresponding sub-region has a significant deviation in quality indicators;

[0028] Abnormal score judgment: If the projection error exceeds the set abnormal score threshold, there is a significant abnormal fluctuation in the quality data of the corresponding sub-region;

[0029] When a sub-region simultaneously meets the conditions that the spatial difference index exceeds the set spatial difference threshold and the projection error exceeds the set abnormal score threshold, it is determined that there is data abnormality in the corresponding region.

[0030] In a preferred embodiment, step S3 includes the following:

[0031] Data missing pattern recognition: After identifying anomalies, it is first necessary to determine the pattern and location of the missing data; by comparing the quality data within the area with the data trends in the surrounding areas, a local time series analysis method is used to identify the missing pattern of the missing data points.

[0032] Multi-level data enhancement: For the data in the missing area, first, perform incremental completion based on spatio-temporal rules for each missing area. Extract spatio-temporal features from adjacent areas and historical data to ensure the effectiveness of data reconstruction; the specific processing method is as follows:

[0033] Spatio-temporal region similarity analysis: Perform local spatial similarity analysis on the missing data points, calculate the similarity with the data points in the surrounding areas; determine the similarity between the missing area and other areas, and then find the areas with qualified similarity as the basis for data reconstruction to obtain similar areas.

[0034] Reconstruction based on neighborhood weighting: After determining the similar areas, use the weighted average method to fuse the valid data in the neighborhood areas to fill the missing values in the missing area; the formula is as follows: ; where is the reconstructed value of the missing data point, is the area and the similarity between them, is the area the data value on the corresponding quality index, is the number of areas participating in the reconstruction.

[0035] The technical effects and advantages of a method for managing construction engineering quality data based on artificial intelligence according to the present invention:

[0036] By dynamically integrating multi-dimensional quality data at the construction site, unified processing and associated mining of different regions and different monitoring indicators are realized, and the deviation caused by uneven data distribution and local information loss is eliminated. Using spatio-temporal sequence analysis and spatial difference analysis methods, inherent associations and potential abnormal fluctuation characteristics are extracted from multi-source data, so as to obtain more accurate and comprehensive basic data under complex conditions. Through data enhancement technology, anomalies or missing data are carefully reconstructed and completed, so that the monitoring data remains continuous and complete in the spatio-temporal dimension, and the overall quality judgment is no longer affected by local blind spots or missing points. Thus, while ensuring data coverage and accuracy, the present invention provides high-confidence data support for subsequent quality assessment, structural performance prediction, and decision optimization, improving the overall level of construction engineering quality management and the rationality of decision-making. Description of the Drawings

[0037] Figure 1 This is a schematic flow diagram of a method for managing construction project quality data based on artificial intelligence according to the present invention. Specific embodiments

[0038] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0039] Embodiment 1: Figure 1 A method for managing construction project quality data based on artificial intelligence according to the present invention is given, including:

[0040] S1, collecting multi-dimensional quality data in different areas of the construction site, and dynamically optimizing the layout of sensors based on spatial data entropy and real-time data density;

[0041] S2, identifying and analyzing the collected quality data, using spatial difference analysis combined with spatio-temporal sequence analysis to extract the correlation and abnormal fluctuations between the data in each area, and analyzing whether there is data abnormality;

[0042] S3, if there is data abnormality in a certain area, filling the quality monitoring blind area through data enhancement technology to ensure the integrity and accuracy of the data.

[0043] In the process of quality management of construction projects, the quality data on the construction site (such as temperature and humidity, stress and vibration, etc.) directly affects the overall quality and safety of the project. Factors such as temperature and humidity changes, stress distribution, and vibration state during concrete pouring need to be monitored and recorded in real time to promptly discover potential quality problems and take effective measures. Therefore, data collection becomes a key link in quality control. The traditional manual collection and data recording methods often cannot meet the requirements of real-time, accuracy, and comprehensiveness in the highly complex construction site, often resulting in some minor abnormalities not being discovered in time, thus affecting the project quality. Through dynamic sensor layout and data synchronous transmission, the accuracy and real-time of data collection can be effectively improved, providing solid data support for subsequent quality analysis and decision-making.

[0044] Step S1 includes the following content:

[0045] First, divide the entire construction site into several sub - regions; within each sub - region, further identify multiple key monitoring points, which are usually located at the center, edge of concrete pouring, and areas vulnerable to environmental impacts. For example, in the concrete foundation construction of a large commercial complex, the site can be divided into several base areas, connection areas, and transition areas, and several key monitoring points are set in each area to cover different construction stages and environmental conditions.

[0046] According to the specific requirements of each key monitoring point, select sensors of corresponding types and precisions, and set their sampling frequencies and data transmission intervals. Temperature and humidity sensors should have high sensitivity and fast response characteristics, stress sensors need to have high precision and stability, and vibration sensors need to have high - frequency sampling capabilities. At the initial deployment, the sensors are evenly arranged according to the pre - determined key monitoring points to ensure coverage of all important areas.

[0047] Based on an optimization algorithm that combines spatial data entropy and real - time data density, dynamically optimize the sensor layout. The specific process is as follows:

[0048] Spatial data entropy calculation: First, calculate the spatial data entropy of each sub - region in the construction site. The formula is as follows: ; where represents the spatial data entropy of the th sub - region, represents within the th sub - region, the probability distribution of the th quality index (such as temperature, humidity, stress, vibration).

[0049] Spatial data entropy represents the non - uniformity of data distribution within a region. The higher the entropy value, the more dispersed and unbalanced the data distribution in that region, and more sensors may be needed to improve data coverage.

[0050] Real - time data density assessment: Subsequently, calculate the real - time data density of each sub - region. The formula is as follows: ; where represents the real - time data density of the th sub - region, is the number of data points received in the corresponding region over a past period of time, is the total time of the corresponding time period.

[0051] Real - time data density represents the data acquisition frequency of this region. The higher the density, the more frequent the data acquisition in this region and the larger the number of data points, and more sensors may not be needed; but if the density is too low, the data quality may decline and the sampling density needs to be increased.

[0052] Optimize the decision function: Combine spatial data entropy and real-time data density to construct an optimized decision function , which is used to determine whether sensor layout needs to be adjusted. The formula is as follows: ;

[0053] : Use the logarithmic function to smooth the real-time data density value to avoid overly suppressing the influence of entropy value at high data density. Adding 1 is to avoid bad mathematical problems in the case of zero density.

[0054] Exponential term: Amplify the influence of spatial data entropy through the exponential function . When the spatial data entropy is high, the corresponding optimized decision function value will also increase significantly, indicating that more sensors may be needed in this area.

[0055] Subtraction of 1 term: Subtract 1 to ensure that the value of the optimized decision is greater than or equal to 0. This ensures that even if the entropy value or density of a certain sub-region is low, the decision function will not cause excessive adjustment.

[0056] Sensor layout adjustment: According to the value of the optimized decision function, if the optimized decision function of a certain sub-region is greater than the maximum value of the preset threshold range, increase the number of sensors or adjust the sensor sampling frequency in the corresponding area to improve data coverage and acquisition accuracy. Conversely, if the optimized decision function is less than the minimum value of the preset threshold range, the sensor density in this area can be reduced to save resources.

[0057] Through wireless network or wired connection, the multi-dimensional quality data collected by each sensor is transmitted to the central data processing unit in real time. To ensure the timeliness and integrity of data transmission, an efficient data compression and transmission protocol is adopted, and a data verification mechanism is implemented during the transmission process to ensure that data is not lost or duplicated.

[0058] The key to step S1 lies in achieving efficient acquisition of multi-dimensional quality data through reasonable sensor layout and dynamic optimization. In this process, first, ensure the comprehensiveness and representativeness of the acquisition points through regional division and identification of key monitoring points, and then dynamically adjust the sensor layout through an optimization algorithm based on spatial data entropy and real-time data density to ensure the comprehensiveness and accuracy of data acquisition. During the data acquisition process, redundant sensors and data interpolation algorithms are also introduced to enhance the robustness and reliability of data acquisition, thus providing high-quality basic data for subsequent quality data analysis. The innovation of this process is not only reflected in the intelligence of sensor layout, but also in effectively improving the accuracy and stability of data acquisition through dynamic adjustment and fault tolerance mechanisms.

[0059] In the quality management process of large-scale construction projects, accurately identifying and analyzing the quality data at the construction site is crucial for ensuring the overall quality and safety of the project. With the development of sensor technology, a large amount of multi-dimensional quality data can be collected in real time at the construction site, such as indicators like temperature and humidity, stress, and vibration. However, this data is often distributed in different spatial regions and changes dynamically with the progress of construction and environmental conditions. In this complex context, how to effectively identify and analyze this multi-source, multi-dimensional quality data and extract the correlations and abnormal fluctuations between regions has become the key to improving the level of project quality management. Due to their limitations, traditional data analysis methods are difficult to fully exploit the potential information in the data, resulting in some quality hazards not being detected in a timely manner. Therefore, the spatial difference analysis and spatio-temporal sequence analysis methods based on artificial intelligence can accurately identify quality anomalies during the construction process in a multi-dimensional data environment and improve the overall intelligence level of quality management.

[0060] Step S2 includes the following:

[0061] In the identification and analysis stage, first, preprocess the collected multi-dimensional quality data, including data cleaning and normalization, to eliminate the influence of sensor noise and data deviation. Then, use the spatial difference analysis method to compare the data in different regions and identify the significant differences in indicators such as temperature and humidity, stress, and vibration between regions. The specific steps are as follows:

[0062] a. Spatial data standardization: To eliminate the dimensional differences between different quality indicators, standardize each quality indicator and calculate the standard score of each data point: ; where is the mean of the th indicator, and is its standard deviation. The standardized data ensures that different quality indicators are compared on the same scale, eliminating the influence of dimensional differences.

[0063] b. Spatial difference index calculation: Calculate the spatial difference index for the standardized data of each sub-region to quantify the overall deviation degree of this region in each quality indicator. The calculation formula is as follows: ; where is the total number of quality indicators, and is the standard score of the th sub-region and the th indicator. The higher the spatial difference index, the greater the deviation degree of this region from the overall mean in multiple quality indicators, and the higher the risk of potential quality problems.

[0064] c. Spatiotemporal Sequence Decomposition and Trend Extraction: For the time-series data of each sub-region perform spatiotemporal sequence decomposition to extract trend, seasonal, and residual components. Multiscale wavelet transform is used for decomposition, and the formula is as follows: ; where represents the trend component, represents the seasonal component, represents the residual component. Through multiscale wavelet transform, the change characteristics of data can be captured at different time scales, improving the sensitivity of anomaly detection.

[0065] d. Spatiotemporal Correlation Matrix Construction: Construct the spatiotemporal correlation matrix between sub-regions , where represents the degree of correlation between the -th sub-region and the -th sub-region in terms of quality indicators. Mutual information is used to measure the non-linear dependence relationship between two sub-regions, and the formula is as follows: ; where is the joint probability distribution of the -th and -th sub-regions on a certain quality indicator, and and are their marginal probability distributions respectively. Mutual information can effectively capture complex non-linear relationships and reveal potential quality correlations between different regions.

[0066] e. Anomaly Fluctuation Detection and Comprehensive Analysis of Spatial Differences: Use principal component analysis to reduce the dimension of the spatiotemporal correlation matrix and extract the main spatiotemporal change patterns. The steps are as follows:

[0067] Principal Component Extraction: Perform singular value decomposition on the spatiotemporal correlation matrix and extract the first principal components to capture the main change patterns in the data.

[0068] Projection Error Calculation: Calculate the projection error of each sub-region in the principal component space, and the formula is as follows: ; where is the predicted value of the -th sub-region on the -th principal component. The larger the projection error, the higher the degree to which the quality data of the sub-region deviates from the overall pattern, and the greater the possibility of abnormal fluctuations.

[0069] f. Combine the spatial difference index and the projection error to comprehensively evaluate each sub-region to determine whether there is data anomaly. Set a dual judgment criterion, specifically as follows:

[0070] Spatial difference determination: If the spatial difference index exceeds the set spatial difference threshold, there is a significant deviation in the quality index for the corresponding sub-region.

[0071] Abnormal score determination: If the projection error exceeds the set abnormal score threshold, there is a significant abnormal fluctuation in the quality data for the corresponding sub-region.

[0072] When a sub-region simultaneously meets the conditions that the spatial difference index exceeds the set spatial difference threshold and the projection error exceeds the set abnormal score threshold, it is determined that there is data abnormality in the corresponding region. Ensure that data abnormality in the region is only confirmed when both spatial difference and abnormal fluctuation exist.

[0073] In step S2, through detailed spatial difference analysis and spatio-temporal sequence analysis, the quality data of each region at the construction site is systematically identified and analyzed, and the correlation and abnormal fluctuations between the data are extracted. Through a series of steps such as standardization processing, spatial difference quantification, time series decomposition, spatio-temporal correlation matrix construction, and abnormal fluctuation detection, the comprehensiveness and accuracy of data analysis are ensured. The spatio-temporal correlation matrix based on mutual information and the abnormal score mechanism based on principal component analysis are introduced, breaking through the limitations of traditional methods in capturing non-linear relationships and processing multi-dimensional data, and effectively improving the sensitivity and accuracy of regional data abnormality detection. This process provides a solid technical foundation for subsequent quality judgment and adjustment, and significantly improves the intelligent level of construction project quality management.

[0074] In the management of construction project quality data, the collected quality data is usually provided by a large number of sensors or detection devices. These data may have blind spots or missing values due to reasons such as equipment failures and environmental interferences. If these missing data are not processed in a timely manner, it will affect the accuracy of subsequent quality assessment and management decisions. In traditional data processing methods, simple interpolation methods or linear estimations are often used to fill these missing values, but these methods often have large errors and cannot effectively restore the authenticity and reliability of the data.

[0075] With the progress of artificial intelligence and data augmentation technologies, the methods of data reconstruction and interpolation completion have become more refined and complex. By adopting more efficient data augmentation technologies, the values of missing regions can be inferred from the data distribution or time series patterns of adjacent regions, thereby effectively filling the blind spots in quality monitoring and restoring the continuity and integrity of the data. The purpose of this step is to accurately reconstruct the missing data through innovative data augmentation technologies, combined with the similarity and spatio-temporal patterns of regional data, to ensure the integrity and accuracy of quality monitoring data, so as to provide more reliable data support for subsequent quality analysis and decision-making.

[0076] Step S3 includes the following:

[0077] During the quality data monitoring process, once data anomalies or deficiencies are identified in a certain area, it is first necessary to determine the nature and scope of the missing data. To complete these missing data, data augmentation techniques are used for processing. This technique not only completes the data through simple interpolation but also combines the data of adjacent areas and spatio-temporal rules for more refined reconstruction. The specific steps are as follows:

[0078] Recognition of data missing patterns: After identifying anomalies, it is first necessary to determine the patterns and locations of the missing data. By comparing the quality data within the area with the data trends in the surrounding areas, local time series analysis methods are used to identify the missing patterns of the missing data points. For example, if the stress data in a certain area is missing, it is necessary to analyze the stress change patterns in the area and adjacent areas within the same time window and construct a trend model of the local time series.

[0079] Multi-level data augmentation: For the data in the missing area, data reconstruction is carried out through multiple levels of augmentation methods. First, incremental completion based on spatio-temporal rules is performed for each missing area. This incremental completion is not based on a single interpolation method but extracts spatio-temporal features from adjacent areas and historical data to ensure the effectiveness of data reconstruction. The specific processing method is as follows:

[0080] Spatio-temporal regional similarity analysis: Perform local spatial similarity analysis on the missing data points to calculate the similarity with the data points in the surrounding areas. Through this similarity measure, the similarity between the missing area and other areas can be determined, and then the areas with qualified similarity can be found as the basis for data reconstruction to obtain similar areas.

[0081] Reconstruction based on neighborhood weighting: After determining the similar areas, the effective data in the neighborhood areas are fused using the weighted average method to fill the missing values in the missing area. The formula is as follows: ; where is the reconstructed value of the missing data point, is the similarity between area and , is the data value of area on the corresponding quality index, is the number of areas participating in the reconstruction. Through this weighting method, it is ensured that the reconstructed value can be as close as possible to the actual quality data in the surrounding areas.

[0082] Verify the filled data. The verification methods include comparison with the surrounding areas and historical data to ensure that the reconstructed data conforms to the overall data trend. For more complex or unexpected data completion, further adjustment measures are taken. Specifically, if the reconstructed data deviates far from the trend in the surrounding areas, it is necessary to return to the original data source for fine manual adjustment or optimization through further spatial difference analysis.

[0083] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0084] Only some exemplary embodiments of the present invention have been described above by way of illustration. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0085] It should be noted that in this text, if there are relational terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0086] As described above, this is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application shall be subject to the protection scope of the claims.

Claims

1. A construction engineering quality data management method based on artificial intelligence, characterized in that: Includes steps: S1, collects multi-dimensional quality data from different areas of the construction site and dynamically optimizes sensor layout based on spatial data entropy and real-time data density, including the following: First, the entire construction site is divided into several sub-areas; multiple key monitoring points are set in each sub-area; according to the specific needs of each key monitoring point, sensors of corresponding types and precisions are selected, and their sampling frequency and data transmission interval are set; Based on the optimization algorithm combining spatial data entropy and real-time data density, the sensor layout is dynamically optimized. The specific process is as follows: First, the spatial data entropy of each sub-area of ​​the construction site is calculated; Subsequently, the real-time data density of each sub-region is calculated; Combining spatial data entropy and real-time data density, an optimization decision function is constructed to determine whether the sensor deployment needs to be adjusted; According to the value of the optimization decision function, if the optimization decision function of a sub-area is greater than the maximum value of the preset threshold range, the number of sensors is increased in the corresponding area; On the contrary, if the optimized decision function is less than the minimum value of the preset threshold range, the sensor density in the corresponding area is reduced to save resources; S2, identify and analyze the collected quality data, use spatial difference analysis combined with spatiotemporal sequence analysis to extract the correlation and abnormal fluctuations between the data in each region, and analyze whether there are data anomalies; Preprocess the collected multi-dimensional quality data; Construct the spatiotemporal correlation matrix between each sub-region; use mutual information to measure the nonlinear dependency between two sub-regions; Perform singular value decomposition on the spatiotemporal correlation matrix, extract the first M principal components, and calculate the projection error of each sub-region in the principal component space; Perform a comprehensive assessment of each sub-region to determine if there are any data anomalies: If the spatial difference index exceeds the set spatial difference threshold, the corresponding sub-region has a significant deviation in quality indicators; If the projection error exceeds the set abnormal score threshold, the quality data of the corresponding sub-region has significant abnormal fluctuations; When a sub-region satisfies both the spatial difference index exceeding the set spatial difference threshold and the projection error exceeding the set abnormal score threshold, it is determined that the corresponding region has data anomalies; S3, if there is data anomaly in a certain area, data enhancement technology is used to fill the quality monitoring blind spot to ensure the integrity and accuracy of the data.

2. The method for managing construction project quality data based on artificial intelligence according to claim 1, characterized in that: Calculate the spatial data entropy of each sub-area of ​​the construction site, the formula is as follows: Among them, SDE j represents the spatial data entropy of the j-th sub-region, Indicates that in the jth sub-region The data distribution probability of the quality indicators; Calculate the real-time data density of each sub-area using the following formula: Among them, RDD j represents the real-time data density of the jth sub-area, N j is the number of data points received by the corresponding area in the past period of time, T j is the total time of the corresponding time period; Constructing the optimization decision function O j , the formula is as follows:

3. The method for managing construction project quality data based on artificial intelligence according to claim 2 is characterized in that: Step S2 includes the following contents: First, the collected multidimensional quality data is preprocessed, including data cleaning and normalization. Then, the spatial difference analysis method is used to compare the data of different regions to identify the significant differences in various indicators between regions. The specific steps are as follows: a. Spatial data standardization: In order to eliminate the dimensional differences between different quality indicators, for each quality indicator X i Perform standardization and calculate the standard score for each data point; b. Spatial difference index calculation: Calculate the spatial difference index SDI for the standardized data of each sub-region j j , in order to quantify the overall deviation of the region in terms of various quality indicators; the calculation formula is as follows: Where N is the total number of quality indicators, Z ij is the standard score of the ith indicator in the jth sub-region; c. Spatiotemporal series decomposition and trend extraction: For each sub-region’s time series data Y j (t) Decompose the spatiotemporal series to extract the trend, seasonality and residual components; use multi-scale wavelet transform for decomposition, the formula is as follows: j (t) = W j (t)+S j (t)+R j (t); where W j (t) represents the trend component, S j (t) represents the seasonal component, R j (t) represents the residual component; d. Construction of spatiotemporal correlation matrix: Construct the spatiotemporal correlation matrix A between each sub-region, where A jk Indicates the degree of correlation between the j-th sub-region and the k-th sub-region in terms of quality indicators; the mutual information I jk To measure the nonlinear dependency between two sub-regions, the formula is as follows: Among them, p(x,y) is the joint probability distribution of the j-th and k-th sub-regions on a certain quality indicator, and p(x) and p(y) are their marginal probability distributions respectively.

4. The method for managing construction project quality data based on artificial intelligence according to claim 3 is characterized in that: Step S2 also includes the following contents: e. Abnormal fluctuation detection and comprehensive analysis of spatial differences: Use principal component analysis to reduce the dimension of the spatiotemporal correlation matrix and extract the main spatiotemporal variation patterns; the steps are as follows: Principal component extraction: Perform singular value decomposition on the spatiotemporal correlation matrix and extract the first M principal components PC1, PC2, ..., PC M , to capture the main patterns of variation in the data; Projection error calculation: Calculate the projection error E of each sub-region in the principal component space j , the formula is as follows: in, is the predicted value of the j-th sub-region on the m-th principal component; f. Combine the spatial difference index and projection error to conduct a comprehensive assessment of each sub-region to determine whether there is data anomaly; set a dual judgment standard, as follows: Spatial difference judgment: If the spatial difference index exceeds the set spatial difference threshold, the corresponding sub-region has a significant deviation in quality indicators; Abnormal score determination: If the projection error exceeds the set abnormal score threshold, the quality data of the corresponding sub-region has significant abnormal fluctuations; When a sub-region simultaneously satisfies the conditions that the spatial difference index exceeds the set spatial difference threshold and the projection error exceeds the set anomaly score threshold, it is determined that data anomalies exist in the corresponding region.

5. The method for managing construction engineering quality data based on artificial intelligence according to claim 4 is characterized in that: Step S3 includes the following contents: Data missing pattern recognition: After identifying the anomaly, the first thing to do is to determine the pattern and location of the missing data; by comparing the quality data in the area with the data trends in the surrounding areas, a local time series analysis method is used to identify the missing pattern of missing data points.

6. The method for managing construction engineering quality data based on artificial intelligence according to claim 5 is characterized in that: Step S3 also includes the following contents: Multi-level data enhancement: For data in missing areas, first, each missing area is incrementally supplemented based on spatiotemporal rules, and spatiotemporal features are extracted from adjacent areas and historical data to ensure the effectiveness of data reconstruction. The specific processing methods are as follows: Spatiotemporal region similarity analysis: perform local spatial similarity analysis on missing data points and calculate the similarity with data points in the surrounding area; determine the similarity between the missing area and other areas, and then find areas that meet the similarity standards as the basis for data reconstruction to obtain similar areas; Neighborhood weighted reconstruction: After determining the similar area, the weighted average method is used to merge the valid data of the neighborhood area to fill the missing values ​​of the missing area; the formula is as follows: Among them, X J is the reconstructed value of the missing data point, S JL is the similarity between regions J and L, Y L is the data value of region L on the corresponding quality index, is the number of regions involved in the reconstruction.

Images