Building enterprise data management and decision-making method based on lake and warehouse integrated architecture

Abstract

The invention provides a building enterprise data management and decision-making method based on a lake-warehouse integrated architecture, and relates to the technical field of building enterprise data management, and the method comprises the steps: collecting data information of a project construction stage in which a building enterprise is responsible, storing the data information in a data lake, and structuring all data in a data warehouse; construction stage labels are calculated according to the spatial feature information and the business feature information, the construction stage of the current building enterprise responsible project is judged according to the construction stage labels, and a space-time demand field is constructed according to the construction stage of the current building enterprise responsible project. S2, predicting resource demand quantity of all spatial positions of a construction site in the construction process, partitioning the construction site by taking a peak point of a space-time demand field as a seed point, specifically including physical partitioning and digital twinborn partitioning, and obtaining the resource demand quantity predicted in S2 and data of a data lake, and calculating an abnormal field in the construction stage according to the obtained data.

Description

Technical Field

[0001] The present invention relates to the technical field of construction enterprise data management, and specifically to a construction enterprise data governance and decision-making method based on a lake-warehouse integrated architecture. Background Art

[0002] Construction enterprise data covers all types of information generated by construction enterprises during their operations. This data runs through the entire life cycle of a project, from project planning, design, construction to operation and maintenance. During the construction phase, because the total project time is fixed, it is necessary to make decisions on the overall construction progress based on changes in construction phase data. To meet the data needs of construction companies, existing enterprises adopt a data lake-warehouse integrated architecture for data governance and decision-making. This architecture combines the flexibility of a data lake with the efficiency of a data warehouse, enabling unified data storage, management, and analysis. Data lakes can store massive amounts of raw data in various formats, preserving data integrity for enterprises. Data warehouses, on the other hand, provide high-quality, structured data by cleaning, organizing, and modeling it, facilitating rapid query and analysis. However, in actual use, because the data in the construction phase is complicated and involves multiple links, the data in each link changes dynamically and rapidly, and the timeliness of decision-making in the construction phase is required to be high, it is difficult to meet the needs of construction companies. Summary of the Invention

[0003] In response to the shortcomings of the existing technology, the present invention provides a data governance and decision-making method for construction enterprises based on a lake-warehouse integrated architecture.

[0004] A data governance and decision-making method for construction enterprises based on a lake-warehouse integrated architecture includes the following steps: S1. Collect data and information about the construction phase of a project by the construction company and store it in the data lake. The data warehouse structures all the data. Extract spatial feature information, including Building Information Model (BIM) data and ultra-wideband (UWB) data from the data lake. Then extract service feature information, including the entropy of the current spectrum of construction equipment and the gradient of construction material consumption. The construction phase label is calculated based on the spatial feature information and business feature information. Based on the construction phase label, the construction phase of the project currently under the responsibility of the construction company is determined. S2. Construct a spatiotemporal demand field based on the construction phase of the project currently under the construction enterprise's responsibility, and predict the resource demand at all spatial locations on the construction site during the construction process; S3. Partition the construction site based on the peak points of the spatiotemporal demand field as seed points, including physical partitioning and digital twin partitioning; When S1 determines that the construction phase of the project currently under the charge of the construction enterprise has changed, S2 is used to reconstruct the demand field and re-partition the project based on the reconstructed demand field to obtain the partitioning result. S4, receiving the partitioning results of S3, obtains the resource demand data of the construction phase of different partitions, and makes corresponding decisions based on the comparison results of the spatiotemporal demand field data; S5. Obtain the resource demand predicted in S2 and the data from the data lake, and calculate the abnormal field during the construction phase based on the acquired data; After obtaining the abnormal field, the abnormal state of the current construction phase is taken as the starting point, and the expected state of the construction phase is set as the end point to obtain the adjustment path, and the abnormal state of the construction phase is repaired according to the adjustment path.

[0005] Preferably, the specific working steps of S1 are as follows: The spatial feature information extraction specifically includes: encoding the building information model BIM coordinates to obtain BIM coordinates , collect the trajectory vector G of the ultra-wideband device and fuse it to express it as ; The business feature information extraction specifically includes: the construction equipment is specifically a concrete mixer, and the entropy value of the current spectrum is obtained by collecting the current information of the construction equipment; The construction materials specifically refer to the prefabricated components during the construction phase. The difference between the construction material consumption at two adjacent time points is calculated and divided by the time interval to obtain the average gradient within the time period, which is used as the gradient of construction material consumption.

[0006] Preferably, the specific steps of calculating the construction stage label in S1 are as follows: First, we build a graph structure and determine the nodes of the graph. We divide the construction phases of the project that the current construction company is responsible for into stages, and use the start and end of each stage as the key points of construction. The feature vector of each key point is composed of spatial feature information and business feature information. Then, the adjacency matrix is ​​constructed to establish the connection relationship between the nodes and form an adjacency matrix; Then initialize the GCN model, define the number of GCN layers, the number of units in each layer, the activation function and other hyperparameters, build the graph convolution layer and classification layer of the GCN model, build the graph convolution layer, input is the node feature matrix and adjacency matrix, and output is the updated node feature matrix; After the last graph convolution layer, a classification layer is added to map node features to the probability distribution of stage labels; The constructed node feature matrix and adjacency matrix are used as input data, and the true stage label corresponding to each node is prepared as the training label; Select the optimizer and loss function to compile the GCN model; Prepare training data, input the training data into the GCN model, perform iterative training, and adjust the weights according to the value of the loss function in each round of training; Input new data into the trained GCN model. For new construction key point data, its spatial features and business features need to be calculated first, and the corresponding graph structure needs to be constructed. The node feature matrix and adjacency matrix of the new data are input into the trained GCN model. The model will output the probability distribution of each node belonging to different stage labels. Based on the probability distribution, the stage label with the highest probability is selected as the predicted stage label output for the construction key point.

[0007] Preferably, the specific steps of S2 in constructing the spatiotemporal demand field are as follows: S21. Divide the time and space of the current project of the construction enterprise into a three-dimensional grid, and set the grid size to unit grid; S22. Define the resource demand calculation formula and construct the time-space demand field. The time is determined by the change rate of the construction stage and the spatial distribution of resources. When the grid Resource demand at ; The specific calculation formula is as follows: ; in Representation Grid exist The resource demand at the moment, in standard man-day equivalents, is the rate of change during the construction phase; For the Basic demand weight of class resources; For the moment The center coordinates of the class resource spatial distribution heat map; and is the regression coefficient, which is used to balance the effects of the change rate of the construction stage and the spatial distribution of resources on the resource demand; is a minimum positive constant used to prevent the denominator from having a zero value; S23. When the construction phase of the construction site changes according to the predicted phase label, a phase change instruction is triggered, and the demand field reconstruction process of S22 is repeated to update the resource demand field at fixed time intervals.

[0008] Preferably, the S2 further includes S24, which is used to modify the resource demand calculation formula according to the business feature information and spatial feature information extracted in real time by S1: S24. Adjust the basic demand weights of various resources based on the business feature information extracted in real time ; According to the extracted spatial feature information, the coordinates of the resource thermal center are Make corrections.

[0009] Preferably, the specific steps of S3 are as follows: S31, identifying peak points in the demand field, performing DBSCAN clustering on the peak points, and generating k seed point sets; S32, dividing the physical construction area with the seed point set as the center to complete the physical partition; Specifically, based on the seed point set, the construction site is divided into k physical construction areas using the Voronoi diagram algorithm. Each seed point corresponds to a Voronoi region, which contains all the grid cells closest to the seed point. Regularly calculate the difference in resource requirements between adjacent partitions. For each adjacent partition, calculate the ratio of the difference in resource requirements to the maximum resource requirement: When the ratio exceeds the set value, the partition boundary is adjusted; The adjustment method is to re-divide the grid cells close to the boundary of the high-demand zone to the low-demand zone until the ratio is lower than the set value or cannot be adjusted any further; S33. Map the physical partitions in the BIM model, and add an additional data layer containing resource requirements D(x, y, z, t) to complete the digital twin partitioning. When the demand field is reconstructed, recalculate the seed points and update the partitions.

[0010] Preferably, the specific working steps of S31 are as follows: Extract the resource demand D(x,y,z,t) of all grid cells from the demand field and calculate its mean and standard deviation ; Traverse each grid cell in the demand field and find the grid cell that satisfies D(x,y,z,t)> +2 The grid points of , these points are the candidate peak points; Steps: Use the DBSCAN clustering algorithm to cluster the candidate peak points; The candidate peak points are divided into multiple clusters, each cluster represents a high-demand area, and the center point is selected from each cluster as the seed point. The seed point set is expressed as .

[0011] Preferably, the specific steps of S4 are as follows: Establish a demand feature vector for each partition k, specifically calculating the average demand of the partition, the maximum demand of the partition, the demand gradient and the material turnover time; Setting thresholds 、 and ; When the maximum value of the demand gradient exceeds the threshold When the ratio of the maximum demand to the average demand of the partition exceeds the threshold, the device scheduling instruction is triggered. When the material turnover time exceeds the threshold, the material replenishment instruction is triggered. , triggers the logistics optimization instruction.

[0012] Preferably, the specific working steps of S5 are as follows: Obtain the predicted resource demand data from step S2, including the predicted resource demand for each grid point ; Obtain real-time resource demand data from the data lake ; For each grid point, the relative error between the observed resource demand and the predicted resource demand is calculated, and then the average value of all grid points is calculated to obtain the spatial anomaly degree. ; Calculate the absolute error of the time derivative of resource demand and integrate it within the set time period to obtain the time anomaly degree ; Combine the spatial anomaly and temporal anomaly to calculate the comprehensive anomaly value of each grid point and construct a three-dimensional anomaly matrix ; The prefabricated component ID in the BIM model is associated with the trajectory data of the UWB device to find the equipment or personnel that enter or approach the abnormal grid during the abnormal time period as the source of the abnormality.

[0013] Preferably, the specific steps of said S5 according to the abnormal state of the repair construction stage according to the adjustment path are as follows: According to the three-dimensional anomaly matrix Set the starting state S0 and the target state ST; Constructing the decision process, action set , define a reward function to evaluate the immediate reward brought by executing action a in state s; The algorithm is used to learn the optimal action selection under different states and obtain the optimal repair path from the starting state to the target state.

[0014] The present invention provides a data governance and decision-making method for construction enterprises based on a lake-warehouse integrated architecture. It has the following beneficial effects: 1. Multi-source data fusion can comprehensively reflect the construction situation, build construction stage labels to accurately grasp the construction progress, build spatiotemporal demand fields to predict resource needs, and reasonable zoning facilitates construction management. The repair mechanism based on abnormal fields can improve the stability and controllability of the construction process. Overall, it can effectively manage construction company data and assist decision-making, thereby improving construction efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] like Figure 1 The present invention proposes a data governance and decision-making method for construction enterprises based on a lake-warehouse integrated architecture, comprising the following steps: The following steps are involved: S1. Collect data and information about the construction phase of a project by the construction company and store it in the data lake. The data warehouse structures all the data. Perform spatial feature information extraction, specifically including the Building Information Model (BIM) data and ultra-wideband device (UWB) data in the data lake, and then perform business feature information extraction, specifically including the entropy value of the current spectrum of construction equipment and the gradient of construction material consumption. It should be noted that the relevant data collected by the construction company during the project construction phase includes BIM data (including information such as the spatial coordinates of the building structure), UWB data (equipment trajectory vector information), construction equipment current spectrum data, and construction material consumption data. The collected data is stored in the data lake for subsequent unified management and analysis.

[0018] Use the data warehouse to structure all the data in the data lake, organizing different types of data into structured data formats that are easy to analyze and extract. For example, BIM data and UWB data can be stored in corresponding structured tables to prepare for the subsequent extraction of spatial features and business features. The construction phase label is calculated based on the spatial feature information and business feature information. Based on the construction phase label, the construction phase of the project currently under the responsibility of the construction company is determined. S2. Construct a spatiotemporal demand field based on the construction phase of the project currently under the construction enterprise's responsibility, and predict the resource demand at all spatial locations on the construction site during the construction process; S3. Partition the construction site based on the peak points of the spatiotemporal demand field as seed points, including physical partitioning and digital twin partitioning; When S1 determines that the construction phase of the project currently under the charge of the construction enterprise has changed, S2 is used to reconstruct the demand field and re-partition the project based on the reconstructed demand field to obtain the partitioning result. S4, receiving the partitioning results of S3, obtains the resource demand data of the construction phase of different partitions, and makes corresponding decisions based on the comparison results of the spatiotemporal demand field data; S5. Obtain the resource demand predicted in S2 and the data from the data lake, and calculate the abnormal field during the construction phase based on the acquired data; After obtaining the abnormal field, the abnormal state of the current construction phase is taken as the starting point, and the expected state of the construction phase is set as the end point to obtain the adjustment path, and the abnormal state of the construction phase is repaired according to the adjustment path.

[0019] It should be noted that based on the lake-warehouse integrated architecture, construction data from construction companies is collected into the data lake and structured using the data warehouse. Multi-source data, including BIM data, UWB data, construction equipment current spectrum entropy, and construction material consumption gradients, is integrated. Feature extraction constructs construction phase labels to determine the construction phase. A spatiotemporal demand field is constructed based on the construction phase to predict resource demand. This is then used as a basis for zoning the construction site. Decisions are then made based on the zoning results and the spatiotemporal demand field. The system also introduces calculations for abnormal fields during the construction phase and adjusts paths to repair abnormal conditions. The lake-warehouse integrated architecture is conducive to the efficient storage, management and analysis of data, breaking down data silos. The integration of multi-source data can comprehensively reflect the construction situation, build construction stage labels to accurately grasp the construction progress, build a spatiotemporal demand field to predict resource needs, and reasonable zoning facilitates construction management. The repair mechanism based on the abnormal field can improve the stability and controllability of the construction process. Overall, it can effectively manage the data of construction enterprises and assist decision-making, thereby improving construction efficiency and quality.

[0020] As an optional embodiment: the specific working steps of S1 are as follows: The spatial feature information extraction specifically includes: encoding the building information model BIM coordinates to obtain BIM coordinates , collect the trajectory vector G of the ultra-wideband device and fuse it to express it as It should be noted that the Geohash algorithm is used to encode BIM coordinates. By dividing the coordinate space into grids and then encoding the coordinates according to the grid where they are located, the encoding length can control the accuracy. The encoded result can represent the approximate location of the geographical location and convert the continuous building space coordinates into discrete Geohash codes. By receiving signals from multiple UWB base stations, the real-time location coordinates of the device are calculated to form a trajectory vector, which can be represented as a series of location coordinate points corresponding to timestamps; The specific fusion method is splicing fusion. For example, the character sequence of the Geohash code and the coordinate point sequence in the UWB trajectory vector are spliced ​​together in a certain order to form a longer feature vector. Similarly, the device current spectrum entropy value and the material consumption gradient value are spliced ​​together. In this way, feature information from different sources is integrated to form a comprehensive feature representation for subsequent data analysis, machine learning modeling and other tasks. The business feature information extraction specifically includes: the construction equipment is specifically a concrete mixer, and the entropy value of the current spectrum is obtained by collecting the current information of the construction equipment; it should be noted that in this embodiment, the entropy value is specifically calculated using the Shannon entropy formula; Construction materials specifically refer to prefabricated components during the construction phase. The average gradient within the time period is obtained by calculating the difference between the construction material consumption at two adjacent time points and dividing it by the time interval. This is used as the gradient of construction material consumption. Prefabricated components specifically include various types of prefabricated concrete parts; It should be noted that the fusion of Geohash coding and UWB trajectory vectors can convert continuous building space coordinates into discrete codes, while grasping the real-time location of the equipment, providing accurate spatial information for subsequent analysis. The detailed calculation of business characteristic information can more accurately extract equipment operating status and material consumption trends, which will help to more accurately judge the construction stage and make resource demand forecasts.

[0021] As an optional embodiment: the specific steps of calculating the construction stage label in S1 are as follows: First, we construct a graph structure and determine the nodes of the graph. We divide the construction phases of the project that the current construction company is responsible for into stages, and use the start and end of each stage as construction key points. The feature vector of each construction key point is composed of spatial feature information and business feature information. For example, if each node has 5-dimensional spatial features and 3-dimensional business features, the shape of the feature matrix X is (N, 8), where N is the number of nodes. Then, an adjacency matrix is ​​constructed to establish the connection relationship between nodes. For example, if there is a sequence or mutual constraint relationship between two construction key points, the corresponding position in the adjacency matrix is ​​marked as 1, otherwise it is marked as 0. For the 10 construction key points, if key point 1 must be completed before key point 2, the position in the first row and second column of the adjacency matrix A is marked as 1, otherwise it is marked as 0. Then initialize the GCN model, define the number of GCN layers, the number of units in each layer, the activation function and other hyperparameters, build the graph convolution layer and classification layer of the GCN model, build the graph convolution layer, input is the node feature matrix and adjacency matrix, and output is the updated node feature matrix; After the last graph convolution layer, a classification layer is added to map node features to the probability distribution of stage labels. It should be noted that the appropriate number of GCN layers should be selected based on the complexity of the problem and the amount of data. Determine the number of units in each layer. Usually, as the number of layers increases, the number of units can be gradually reduced so that the model can gradually extract higher-level feature representations; In this embodiment, a GCN model with two graph convolutional layers is initialized, with 16 units in the first layer and 8 units in the second layer, both using the ReLU activation function; followed by a classification layer using the Softmax activation function; The constructed node feature matrix and adjacency matrix are used as input data, and the true stage label corresponding to each node is prepared as the training label. It should be noted that the true stage label corresponding to each node is the historical construction stage label of the construction company. Select the optimizer and loss function to compile the GCN model; Select a suitable optimizer. In this example, the Adam optimizer is used. Select the loss function according to the problem type. In this embodiment, the cross entropy loss function is used. Associate the optimizer and loss function with the GCN model to complete the model compilation process; Prepare training data and input it into the GCN model for iterative training. In each round of training, adjust the weights according to the value of the loss function. It should be noted that the training data is the node feature matrix and the adjacency matrix. The spatial features and business features of each key point extracted are combined into a feature vector to form the feature matrix X. Ensure that the construction of the adjacency matrix A is consistent with the actual construction logic dependencies, and that the dimension of the matrix corresponds to the number of nodes in the feature matrix X; It should also be noted that the number of training rounds and batch size are set. In this embodiment, a larger batch size is set, such as the same as the number of nodes, because graph data is usually processed as a whole; In each round of training, the model automatically adjusts the weights according to the value of the loss function to minimize the difference between the predicted results and the true labels; Train the model for 100 rounds, using all nodes for training in each round; Input new data into the trained GCN model. For new construction key point data, its spatial features and business features need to be calculated first, and the corresponding graph structure needs to be constructed. The node feature matrix and adjacency matrix of the new data are input into the trained GCN model. The model will output the probability distribution of each node belonging to different stage labels. Based on the probability distribution, the stage label with the highest probability is selected as the predicted stage label output for the construction key point.

[0022] It should be noted that the construction phase labels are calculated by building a graph structure and using the GCN model, including specific steps such as determining the nodes of the graph, building the adjacency matrix, initializing the GCN model, and training the GCN model. The graph structure can intuitively represent the logical relationship between key points in the construction phase. The GCN model can comprehensively consider spatial feature information and business feature information, learn the deep feature representation of key points, and thus more accurately classify the construction phases, improve the accuracy of construction phase judgment, and provide a reliable basis for subsequent resource demand forecasting and decision-making.

[0023] As an optional embodiment: the specific steps of constructing the spatiotemporal demand field in S2 are as follows: S21. Divide the time and space of the current project of the construction enterprise into a three-dimensional grid, and set the grid size to It should be noted that in this embodiment, the size of each unit grid does not exceed ; S22. Define the resource demand calculation formula and construct the time-space demand field. The time is determined by the change rate of the construction stage and the spatial distribution of resources. When the grid Resource demand at ; The specific calculation formula is as follows: ; in Representation Grid exist The resource demand at the moment, in standard man-day equivalents, The rate of change of the construction phase reflects the speed of the construction phase and is obtained by differencing the time series of the phase labels. For the Basic demand weight of resource type; represents the contribution weight of different prefabricated component types to resource demand; For the moment The center coordinates of the class resource spatial distribution heat map; it should be noted that the specific method of obtaining it is to add a construction status mark for each prefabricated component in the BIM model, including states such as "not under construction", "under construction", and "completed"; According to the construction schedule and actual construction situation, prefabricated components in the "under construction" state are selected; For prefabricated components in the "under construction" state, the geometric center coordinates are calculated and used as the construction hotspot coordinates of the resource; and is the regression coefficient, which is used to balance the impact of the construction phase change rate and resource spatial distribution on resource demand. This is achieved by collecting the actual resource demand data of multiple historical projects and the corresponding construction phase change rate and resource spatial distribution data, establishing a regression model, and solving it through regression analysis methods such as the least squares method. is a minimum positive constant used to prevent the denominator from being zero; in this embodiment, it is set to 1×10^-6; S23. When the construction phase of the construction site changes according to the predicted phase label, a phase change instruction is triggered, and the demand field reconstruction process of S22 is repeated to update the resource demand field at fixed time intervals.

[0024] In this embodiment, the fixed time interval is 5 minutes. Within 5 minutes after receiving the phase change instruction, the resource demand D(x, y, z, t) of each grid unit is recalculated based on the latest construction phase information and resource demand situation, including: Update construction phase change rate ; Adjust the basic demand weights of various resources according to the new construction stage ; Re-determine the center coordinates of the resource space distribution heat map ; The calibrated regression coefficients are substituted into the resource demand forecasting model to calculate the new resource demand field distribution; The reconstructed resource demand field data is pushed to construction management personnel, resource scheduling systems, and on-site workers in real time. Construction management personnel can intuitively view changes in the resource demand field through mobile terminals or management platforms. The resource scheduling system optimizes resource allocation plans based on the updated resource demand field. On-site workers can rationally arrange construction tasks and resource usage based on the latest resource demand information. It should be noted that the paper describes the division of the construction site into a three-dimensional grid when constructing the spatiotemporal demand field, defines the resource demand calculation formula, and determines the resource demand based on the construction phase change rate and resource spatial distribution. It also proposes a process for updating the construction phase change rate after triggering the phase change instruction, adjusting the resource base demand weight, re-determining the center coordinates of the resource spatial distribution heat map, and reconstructing the resource demand field. Three-dimensional grid division can discretize complex construction sites, facilitating the precise calculation of resource requirements in each area. The defined resource demand calculation formula comprehensively considers construction progress and spatial factors, and can more accurately predict resource demand. When the construction stage changes, relevant parameters are updated and the demand field is reconstructed in a timely manner, ensuring the real-time and accuracy of resource demand predictions, and helping to optimize resource allocation.

[0025] As an optional embodiment: the S2 further includes S24, which is used to modify the resource demand calculation formula according to the service feature information and spatial feature information extracted in real time by S1: S24. Adjust the basic demand weights of various resources based on the business feature information extracted in real time ; It should be noted that the specific formula is as follows: For the initial The basic demand weight of the resource type indicates the basic demand for the resource under normal construction conditions; ; For the initial The basic demand weight of the resource type indicates the basic demand for the resource under normal construction conditions; After dynamic calibration Class resource base demand weight; is the device current spectrum entropy value at the current moment. When the device is abnormal, the current spectrum entropy value will increase, indicating that the device's operating state has become unstable; Specifically, a high-precision current sensor, such as a Hall current sensor, is connected to the power supply circuit of the device, and the voltage signal output by the current sensor is connected to the data acquisition card; Collect current signals over a period of time. The length of this period needs to be determined based on the equipment's operating cycle and possible abnormal conditions. The collected time domain current signal is converted into the frequency domain through fast Fourier transform; For example, for a collected current signal sequence x(n), whose length is N, after FFT, the frequency domain sequence X(k) is obtained, where k = 0, 1, ..., N-1. Each point in the frequency domain sequence corresponds to the amplitude and phase information of a frequency component; The power spectral density (PSD) reflects the power distribution of the signal at different frequencies. It can be calculated by the square of the modulus of the FFT result, that is, PSD(k)=|X(k)|². For example, if the complex amplitude of a frequency point k after FFT is 3+4i, then its power spectral density is (3)²+(4)²=25; Normalize the power spectral density so that its sum is 1, which is used as the probability density function of the frequency component, that is, p(k)=PSD(k) / ΣPSD(k), where ΣPSD(k) is the sum of the power spectral density of all frequency points; Calculate the spectrum entropy according to the information entropy formula: the device current spectrum entropy Hcurrent = -Σp(k) × log2p(k). If the probability density of m frequency points in the normalized probability density sequence is greater than 0, then sum these m points. Specifically, at the construction site or production site, arrange for a dedicated person to regularly record the material consumption. The recording frequency can be determined according to the project progress and the speed of material consumption. The collected material consumption data are arranged in chronological order to construct a time series. Regression analysis is used to fit the material consumption time series to obtain ; The characteristic impact coefficient is used to adjust the influence of business characteristics on resource weights; In this embodiment The values ​​of are 0.15 and 0.2 respectively, which are used to balance the changes in equipment status and material consumption. In actual use, they are adjusted according to needs; According to the extracted spatial feature information, the coordinates of the resource thermal center are It should be noted that the specific formula is as follows: ; : The center coordinates of the corrected spatial distribution heat map of the j-th resource at time t.

[0026] Based on the center coordinates obtained by clustering UWB device trajectory points, UWB devices can track the locations of personnel and equipment on the construction site in real time. Clustering analysis of UWB trajectory points using the DBSCAN clustering algorithm can reveal the concentrated distribution areas of resources in actual construction. : The construction hotspot coordinates identified from the BIM component construction status mark. The BIM model contains the detailed design information and construction schedule of the building. By analyzing the construction status of the components, the resource demand hotspot location can be predicted.

[0027] : Dynamic adjustment factor, used to control the weight of UWB data and BIM data in the calculation of thermal center coordinates. In this embodiment, the specific value is: when the construction phase label belongs to the main construction or equipment installation phase, Take 0.7, at this time, more emphasis is placed on the actual location data provided by the UWB device, because the construction activities in these two stages are more complex and the location changes frequently; in other stages Taking 0.3, it is relatively more dependent on the predicted position of the BIM model, because at this time the construction activities are relatively regular and the prediction accuracy of the BIM model is higher; By combining the real-time location data provided by UWB devices with the predicted location information from the BIM model, the coordinates of the resource thermal center are corrected. This fusion approach fully leverages the strengths of both data sources, improving the accuracy and robustness of resource spatial distribution predictions. During complex phases such as main construction and equipment installation, the real-time data from UWB devices can better reflect the actual construction situation; while during other relatively stable construction phases, the predicted locations from the BIM model are more reliable. It's important to note that adjusting resource weights based on real-time equipment current spectrum entropy and material consumption gradients can make resource demand forecasts more aligned with actual construction conditions, such as in response to equipment failures or material consumption anomalies. Correcting the coordinates of resource thermal centers, integrating real-time UWB device location data with predicted location information from the BIM model, improves the accuracy and robustness of resource spatial distribution forecasts and enhances the reliability of resource demand forecasts.

[0028] As an optional embodiment: the specific steps of S3 are as follows: S31, identifying peak points in the demand field, performing DBSCAN clustering on the peak points, and generating k seed point sets; S32, dividing the physical construction area with the seed point set as the center to complete the physical partition; Specifically, based on the seed point set, the construction site is divided into k physical construction areas using the Voronoi diagram algorithm. Each seed point corresponds to a Voronoi region, which contains all the grid cells closest to the seed point. Regularly calculate the difference in resource requirements between adjacent partitions. For each adjacent partition, calculate the ratio of the difference in resource requirements to the maximum resource requirement: When the ratio exceeds the set value, the partition boundary is adjusted; it should be noted that the value set in this embodiment is 0.3, and the specific value range is between 0.2 and 0.35, which is set by the construction personnel; The adjustment method is to re-divide the grid cells close to the boundary of the high-demand zone to the low-demand zone until the ratio is lower than the set value or cannot be adjusted any further; S33. Map the physical partitions in the BIM model. The additional data layer contains the resource demand D(x, y, z, t) to complete the digital twin partition. When the demand field is reconstructed, the seed point is recalculated and the partition is updated. It should be noted that the following additional data layers are integrated into the digital twin partition: Resource demand matrix D(x,y,z,t): stores the resource demand of each grid cell and is used to visually display the resource demand distribution; Generate heat maps showing device density and distribution based on UWB device trajectories and device types to optimize device scheduling; Calculate and display the safety risk density of each area based on safety monitoring data (such as height operations and temporary power consumption) to assist in safety management; It should be noted that zoning the construction site and dynamically adjusting the zoning boundaries according to resource demand can make the zoning more in line with actual construction needs. Mapping the physical partitions in the BIM model and adding a data layer realizes the combination of physical construction and digital twins, which facilitates the intuitive display and analysis of the construction situation and provides more comprehensive information support for construction management.

[0029] As an optional embodiment: the specific working steps of S31 are as follows: Extract the resource demand D(x,y,z,t) of all grid cells from the demand field and calculate its mean and standard deviation ; Traverse each grid cell in the demand field and find the grid cell that satisfies D(x,y,z,t)> +2 The grid points of , these points are the candidate peak points; Steps: Use the DBSCAN clustering algorithm to cluster the candidate peak points; Parameter settings: The neighborhood radius is set to 2-3 meters (the specific value is adjusted according to the resource density of the construction site), and the minimum number of samples MinPts is set to 5-10 (depending on the cluster size); The candidate peak points are divided into multiple clusters, each cluster represents a high-demand area, and the center point is selected from each cluster as the seed point. The seed point set is expressed as .

[0030] It should be noted that the ability to accurately identify high-demand areas in the demand field provides a key basis for the subsequent reasonable division of construction zones. Reasonable parameter settings ensure the accuracy and reliability of the clustering results, thereby improving the rationality of the zones and contributing to the effective allocation and management of construction resources.

[0031] As an optional embodiment: the specific steps of S4 are as follows: Establish a demand feature vector for each partition k, specifically calculating the average demand of the partition, the maximum demand of the partition, the demand gradient and the material turnover time; Setting thresholds 、 and ; When the maximum value of the demand gradient exceeds the threshold When the ratio of the maximum demand to the average demand of the partition exceeds the threshold, the device scheduling instruction is triggered. When the material turnover time exceeds the threshold, the material replenishment instruction is triggered. It should be noted that the demand gradient values ​​corresponding to the time when the construction efficiency of all projects begins to decline significantly are statistically analyzed, and the average value is taken as When the maximum value of the demand gradient exceeds the threshold When , it indicates that the demand changes rapidly in space, and equipment may need to be rescheduled to meet the construction needs of different areas; When the ratio of the maximum demand to the average demand of the partition exceeds the threshold When the material shortage occurs, it is necessary to replenish the materials in time to prevent the construction delay caused by material shortage. Specifically, the demand ratio corresponding to the material shortage risk in all projects is statistically analyzed and the average value is taken as The initial value of When the material turnover time exceeds the threshold When the material inventory is overstocked or consumed too slowly, it is necessary to optimize the logistics distribution plan to reduce inventory costs and capital occupation. Specifically, statistical analysis is conducted on the material turnover time corresponding to the time when the cost and risk begin to increase significantly in all projects, and the average value is taken as The initial value of It should be noted that by establishing demand feature vectors for partitions, we can fully understand the construction demand situation of each partition. Setting thresholds and triggering corresponding instructions can timely detect abnormal situations in the construction process and take measures, such as equipment scheduling, material replenishment, logistics optimization, etc., to ensure the smooth progress of the construction process and improve construction efficiency and quality.

[0032] As an optional embodiment: the specific working steps of S5 are as follows: Obtain the predicted resource demand data from step S2, including the predicted resource demand for each grid point ; Obtain real-time resource demand data from the data lake ; For each grid point, the relative error between the observed resource demand and the predicted resource demand is calculated, and then the average value of all grid points is calculated to obtain the spatial anomaly degree. ; It should be noted that the specific calculation formula is: ; Where N is the number of grid points, represents the resource demand data of the i-th grid point, Represents the predicted resource demand at the i-th grid point ; Calculate the absolute error of the time derivative of resource demand and integrate it within the set time period to obtain the time anomaly degree ; It should be noted that, for example, if the time period is set to The specific formula for integration is ; Combine the spatial anomaly and temporal anomaly to calculate the comprehensive anomaly value of each grid point and construct a three-dimensional anomaly matrix It should be noted that, at time t, the spatial anomaly degree and the temporal anomaly degree are combined to calculate the comprehensive anomaly value, which is obtained by assigning the same weight to the spatial anomaly degree and the temporal anomaly degree. Set a threshold for the comprehensive outlier value, and mark the grid points whose comprehensive outlier values ​​are greater than the threshold as abnormal grids; The threshold of the composite outlier is obtained by calculating the distribution of historical outliers and selecting a high percentile (e.g., the 95th percentile); The prefabricated component ID in the BIM model is associated with the trajectory data of the UWB device to find the equipment or personnel that enter or approach the abnormal grid during the abnormal time period as the source of the abnormality.

[0033] It should be noted that by reflecting abnormal conditions during the construction process and discovering potential problems in a timely manner, the cause of the abnormality can be quickly located by comparing it with historical data and combining it with BIM models and UWB device data, providing strong support for subsequent targeted measures to repair the abnormal state.

[0034] As an optional embodiment, the specific steps of said S5 in repairing the abnormal state in the construction phase according to the adjusted path are as follows: According to the three-dimensional anomaly matrix Set the starting state S0 and the target state ST; it should be noted that,

[0035] A(x, y, z, t): The three-dimensional anomaly matrix at the current moment, representing the comprehensive anomaly degree of each grid point in the construction site, obtained by fusing the spatial anomaly degree and the temporal anomaly degree; : The observed resource demand gradient indicates the rate of change of resource demand in space, reflecting the spatial unevenness of resource demand; : The delay time of the current construction phase relative to the planned progress, used to measure the degree of lag in the construction progress; The target state is a desired construction state, which is defined as follows: ; : The expected abnormality level does not exceed 10% of the abnormality threshold, where It is the threshold used to judge abnormalities. It can be determined through statistical analysis of historical data and is the average value of historical data. This condition ensures that abnormal conditions during the construction process are effectively controlled and the construction status is close to the normal level. |∇D_obs-∇D_pred|< : The difference between the expected resource demand gradient and the predicted gradient is less than a set threshold , The value of can be determined based on the fluctuation range of the resource demand gradient in historical data; Constructing the decision process, action set , define a reward function to evaluate the immediate reward brought by executing action a in state s; It should be noted that the decision-making process specifically defines three main actions that can be taken during the anomaly repair process: Equipment adjustment: Adjustment of construction equipment, including equipment maintenance, rescheduling or replacement of equipment types, to improve equipment efficiency or adapt to new construction requirements; Material reallocation: Reallocate construction materials, including urgently allocating materials from inventory to areas with high demand, or optimizing material distribution plans to ensure timely and reasonable material supply; Process optimization: Optimize the construction process, including adjusting the construction sequence, merging or splitting construction tasks, adding or reducing construction teams, etc., to improve construction efficiency and reduce construction delays; It should also be noted that, in this embodiment, a reward function is defined to evaluate the immediate reward brought by executing action a in state s, and its expression is:

[0036] : Indicates the negative impact of abnormality on rewards, where It is a penalty coefficient used to measure the weight of the impact of the degree of abnormality on the construction process. Its value can be determined based on the historical impact data of abnormal situations on construction costs, progress and quality. The higher the degree of abnormality, the lower the reward.

[0037] : Indicates the effect of the time required to perform action a on the reward, where is the time cost coefficient, which is used to measure the cost weight of time in the abnormality repair process. Its value can be set according to the tightness of the construction project schedule. t(a) is the estimated time required to perform action a. The longer the time, the lower the reward. The algorithm is used to learn the optimal action selection under different states and obtain the optimal repair path from the starting state to the target state.

[0038] It should be noted that the specific steps are to create a Q table to store the Q values ​​corresponding to different combinations of states (S) and actions (A). Initially, all Q values ​​are set to 0. The rows of the Q table represent states and the columns represent actions. Set the abnormal state of the current construction phase to the current state Scurrent=S0; In state Scurrent, select an action a from the action set, randomly select an action with a fixed probability to explore the environment, and select the action with the largest current Q value with probability 1-fixed to utilize existing knowledge; Execute the action and observe the new state and reward, execute the selected action a, observe the new state Snew and immediate reward R(Scurrent,a) fed back by the environment; According to the Q-learning update formula, update the value of Q(Scurrent,a) in the Q table:

[0039]

[0040] in: The learning rate controls the degree to which new information replaces old information. It is usually between 0 and 1 and can be adjusted based on experimental results. In this embodiment, it is set to 0.5. is a discount factor used to balance the weights of immediate rewards and future rewards. It is usually between 0 and 1 and can be adjusted according to the importance attached to future rewards. In this embodiment, it is set to 0.5; Update the current state to the new state: Scurrent=Snew; If the current state satisfies the target state ST or reaches the maximum number of iterations, the learning process is terminated; otherwise, it returns to continue learning; Extract the optimal strategy from the updated Q-table, that is, for each state, select the action with the largest Q value as the optimal action, forming the optimal repair path from the starting state to the target state; It should be noted that intelligently selecting the optimal repair action based on the current construction status and abnormal conditions to achieve an effective transition from the abnormal state to the target state can improve the efficiency and accuracy of abnormal repair and ensure the smooth implementation of the construction project.

[0041] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A data governance and decision-making method for construction enterprises based on a lake-warehouse integrated architecture, characterized in that: The following steps are involved: S1. Collect data and information about the construction phase of a project by the construction company and store it in the data lake. The data warehouse structures all the data. Extract spatial feature information, including Building Information Model (BIM) data and ultra-wideband (UWB) data from the data lake. Then extract service feature information, including the entropy of the current spectrum of construction equipment and the gradient of construction material consumption. The construction phase label is calculated based on the spatial feature information and business feature information. Based on the construction phase label, the construction phase of the project currently under the responsibility of the construction company is determined. S2. Construct a spatiotemporal demand field based on the construction phase of the project currently under the construction enterprise's responsibility, and predict the resource demand at all spatial locations on the construction site during the construction process; S3. Partition the construction site based on the peak points of the spatiotemporal demand field as seed points, including physical partitioning and digital twin partitioning; When S1 determines that the construction phase of the project currently under the charge of the construction enterprise has changed, S2 is used to reconstruct the demand field and re-partition the project based on the reconstructed demand field to obtain the partitioning result. S4, receiving the partitioning results of S3, obtains the resource demand data of the construction phase of different partitions, and makes corresponding decisions based on the comparison results of the spatiotemporal demand field data; S5. Obtain the resource demand predicted in S2 and the data from the data lake, and calculate the abnormal field during the construction phase based on the acquired data; After obtaining the abnormal field, the abnormal state of the current construction phase is taken as the starting point, and the expected state of the construction phase is set as the end point to obtain the adjustment path, and the abnormal state of the construction phase is repaired according to the adjustment path.

2. The data management and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 1 is characterized by: The specific working steps of S1 are as follows: The spatial feature information extraction specifically includes: encoding the building information model BIM coordinates to obtain BIM coordinates , collect the trajectory vector G of the ultra-wideband device and fuse it to express it as ; The business feature information extraction specifically includes: the construction equipment is specifically a concrete mixer, and the entropy value of the current spectrum is obtained by collecting the current information of the construction equipment; The construction materials specifically refer to the prefabricated components during the construction phase. The difference between the construction material consumption at two adjacent time points is calculated and divided by the time interval to obtain the average gradient within the time period, which is used as the gradient of construction material consumption.

3. The data management and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 2 is characterized by: The specific steps of calculating the construction stage label in S1 are as follows: First, we build a graph structure and determine the nodes of the graph. We divide the construction phases of the project that the current construction company is responsible for into stages, and use the start and end of each stage as the key points of construction. The feature vector of each key point is composed of spatial feature information and business feature information. Then, the adjacency matrix is ​​constructed to establish the connection relationship between the nodes and form an adjacency matrix; Then initialize the GCN model, define the number of GCN layers, the number of units in each layer, the activation function and other hyperparameters, build the graph convolution layer and classification layer of the GCN model, build the graph convolution layer, input is the node feature matrix and adjacency matrix, and output is the updated node feature matrix; After the last graph convolution layer, a classification layer is added to map node features to the probability distribution of stage labels; The constructed node feature matrix and adjacency matrix are used as input data, and the true stage label corresponding to each node is prepared as the training label; Select the optimizer and loss function to compile the GCN model; Prepare training data, input the training data into the GCN model, perform iterative training, and adjust the weights according to the value of the loss function in each round of training; Input new data into the trained GCN model. For new construction key point data, its spatial features and business features need to be calculated first, and the corresponding graph structure needs to be constructed. The node feature matrix and adjacency matrix of the new data are input into the trained GCN model. The model will output the probability distribution of each node belonging to different stage labels. Based on the probability distribution, the stage label with the highest probability is selected as the predicted stage label output for the construction key point.

4. The data management and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 2 is characterized by: The specific steps of S2 in constructing the space-time demand field are as follows: S21. Divide the time and space of the current project of the construction enterprise into a three-dimensional grid, and set the grid size to unit grid; S22. Define the resource demand calculation formula and construct the time-space demand field. The time is determined by the change rate of the construction stage and the spatial distribution of resources. When the grid Resource demand at ; The specific calculation formula is as follows: ； in Representation Grid exist The resource demand at the moment, in standard man-day equivalents, is the rate of change during the construction phase; For the Basic demand weight of class resources; For the moment The center coordinates of the class resource spatial distribution heat map; and is the regression coefficient, which is used to balance the effects of the change rate of the construction stage and the spatial distribution of resources on the resource demand; is a minimum positive constant used to prevent the denominator from having a zero value; S23. When the construction phase of the construction site changes according to the predicted phase label, a phase change instruction is triggered, and the demand field reconstruction process of S22 is repeated to update the resource demand field at fixed time intervals.

5. The data management and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 4 is characterized by: The S2 further includes S24, which is used to modify the resource demand calculation formula according to the business feature information and spatial feature information extracted in real time by S1: S24. Adjust the basic demand weights of various resources based on the business feature information extracted in real time ; According to the extracted spatial feature information, the coordinates of the resource thermal center are Make corrections.

6. The data governance and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 4 is characterized by: The specific steps of S3 are as follows: S31, identifying peak points in the demand field, performing DBSCAN clustering on the peak points, and generating k seed point sets; S32, dividing the physical construction area with the seed point set as the center to complete the physical partition; Specifically, based on the seed point set, the construction site is divided into k physical construction areas using the Voronoi diagram algorithm. Each seed point corresponds to a Voronoi region, which contains all the grid cells closest to the seed point. Regularly calculate the difference in resource requirements between adjacent partitions. For each adjacent partition, calculate the ratio of the difference in resource requirements to the maximum resource requirement: When the ratio exceeds the set value, the partition boundary is adjusted; The adjustment method is to re-divide the grid cells close to the boundary of the high-demand zone to the low-demand zone until the ratio is lower than the set value or cannot be adjusted any further; S33. Map the physical partitions in the BIM model, and add an additional data layer containing resource requirements D(x, y, z, t) to complete the digital twin partitioning. When the demand field is reconstructed, recalculate the seed points and update the partitions.

7. The data governance and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 6 is characterized by: The specific working steps of S31 are as follows: Extract the resource demand D(x,y,z,t) of all grid cells from the demand field and calculate its mean and standard deviation ; Traverse each grid cell in the demand field and find the grid cell that satisfies D(x,y,z,t)> +2 The grid points of , these points are the candidate peak points; Steps: Use the DBSCAN clustering algorithm to cluster the candidate peak points; The candidate peak points are divided into multiple clusters, each cluster represents a high-demand area, and the center point is selected from each cluster as the seed point. The seed point set is expressed as .

8. The data management and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 6 is characterized by: The specific steps of S4 are as follows: Establish a demand feature vector for each partition k, specifically calculating the average demand of the partition, the maximum demand of the partition, the demand gradient and the material turnover time; Setting thresholds 、 and ; When the maximum value of the demand gradient exceeds the threshold When the ratio of the maximum demand to the average demand of the partition exceeds the threshold, the device scheduling instruction is triggered. When the material turnover time exceeds the threshold, the material replenishment instruction is triggered. , triggers the logistics optimization instruction.

9. The data management and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 3 is characterized by: The specific working steps of S5 are as follows: Obtain the predicted resource demand data from step S2, including the predicted resource demand for each grid point ; Obtain real-time resource demand data from the data lake ; For each grid point, the relative error between the observed resource demand and the predicted resource demand is calculated, and then the average value of all grid points is calculated to obtain the spatial anomaly degree. ; Calculate the absolute error of the time derivative of resource demand and integrate it within the set time period to obtain the time anomaly degree ; Combine the spatial anomaly and temporal anomaly to calculate the comprehensive anomaly value of each grid point and construct a three-dimensional anomaly matrix ; The prefabricated component ID in the BIM model is associated with the trajectory data of the UWB device to find the equipment or personnel that enter or approach the abnormal grid during the abnormal time period as the source of the abnormality.

10. The data management and decision-making method for construction enterprises based on the lake-warehouse integrated architecture according to claim 9 is characterized by: The specific steps of S5 in repairing the abnormal state during the construction phase according to the adjusted path are as follows: According to the three-dimensional anomaly matrix Set the starting state S0 and the target state ST; Constructing the decision process, action set , define a reward function to evaluate the immediate reward brought by executing action a in state s; The algorithm is used to learn the optimal action selection under different states and obtain the optimal repair path from the starting state to the target state.

Images