A cost auxiliary accounting and risk prediction system for construction engineering

Abstract

This invention relates to the field of information management technology for construction projects, and discloses a cost-assisted accounting and risk prediction system for construction projects. The system includes a multi-source heterogeneous data sensing and access module, an atomic work unit construction module, a topology network adaptive reconstruction module, a cost sensitivity cascade analysis module, and a risk damping and reverse potential energy control module. The invention achieves physical-level accounting by using the atomic work unit construction module to calculate real-time reconstruction costs based on the measured values ​​of physical quantities of resource consumption and the mapping between real-time library unit prices. The topology network adaptive reconstruction module dynamically adjusts the network structure. The cost sensitivity cascade analysis module quantifies the impact of upstream node schedule deviations on downstream node costs, simulates risk cascade diffusion, and predicts accumulated risk energy. The risk damping and reverse potential energy control module generates virtual resource tokens for reverse transmission, and executes physical access control when the reverse locking signal strength exceeds a threshold, achieving automated risk blocking.

Description

Technical Field

[0001] This invention relates to the field of information management technology for construction projects, specifically a cost-aid accounting and risk prediction system for construction projects. Background Technology

[0002] Construction project management involves a vast resource allocation system and complex capital flow logic. Traditional construction cost accounting mainly relies on phased financial statement statistics and manual verification of bills of quantities. This manual verification method leads to a misalignment between financial value data and physical progress data at the construction site in terms of time. The lagging cost data cannot reflect the instantaneous resource consumption generated by physical operations such as hoisting and material transportation. There is a disconnect between the actual input at the physical operation level and the book cost accounting results, making it difficult for the project management system to obtain cost feedback immediately when physical operations occur, and making it impossible to establish a real-time mapping relationship between physical consumption and value loss.

[0003] Construction sites involve not only cash flow but also complex topological networks composed of multiple processes. Existing project risk management methods focus on monitoring the progress of individual work nodes or providing safety warnings for isolated areas. This isolated monitoring approach ignores the nonlinear coupling characteristics of the construction process network. Schedule deviations generated by upstream work nodes can be transmitted to downstream work nodes through resource dependencies, and in the process of transmission, they can be transformed into cost increments or risk accumulation for downstream work nodes. The lack of quantitative analysis of cascading effects makes it impossible for management systems to predict the global impact of small disturbances spreading in the topological network and their cumulative destructive power on end nodes.

[0004] Existing digital construction management platforms mostly focus on one-way data collection and visualization. When data display platforms identify potential risk trends or signs of cost overruns, they mainly rely on manual hierarchical reporting and administrative orders for intervention. The feedback path of information flow from the risk identification end to the on-site execution end is too long, and there is a lack of automated blocking mechanisms that directly affect physical equipment. When the risk energy accumulates to a critical state, the machinery and personnel on the work site can still continue to generate new workload, leading to an increase in the probability of risk accidents and the continuous investment of ineffective costs, making it impossible to achieve closed-loop management of risk prediction and physical control. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a cost-aided accounting and risk prediction system for construction projects, which solves the problems of traditional construction project cost accounting lagging behind the physical progress of construction sites and lacking the ability to predict global risks based on the cascading effect of process topology networks.

[0006] To achieve the above objectives, the present invention provides a cost-aid accounting and risk prediction system for construction engineering, comprising: a multi-source heterogeneous data sensing and access module, an atomic operation unit construction module, a topology network adaptive reconstruction module, a cost sensitivity cascade analysis module, and a risk damping and reverse potential energy control module.

[0007] The multi-source heterogeneous data sensing and access module is deployed at the construction site and configured to collect data from crane load sensors, personnel identification gates, material weighbridges, and environmental parameter monitoring stations. This module utilizes an edge computing gateway to receive these data, performs time alignment processing on the data from the crane load sensors, personnel identification gates, material weighbridges, and environmental parameter monitoring stations using an internally integrated clock source, and cleans the data to output a standardized spatiotemporal physical state data stream.

[0008] The atomic job unit construction module is connected to the multi-source heterogeneous data sensing and access module and configured to receive spatiotemporal physical state data streams. The atomic job unit construction module uses preset spatiotemporal constraint verification logic to identify physical job events that conform to preset rules, and generates atomic job units based on the physical job events.

[0009] During the generation of atomic work units, the atomic work unit construction module parses the resource types involved in the atomic work unit, obtaining the measured physical consumption value and the real-time unit price of each resource type. Combining the timestamp, spatial coordinate vector, and action feature vector of the physical work event, the atomic work unit construction module verifies the physical work event using a spatiotemporal verification logic function and a Kronecker function. After successful verification, the atomic work unit construction module performs dynamic cost aggregation based on the mapping relationship between the measured physical consumption value of the resource and the real-time unit price, accumulating the product of the measured physical consumption value and the real-time unit price of each resource type to calculate the real-time reconstruction cost, thus achieving cost-assisted accounting at the physical work level.

[0010] The adaptive topology reconstruction module connects to the atomic job unit construction module and is configured to store and manage the directed weighted engineering topology network. The adaptive topology reconstruction module receives atomic job units in real time and calculates the arithmetic mean of the key attribute observations of the atomic job units belonging to the job node within the current time window, as well as the average of the squares of the differences between each key attribute observation and the arithmetic mean, to obtain the key attribute dispersion. When the key attribute dispersion exceeds a preset structural splitting threshold, the adaptive topology reconstruction module splits the job node into multiple child nodes and distributes the incoming and outgoing edge weights of the original job node to the multiple child nodes according to the proportion of total resource consumption contained in each child node.

[0011] The adaptive topology reconstruction module filters adjacent job node pairs whose key attribute dispersion is lower than a preset aggregation threshold, and calculates the cross-correlation coefficient of job attributes between adjacent job node pairs within a historical time step. When the cross-correlation coefficient of job attributes meets the preset correlation judgment condition, the adaptive topology reconstruction module merges adjacent job node pairs into a single aggregate node.

[0012] The cost sensitivity cascade analysis module is connected to the topology network adaptive reconstruction module and configured to calculate the cost sensitivity weights between work nodes based on the adjusted directed weighted engineering topology network. The cost sensitivity cascade analysis module uses linear regression analysis to determine the partial derivative of the cost of the downstream work node with respect to the schedule deviation of the upstream work node, and analyzes the resource sets occupied by the upstream and downstream work nodes. It calculates the nonlinear coupling factor based on the ratio of the intersection to the union of the resource sets. The cost sensitivity cascade analysis module multiplies the partial derivative with the coefficient corrected by the nonlinear coupling factor to obtain the cost sensitivity weight.

[0013] The cost sensitivity cascade analysis module transforms the directed weighted engineering topology network into a global risk transmission adjacency matrix and constructs an initial risk disturbance vector containing the actual schedule deviations and cost deviations of each work node. The module then inverts the matrix obtained by subtracting the product of the risk transmission attenuation coefficient and the global risk transmission adjacency matrix from the identity matrix. This inverted matrix is ​​then multiplied by the initial risk disturbance vector to calculate the cumulative risk energy distribution vector after the cascade effect stabilizes. Finally, the module calculates the cumulative risk energy of the terminal nodes, thus enabling risk prediction under the cascade effect.

[0014] The risk damping and reverse potential energy control module is connected to the cost sensitivity cascade analysis module. Based on the reverse connection relationship of the directed weighted engineering topology network, the risk damping and reverse potential energy control module identifies upstream transport edges pointing to high-risk end nodes. Based on the accumulated risk energy of the starting node of the upstream transport edge, the cost sensitivity weight of the connecting edge, the available risk buffer resources of the system, and the estimated intervention cost, the module calculates the damping intervention effectiveness index and selects paths with a damping intervention effectiveness index higher than a preset intervention initiation threshold as the optimal damping control path.

[0015] The risk damping and reverse potential energy control module generates virtual resource tokens based on accumulated risk energy. These virtual resource tokens encapsulate reverse potential energy values. The reverse potential energy values ​​are calculated based on the difference between the accumulated risk energy of downstream work nodes and the safety threshold along the optimal damping control path, the resource response sensitivity coefficient of upstream work nodes, and the negative exponential decay factor of the topological path distance between upstream and downstream work nodes.

[0016] The risk damping and reverse potential energy control module transmits the virtual resource token in reverse along the directed weighted engineering topology network to the currently executing job node and calculates the reverse locking signal strength acting on the currently executing job node. When the reverse locking signal strength exceeds a preset forced circuit breaker threshold, the risk damping and reverse potential energy control module sends a permission locking command to the atomic job unit construction module to execute physical permission locking and stop generating new atomic job units. The duration of the physical permission locking hardware lock is determined based on the ratio of the reverse potential energy value to the forced circuit breaker threshold, the risk response intensity index, and the baseline cooling time unit.

[0017] This invention provides a cost-assistance accounting and risk prediction system for construction projects. It has the following beneficial effects: 1. This invention collects multi-dimensional physical data from the construction site through a multi-source heterogeneous data sensing and access module, and transforms the physical data into atomic operation units using an atomic operation unit construction module. The atomic operation unit construction module performs dynamic cost aggregation based on the mapping relationship between the measured value of physical quantities of resource consumption and the real-time unit price in the library, calculates real-time reconstruction costs, refines the cost accounting granularity to the level of physical operation events, establishes a real-time correlation between physical progress and value consumption, solves the problem of traditional construction engineering cost accounting lagging behind construction progress, and realizes cost-assisted accounting at the physical operation level.

[0018] 2. This invention uses a topology network adaptive reconstruction module to dynamically adjust the directed weighted engineering topology network based on the key attribute dispersion of atomic work units, thus truly reflecting the dynamic organizational structure of the construction process. The cost sensitivity cascade analysis module quantifies the impact of upstream work node schedule deviations on downstream work node costs. By simulating the risk cascade diffusion process through the operation of the global risk transmission adjacency matrix, it overcomes the limitations of single-point risk monitoring, quantifies the nonlinear risk transmission effect in complex networks, predicts the cumulative risk energy of end nodes, and realizes risk prediction under cascade effects.

[0019] 3. This invention constructs a reverse feedback path from the risk prediction end to the operation execution end through a risk damping and reverse potential energy control module. It uses virtual resource tokens to transmit reverse potential energy and directly sends an access control command when the reverse locking signal strength exceeds the safety threshold. This transforms the risk prediction results at the data level into mandatory intervention measures at the physical level. It can automatically block the generation of new atomic operation units before the risk energy accumulates to the critical point, thus preventing the further expansion of risk accidents and cost overruns. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall system architecture and hardware environment of the present invention; Figure 2 This is a flowchart illustrating the construction process of the data access and operation unit of the present invention; Figure 3 This is a schematic diagram of the adaptive topology reconstruction of the present invention; Figure 4 This is a schematic diagram illustrating the cost sensitivity and risk transmission of the present invention; Figure 5 This is a flowchart of the risk damping and reverse control of the present invention.

[0021] The system comprises the following components: 10. Multi-source heterogeneous data sensing and access module; 11. Physical state acquisition unit; 12. Data cleaning and alignment unit; 20. Atomic operation unit construction module; 21. Spatiotemporal constraint verification unit; 22. Dynamic cost aggregation unit; 30. Topology network adaptive reconstruction module; 31. Discrete variance analysis unit; 32. Node fission and fusion execution unit; 40. Cost sensitivity cascade analysis module; 41. Elastic weight calculation unit; 42. Risk energy transmission unit; 50. Risk damping and reverse potential energy control module; 51. Damping path solution unit; 52. Virtual token generation and reverse transmission unit; 53. Physical permission locking execution unit; 60. Front-end sensing device group; 61. Lifting machinery load sensor; 62. Personnel identification gate; 63. Material weighing scale; 64. Environmental parameter monitoring station; 70. Edge computing gateway; 80. Data transmission network; 90. Central processing server; 91. Processor; 92. Memory; 93. Visualization terminal. Detailed Implementation

[0022] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] See attached document Figure 1 This invention provides a cost-aid accounting and risk prediction system for construction engineering, including a multi-source heterogeneous data sensing and access module 10, an atomic operation unit construction module 20, a topology network adaptive reconstruction module 30, a cost sensitivity cascade analysis module 40, and a risk damping and reverse potential energy control module 50.

[0024] The multi-source heterogeneous data sensing and access module 10 is deployed at the construction site of the building project. The multi-source heterogeneous data sensing and access module 10 collects data from the load sensor of the lifting equipment, the personnel positioning gate, the material weighbridge, and the environmental monitoring data. The multi-source heterogeneous data sensing and access module 10 cleans and aligns the collected data with time, and outputs a standardized spatiotemporal physical state data stream.

[0025] The atomic job unit construction module 20 is signal-connected to the multi-source heterogeneous data sensing and access module 10. The atomic job unit construction module 20 receives spatiotemporal physical state data streams. The atomic job unit construction module 20 uses preset spatiotemporal constraint verification logic to identify valid physical job events and generates atomic job units based on the valid physical job events.

[0026] Atomic work unit construction module 20 Calculation Real-time reconfiguration cost of individual atomic work units The calculation formula is as follows: ; in: Indicates the first The real-time reconfiguration cost of an atomic work unit Indicates the first The total number of resource types involved in each atomic work unit; Indicates the first In the atomic work unit, the first Measured physical consumption of the resource type; Indicates the first Real-time unit price of the resource category; Indicates the first The timestamp of the physical event that occurred for each atomic work unit; Indicates the first The spatial coordinate vector of the physical event occurring in each atomic work unit; Indicates the first Action feature vectors of physical events corresponding to each atomic work unit; This represents a spatiotemporal verification logic function; This represents the Kronecker function.

[0027] The topology network adaptive reconstruction module 30 is signal-connected to the atomic job unit construction module 20. The topology network adaptive reconstruction module 30 stores the directed weighted engineering topology network. The topology network adaptive reconstruction module 30 receives atomic job units in real time. The topology network adaptive reconstruction module 30 dynamically adjusts the structure of job nodes in the directed weighted engineering topology network according to the discreteness of the key attribute of the atomic job unit.

[0028] The adaptive topology reconstruction module 30 calculates the dispersion of job nodes. The calculation formula is as follows: ; in: This indicates the total number of atomic job units belonging to the job node within the current time window; Indicates the first Key attribute observations for each atomic work unit; This represents the arithmetic mean of the key attribute observations of all atomic job cells within a job node, when the dispersion... When the number of nodes exceeds the preset structural splitting threshold, the topology network adaptive reconstruction module 30 splits the work node into multiple child nodes.

[0029] The cost sensitivity cascade analysis module 40 is connected to the topology network adaptive reconstruction module 30 by signal. The cost sensitivity cascade analysis module 40 calculates the edge weights between operation nodes based on the adjusted directed weighted engineering topology network.

[0030] The cost sensitivity cascade analysis module 40 calculates the cost sensitivity weights of work nodes pointing to other work nodes. The calculation formula is as follows: ; in: Indicates the work node Cost For work nodes Construction period deviation The partial derivatives; Indicates the work node With work nodes The nonlinear coupling factor between them.

[0031] The risk damping and reverse potential energy control module 50 is connected to the cost sensitivity cascade analysis module 40. The risk damping and reverse potential energy control module 50 calculates the risk energy of the end node in the directed weighted engineering topology network according to the cost sensitivity weight. The risk damping and reverse potential energy control module 50 generates a virtual resource token based on the risk energy and transmits the virtual resource token in reverse to the currently executing job node.

[0032] The risk damping and reverse potential energy control module 50 calculates the strength of the reverse locking signal acting on the current execution node. The calculation formula is as follows: ; in: Indicates the control gain coefficient; This represents the set of paths from the current execution node to all potentially high-risk end nodes; This represents the predicted overspending risk energy at the end node; Representing a path Directed edges in; Represents a directed edge The damping attenuation coefficient is set to a value greater than 1 to characterize the signal loss that occurs as the propagation path length increases. The value of the damping attenuation coefficient is proportional to the spatial physical distance between the working nodes.

[0033] The risk damping and reverse potential energy control module 50 is connected to the atomic job unit construction module 20. When the reverse locking signal strength exceeds the preset safety threshold, the risk damping and reverse potential energy control module 50 sends an access control command to the atomic job unit construction module 20. The atomic job unit construction module 20 responds to the access control command and stops generating new atomic job units.

[0034] See attached document Figure 1 The hardware environment of a cost-aid accounting and risk prediction system for construction projects includes a front-end sensing device group 60, an edge computing gateway 70, a data transmission network 80, and a central processing server 90.

[0035] The front-end sensing device group 60 is physically deployed in the work area of ​​the construction site. The front-end sensing device group 60 includes a crane load sensor 61, a personnel identification gate 62, a material weighing scale 63, and an environmental parameter monitoring station 64. The crane load sensor 61 is connected to the boom or hook assembly of the crane through a signal cable to collect the load weight analog signal during the lifting operation. The personnel identification gate 62 is fixedly installed at the physical boundary entrance and exit of the work area to read the radio frequency identification tag data of the personnel entering and exiting the work area. The material weighing scale 63 is embedded in the ground of the material transportation channel of the construction site to measure the mass data of the transport vehicles and the loaded materials. The environmental parameter monitoring station 64 is distributed at the perimeter of the construction site and the high-altitude work surface to collect the air temperature, air humidity and wind speed values ​​of the site.

[0036] The edge computing gateway 70 is physically connected to the crane load sensor 61, the personnel identification gate 62, the material weighing scale 63, and the environmental parameter monitoring station 64. The edge computing gateway 70 is equipped with an analog-to-digital conversion circuit and a digital signal processing circuit to receive the raw data signals output by the front-end sensing device group 60 and to perform noise filtering and format standardization processing on the raw data signals. The edge computing gateway 70 integrates a high-precision clock source to add a unified timestamp to the processed data to ensure the timing alignment of multi-source data.

[0037] The data transmission network 80 is communicatively connected to the edge computing gateway 70. The data transmission network 80 is used to establish a high-bandwidth bidirectional data transmission channel between the edge computing gateway 70 and the central processing server 90.

[0038] The central processing server 90 is communicatively connected to the data transmission network 80. The central processing server 90 includes a processor 91 and a memory 92. The memory 92 is used to store computer program instructions and engineering project databases. The processor 91 is used to call and execute the computer program instructions stored in the memory 92 to run the logical functions of the multi-source heterogeneous data sensing and access module 10, the atomic operation unit construction module 20, the topology network adaptive reconstruction module 30, the cost sensitivity cascade analysis module 40, and the risk damping and reverse potential energy control module 50. The central processing server 90 is connected to a visualization terminal 93 through a data interface. The visualization terminal 93 is used to present dynamic topology view and risk warning information.

[0039] See attached document Figure 2The multi-source heterogeneous data sensing and access module 10 includes a physical state acquisition unit 11. The physical state acquisition unit 11 establishes digital communication connections with the crane load sensor 61, the personnel identification gate 62, the material weighing scale 63, and the environmental parameter monitoring station 64, respectively. The physical state acquisition unit 11 is equipped with a multi-channel data acquisition interface for receiving raw physical signals from different hardware devices in parallel.

[0040] The physical status acquisition unit 11 reads mechanical operation status data from the crane load sensor 61 via a serial communication protocol. The mechanical operation status data includes the instantaneous load value of the crane hook, the horizontal rotation angle value of the crane boom, and the luffing radius value of the crane trolley. The physical status acquisition unit 11 reads personnel access log data from the personnel identification gate 62 via a network socket interface. The personnel access log data includes the unique RFID tag identification code of the operator, the time point when the operator passed through the gate, and the grid area code where the gate is located. The physical status acquisition unit 11 reads material flow data from the material weighing scale 63 via an industrial bus interface. The material flow data includes the total mass value of the transport vehicle when fully loaded, the empty tare weight value of the transport vehicle, and the material type identification code. The physical status acquisition unit 11 reads on-site environmental data from the environmental parameter monitoring station 64 via an analog input interface. The on-site environmental data includes the air temperature value, relative humidity value, and instantaneous wind speed value of the construction site.

[0041] The physical state acquisition unit 11 encapsulates the mechanical operation status data, personnel access log data, material flow data and on-site environmental data into a raw physical state vector according to a unified time base. The raw physical state vector aggregates multi-dimensional physical parameters at the same sampling time. The physical state acquisition unit 11 transmits the raw physical state vector to the subsequent processing unit of the multi-source heterogeneous data sensing and access module 10.

[0042] The multi-source heterogeneous data sensing and access module 10 includes a data cleaning and alignment unit 12, which is communicatively connected to the physical state acquisition unit 11. The data cleaning and alignment unit 12 receives the original physical state vector output by the physical state acquisition unit 11. The data cleaning and alignment unit 12 stores a preset system synchronization clock frequency and generates a series of discrete standard synchronization time nodes according to the preset system synchronization clock frequency.

[0043] The data cleaning and alignment unit 12 performs linear interpolation operations on the mechanical operation status data, material flow data and on-site environmental data contained in the original physical state vector. The data cleaning and alignment unit 12 calculates the alignment value of each physical parameter under the standard synchronization time node according to the time difference ratio between the standard synchronization time node and the adjacent original sampling time, thereby mapping asynchronous data with different sampling frequencies to the standard synchronization time node.

[0044] The data cleaning and alignment unit 12 stores the coordinate system transformation matrix of the construction project. The data cleaning and alignment unit 12 parses the spatial position coordinate information contained in the original physical state vector. The data cleaning and alignment unit 12 uses the coordinate system transformation matrix of the construction project to uniformly map the spatial position coordinate information of the personnel identification gate 62 and the spatial position coordinate information of the lifting machinery load sensor 61 to the three-dimensional construction project coordinate system. The data cleaning and alignment unit 12 packages the data after linear interpolation and coordinate mapping into a standardized spatiotemporal physical state data stream. The data cleaning and alignment unit 12 sends the standardized spatiotemporal physical state data stream to the atomic operation unit construction module 20.

[0045] The atomic operation unit construction module 20 includes a spatiotemporal constraint verification unit 21. The spatiotemporal constraint verification unit 21 is unidirectionally connected to the data cleaning and alignment unit 12. The spatiotemporal constraint verification unit 21 receives standardized spatiotemporal physical state data streams. The spatiotemporal constraint verification unit 21 has a built-in engineering operation logic judgment rule library. The engineering operation logic judgment rule library contains effective operation space range parameters and effective mechanical action threshold parameters corresponding to different types of work.

[0046] The spatiotemporal constraint verification unit 21 parses the personnel coordinate data, mechanical action data, and material change data in the standardized spatiotemporal physical state data stream. For each sampling time point, the spatiotemporal constraint verification unit 21 calculates the spatiotemporal coupling judgment value of multi-dimensional features. The calculation logic is as follows: only when the instantaneous load change rate uploaded by the crane load sensor 61 exceeds the preset effective load action threshold, the quality change value of material flow exceeds the preset minimum material consumption measurement threshold, and the job matching confidence recorded by the personnel identification gate 62 exceeds the preset job verification pass threshold, combined with the spatial proximity of the operator and the construction machinery, an effective spatiotemporal coupling judgment value is generated. The spatiotemporal constraint verification unit 21 compares the calculated spatiotemporal coupling judgment value with the preset event trigger threshold.

[0047] The spatiotemporal constraint verification unit 21 compares the calculated spatiotemporal coupling judgment value with the preset event triggering threshold. When the spatiotemporal coupling judgment value is greater than the preset event triggering threshold, the spatiotemporal constraint verification unit 21 determines that the current physical state constitutes a valid physical operation event and sends the spatiotemporal coordinate data and resource consumption data corresponding to the valid physical operation event to the dynamic cost collection unit 22 inside the atomic operation unit construction module 20.

[0048] The atomic work unit construction module 20 includes a dynamic cost collection unit 22, which is connected to the spatiotemporal constraint verification unit 21. The dynamic cost collection unit 22 is equipped with a local memory, which stores a real-time unit price database for construction engineering resources. The real-time unit price database for construction engineering resources includes the real-time market unit price corresponding to the type of construction materials, the real-time rental unit price corresponding to the model of construction machinery, and the real-time hourly unit price corresponding to the type of work.

[0049] The dynamic cost collection unit 22 receives the valid physical operation events output by the spatiotemporal constraint verification unit 21. The dynamic cost collection unit 22 parses the set of resource consumption physical quantities contained in the valid physical operation events. Based on the resource type identifier in the set of resource consumption physical quantities, the dynamic cost collection unit 22 performs a search operation in the real-time unit price database of building engineering resources to obtain the matching resource unit price value.

[0050] Dynamic cost collection unit 22 executes for the first The instantaneous cost calculation task of each atomic work unit, and the dynamic cost collection unit 22 accumulates the cost of the first atomic work unit. The product of the physical consumption value of each type of resource consumed by each atomic work unit and the system library unit price value of that type of resource at the current time point yields the product of the physical consumption value of each type of resource. The dynamic cost aggregation unit 22 generates an atomic operation unit data object containing the instantaneous aggregation cost, the timestamp of the physical event, the spatial coordinates of the physical event, and the set of physical quantities of resource consumption. The dynamic cost aggregation unit 22 assigns a globally unique operation unit identification code to the atomic operation unit data object and sends the instantiated atomic operation unit data object to the topology network adaptive reconstruction module 30.

[0051] See attached document Figure 3 The topology network adaptive reconstruction module 30 includes a discrete variance analysis unit 31. The discrete variance analysis unit 31 establishes a data transmission channel with the atomic job unit construction module 20. The discrete variance analysis unit 31 accesses and reads the current directed weighted engineering topology network structure stored in the system database. The discrete variance analysis unit 31 is configured with job data statistical observation time window parameters.

[0052] Discrete variance analysis unit 31 receives atomic job unit data objects output by atomic job unit construction module 20 in real time. Based on the job affiliation identifier in the atomic job unit data object, it maps the atomic job unit data object to a specific job node in the directed weighted engineering topology network. Within the job data statistical observation time window, discrete variance analysis unit 31 extracts the key attribute observation values ​​of all atomic job units belonging to the job node. The key attribute observation values ​​include the resource consumption rate value and job output efficiency value per unit time.

[0053] Discrete variance analysis unit 31 performs the task of calculating the internal job heterogeneity of job nodes. Based on the key attribute observations of all atomic job units belonging to the job node within the job data statistical observation time window, discrete variance analysis unit 31 uses statistical variance algorithm to calculate the dispersion of job nodes, quantifies the fluctuation of the internal attribute distribution of job nodes, and compares the calculated dispersion with the preset structural splitting threshold to generate a node stability analysis report containing job node identifiers and dispersion values. The node stability analysis report is then sent to the node fission and fusion execution unit 32 inside the topology network adaptive reconstruction module 30.

[0054] The topology network adaptive reconstruction module 30 includes a node fission and fusion execution unit 32, which is data-connected to the discrete variance analysis unit 31. The node fission and fusion execution unit 32 is configured with a graph database management interface, which is used to perform dynamic addition and deletion operations on the physical structure of the directed weighted engineering topology network.

[0055] The node fission and fusion execution unit 32 receives the node stability analysis report and reads the dispersion of the working nodes in the report. When the dispersion is greater than the preset structural splitting threshold, the node fission and fusion execution unit 32 triggers the node fission logic, calls the clustering algorithm to group the atomic working unit sets belonging to the working nodes, generates the corresponding child node sets in the directed weighted engineering topology network according to the grouping results, re-indexes the atomic working unit sets to associate them with the child node sets, and inherits the inbound and outbound edge connections of the original working nodes to the child node sets. Specifically, the inheritance method is as follows: calculate the ratio of the total resource consumption of the atomic working units contained in each child node to the total resource consumption of the original working node, and multiply the inbound and outbound edge weights of the original working node by the ratio, thereby distributing the weights to the corresponding child nodes to maintain the local flow conservation before and after network splitting.

[0056] The node fission and fusion execution unit 32 selects adjacent work node pairs whose dispersion values ​​are all lower than the preset aggregation threshold, calculates the cross-correlation coefficient of work attributes between adjacent work node pairs, and calculates the synchronous change trend of adjacent work node behavior by statistically analyzing the ratio of covariance to standard deviation based on the key attribute value sequence of adjacent work nodes within the historical time step. When the cross-correlation coefficient of work attributes is greater than the preset correlation judgment threshold, the node fission and fusion execution unit 32 executes the node fusion logic.

[0057] The node fission and fusion execution unit 32 merges adjacent work node pairs into a single aggregate node, sets the attribute value of the aggregate node to the weighted sum of the attribute values ​​of adjacent work node pairs, and sends the directed weighted engineering topology network after the structural update to the cost sensitivity cascade analysis module 40.

[0058] See attached document Figure 4 The cost sensitivity cascade analysis module 40 includes an elastic weight calculation unit 41. The elastic weight calculation unit 41 establishes a data communication connection with the topology network adaptive reconstruction module 30 and receives the directed weighted engineering topology network containing the latest node structure output by the topology network adaptive reconstruction module 30. The elastic weight calculation unit 41 is configured with an engineering history database reading interface, which is used to access historical operation cost data and historical operation duration data stored in the system background.

[0059] The flexible weight calculation unit 41 traverses each directed edge in the directed weighted engineering topology network, identifies the upstream and downstream operation nodes connected by the directed edges, extracts a set of historical operation period deviation samples similar to the upstream operation nodes from historical operation cost data and historical operation period data, and extracts a set of historical cost change samples similar to the downstream operation nodes from historical operation cost data and historical operation period data. The flexible weight calculation unit 41 uses the least squares method to perform linear regression analysis on the historical operation period deviation sample set and the historical cost change sample set to obtain the marginal cost impact slope. When the number of retrieved historical samples is insufficient to construct a statistical regression model, the flexible weight calculation unit 41 calls the system's preset industry standard quota database and directly reads the benchmark cost period impact coefficient matching the operation type as the marginal cost impact slope.

[0060] The elastic weight calculation unit 41 analyzes the first resource set occupied by the upstream operation node and the second resource set occupied by the downstream operation node, and calculates the nonlinear coupling factor based on the degree of overlap between the first resource set and the second resource set. The calculation logic is to take the ratio of the cardinality of the intersection of the first resource set and the second resource set to the cardinality of the union, and multiply the ratio of the cardinality by the preset resource competition intensity coefficient.

[0061] Elastic weight calculation unit 41 combines the marginal cost impact slope and nonlinear coupling factor to calculate the cost sensitivity weight from upstream operation node to downstream operation node. The calculation formula is as follows: ; in: Indicates downstream operation node Cost For upstream operation nodes Construction period deviation The approximate value of the partial derivative is determined by the slope of the marginal cost impact obtained from the aforementioned linear regression analysis; This indicates the upstream work node obtained from the aforementioned calculation. With downstream operation nodes The nonlinear coupling factor between them.

[0062] The elastic weight calculation unit 41 assigns the calculated cost sensitivity weights to the directed edges connecting the upstream and downstream operation nodes in the directed weighted engineering topology network. After completing the weight assignment operation for all directed edges in the directed weighted engineering topology network, the weighted risk transmission network is generated and sent to the risk energy transmission unit 42 inside the cost sensitivity cascade analysis module 40.

[0063] The cost sensitivity cascade analysis module 40 includes a risk energy transmission unit 42. The risk energy transmission unit 42 establishes a data transmission channel with the elastic weight calculation unit 41. The risk energy transmission unit 42 receives the weighted risk transmission network output by the elastic weight calculation unit 41. The risk energy transmission unit 42 is equipped with a linear algebra operation acceleration processor. The risk energy transmission unit 42 parses all the job nodes and the weights of the directed edges connecting the job nodes contained in the weighted risk transmission network.

[0064] The risk energy transmission unit 42 converts the weighted risk transmission network into a corresponding global risk transmission adjacency matrix. The row index and column index of the global risk transmission adjacency matrix correspond to the job node numbers in the topology network, and the element values ​​of the global risk transmission adjacency matrix correspond to the cost sensitivity weights calculated above.

[0065] The risk energy transmission unit 42 obtains the initial risk disturbance vector at the current moment. The initial risk disturbance vector consists of the actual schedule deviation and cost deviation values ​​of each work node. The risk energy transmission unit 42 uses the global risk transmission adjacency matrix to perform cascade diffusion simulation on the initial risk disturbance vector and calculates the cumulative risk energy distribution vector after the cascade effect stabilizes. The calculation formula is as follows: ; in: The identity matrix has the same dimension as the global risk transmission adjacency matrix; This represents the preset risk transmission attenuation coefficient, which is used to simulate the energy dissipation of risk during the transmission process. This represents the adjacency matrix representing the global risk transmission. The matrix inversion operator is represented by the symbol. This represents the initial risk disturbance vector.

[0066] The risk energy transmission unit 42 extracts the key risk nodes whose cumulative risk energy values ​​exceed the preset safety warning line based on the cumulative risk energy distribution vector, generates a cascaded risk assessment report containing the key risk node number, the cumulative risk energy value of the key risk node, and the risk propagation path, and sends the cascaded risk assessment report to the risk damping and reverse potential energy control module 50.

[0067] See attached document Figure 5 The risk damping and reverse potential energy control module 50 includes a damping path solving unit 51. The damping path solving unit 51 establishes a one-way data receiving channel with the risk energy transmission unit 42. The damping path solving unit 51 is equipped with a risk intervention resource database interface. The risk intervention resource database interface is used to access the emergency fund amount data, the spare mechanical equipment list data, and the standby work team data that can be called on the construction site.

[0068] The damping path solving unit 51 receives the cascaded risk assessment report output by the risk energy transmission unit 42, analyzes the key risk nodes and risk propagation paths marked in the cascaded risk assessment report, constructs a risk reverse tracing graph based on the reverse connection relationship of the directed weighted engineering topology network, and identifies all upstream transmission edges pointing to key risk nodes in the risk reverse tracing graph.

[0069] Damping path solving unit 51 matches potential intervention measures and estimated intervention costs corresponding to upstream transport edges from the risk intervention resource database, and calculates the damping intervention effectiveness index for each upstream transport edge. The calculation formula is as follows: ; in: Indicates the starting node number of the upstream conveying edge; Indicates the termination node number of the upstream conveyor edge; Indicates the starting job node The current accumulated risk energy value; This indicates the connection of the starting work node in the directed weighted engineering topology network. With termination of work node Cost sensitivity weighting; The operator represents the natural logarithm operator; This indicates the total amount of available risk buffer resources in the current construction project account; This indicates blocking or lowering the starting work node. To the end of the operation node The estimated intervention cost required for risk transmission between them.

[0070] The damping path solving unit 51 compares the calculated damping intervention efficiency index with the preset intervention initiation threshold, selects the upstream transmission edge whose damping intervention efficiency index is higher than the preset intervention initiation threshold, and forms the optimal damping control path set. The optimal damping control path set is then sent to the reverse potential energy generation unit inside the risk damping and reverse potential energy control module 50.

[0071] The risk damping and reverse potential energy control module 50 includes a virtual token generation and reverse transmission unit 52. The virtual token generation and reverse transmission unit 52 establishes a data transmission connection with the damping path solving unit 51. The virtual token generation and reverse transmission unit 52 is configured with a standardized data encapsulation protocol to construct a digital token object containing control command information.

[0072] The virtual token generation and reverse transmission unit 52 receives the set of optimal damping control paths output by the damping path solving unit 51, generates a unique virtual control token for each upstream transmission edge in the set of optimal damping control paths, and calculates the reverse potential energy value encapsulated inside the virtual control token. The calculation formula is as follows: ; in: Indicates the upstream operation node number on the optimal damping control path; Indicates the downstream operation node number on the optimal damping control path; Indicates the reverse transmission direction indicator; Indicates downstream operation node The accumulated risk energy value at the current moment; This indicates the preset safe energy threshold for the work node; Indicates upstream operation node Sensitivity coefficient to external resource input; Represented by natural constant Operations on exponential functions with base 0; This represents the preset signal transmission attenuation compensation factor; Indicates upstream operation node With downstream operation nodes The topological path distance between them.

[0073] The virtual token generation and reverse transmission unit 52 writes the calculated reverse potential energy value into the data load area of ​​the virtual control token, uses the virtual control token to perform resource scheduling correction operations on the upstream work node, calculates the amount of physical resource regulation that the upstream work node needs to add, and calculates the reverse potential energy value calculated above by multiplying it by a preset conversion coefficient from potential energy value to physical resource unit, and then performing a floor operation on the product result.

[0074] The virtual token generation and reverse transmission unit 52 generates a resource injection instruction based on the physical resource control amount and sends the resource injection instruction to the atomic job unit construction module 20 to update the resource configuration parameters of the atomic job unit corresponding to the upstream job node, thereby suppressing the forward transmission of risk energy at the source.

[0075] The risk damping and reverse potential energy control module 50 includes a physical access control unit 53. The physical access control unit 53 and the virtual token generation and reverse transmission unit 52 maintain a data synchronization connection. The physical access control unit 53 is equipped with an Internet of Things (IoT) hardware control interface. The IoT hardware control interface is physically connected to the mechanical equipment ignition control circuit controller and the personnel access gate controller of the work area on the construction site through the industrial fieldbus protocol.

[0076] The physical permission lock execution unit 53 reads the virtual control token output by the virtual token generation and reverse transmission unit 52 in real time, parses the reverse potential energy value stored in the virtual control token data load area, compares the reverse potential energy value with the preset forced circuit breaker threshold, and determines that the upstream operation node is in a critical risk state when the reverse potential energy value is greater than the preset forced circuit breaker threshold.

[0077] The physical access control execution unit 53 performs a job freeze duration calculation task for upstream job nodes that are in a critical risk state, calculating the duration of the hardware lock on the upstream job nodes. The calculation formula is as follows: ; in: Indicates the preset reference cooling time unit; This indicates that the virtual token is generated and transmitted by the reverse transmission unit 52 for the upstream working node. The value of the reverse potential energy; This indicates the preset mandatory circuit breaker threshold. This indicates the preset risk response intensity index, which is used to adjust the sensitivity of the lock-in time to the extent of risk energy exceeding the limit; This indicates the minimum fixed time required to complete an on-site safety inspection.

[0078] The physical access control unit 53 generates a physical blocking command that includes the duration of the hardware lock. This command is sent to the ignition control circuit controller of the upstream work node via the IoT hardware control interface. The ignition control circuit controller detects the current operating status of the machinery. If the machinery is in a stationary and unloaded safe state, it performs signal shielding of the ignition circuit to cut off power activation. If the machinery is running or unloaded, it prohibits the input of new action commands until the current action cycle ends and the machinery returns to a safe state before performing the cut-off operation. Simultaneously, the physical access control unit 53 sends the physical blocking command to the personnel access gate controller of the work area associated with the upstream work node. The personnel access gate controller of the work area refuses any access verification request to enter the controlled work area during the duration of the hardware lock.

[0079] The multi-source heterogeneous data sensing and access module 10 monitors the physical environment of the construction site in real time, collects multi-source sensing data from sensor devices installed on construction machinery and workers, cleans and extracts features from the multi-source sensing data, generates physical operation events containing time, space, object and behavioral attributes, and sends the physical operation events to the atomic operation unit construction module 20.

[0080] The atomic operation unit construction module 20 receives physical operation events, calls the spatiotemporal constraint verification unit 21 to verify the legality of physical operation events, calls the dynamic cost aggregation unit 22 to match the resource unit price corresponding to the physical operation event, calculates the instantaneous aggregation cost of the physical operation event, generates standardized atomic operation unit data objects, and transmits the atomic operation unit data objects to the topology network adaptive reconstruction module 30.

[0081] The topology network adaptive reconstruction module 30 receives atomic job unit data objects, uses discrete variance analysis unit 31 to evaluate the internal job heterogeneity of job nodes, and uses node fission and fusion execution unit 32 to perform structural update operation on the directed weighted engineering topology network, and sends the updated directed weighted engineering topology network to cost sensitivity cascade analysis module 40.

[0082] The cost sensitivity cascade analysis module 40 receives the updated directed weighted engineering topology network, calculates the cost sensitivity weights between operation nodes using the elastic weight calculation unit 41, constructs a global risk transmission adjacency matrix using the risk energy transmission unit 42, simulates the cascade diffusion process of risk energy in the network, outputs a cascade risk assessment report, and sends the cascade risk assessment report to the risk damping and reverse potential energy control module 50.

[0083] The risk damping and reverse potential energy control module 50 receives the cascaded risk assessment report, uses the damping path solving unit 51 to determine the optimal damping control path set, uses the virtual token generation and reverse transmission unit 52 to calculate the reverse potential energy value, and then uses the physical permission lock execution unit 53 to generate a physical blocking command or resource injection command, and sends the physical blocking command or resource injection command to the IoT hardware controller at the construction site. After the IoT hardware controller executes the physical blocking command or resource injection command, it changes the physical operation status of the construction site. The multi-source heterogeneous data sensing and access module 10 collects the changed physical operation status again, thereby forming a closed-loop risk control loop.

[0084] During data interaction and collaborative work, in order to ensure the consistency between risk prediction results and the physical site conditions, the system calculates the closed-loop collaborative convergence coefficient. This is achieved by calculating the Euclidean norm of the difference between the cumulative risk energy distribution vector at the current time step and the previous time step, and comparing it with the Euclidean norm of the vector at the previous time step. When the closed-loop collaborative convergence coefficient is greater than the preset system oscillation threshold, the system triggers a full data recalibration process to eliminate the impact of accumulated errors on the accuracy of risk prediction.

Claims

1. A cost-assistance accounting and risk prediction system for construction projects, characterized in that, include: The multi-source heterogeneous data sensing and access module is used to collect multi-source sensor data from the construction site, clean and align it in time and space, and output a standardized time and space physical state data stream. The atomic job unit construction module is used to receive the spatiotemporal physical state data stream, convert the spatiotemporal physical state data stream into atomic job units containing spatiotemporal coordinates and resource consumption data, and calculate the real-time reconstruction cost of each atomic job unit based on the mapping between the measured value of the physical quantity of resource consumption and the real-time library unit price. The topology network adaptive reconstruction module is used to receive the atomic job units and map the atomic job units to job nodes in the directed weighted engineering topology network, dynamically adjust the node structure of the directed weighted engineering topology network according to the key attribute dispersion of the atomic job units belonging to the job nodes, and output the updated directed weighted engineering topology network. The cost sensitivity cascade analysis module is used to calculate the cost sensitivity weights between the work nodes based on the coupling relationship between the schedule and cost, construct the weighted risk transmission network, and simulate the risk cascade diffusion process based on the risk transmission network to predict the cumulative risk energy of the end nodes. The risk damping and reverse potential energy control module is used to generate a virtual resource token based on the accumulated risk energy, transmit the virtual resource token in reverse along the directed weighted engineering topology network to the currently executing job node to adjust the resource configuration, and send an access control instruction to the atomic job unit construction module to perform physical access control when the calculated reverse locking signal strength acting on the currently executing job node exceeds a preset threshold.

2. The cost-aided accounting and risk prediction system for construction projects according to claim 1, characterized in that, The hardware foundation of the multi-source heterogeneous data sensing and access module includes: Crane load sensors are used to collect analog signals of load weight during lifting operations; Personnel identification gates are used to read the radio frequency identification tag data of personnel entering and exiting the work area; Material weighbridges are used to measure the mass data of transport vehicles and loaded materials. Environmental parameter monitoring station, used to collect on-site temperature, humidity and wind speed values; An edge computing gateway is used to receive raw data, perform time alignment processing using an internally integrated high-precision clock source, and output the spatiotemporal physical state data stream.

3. The cost-assistance accounting and risk prediction system for construction projects according to claim 1, characterized in that, When calculating the real-time reconstruction cost, the atomic task unit construction module analyzes the resource types involved in the atomic task unit, obtains the measured physical consumption value and real-time library unit price of each type of resource, combines the timestamp, spatial coordinate vector, and action feature vector of the physical event, and confirms the validity of the physical task event corresponding to the atomic task unit through a spatiotemporal verification logic function and a Kronecker function. Then, it accumulates the product of the consumption of each type of resource and the real-time library unit price to obtain the real-time reconstruction cost. The real-time reconstruction cost is transmitted as a key attribute to the topology network adaptive reconstruction module.

4. The cost-assistance accounting and risk prediction system for construction projects according to claim 1, characterized in that, The adaptive reconstruction module of the topology network calculates the discreteness and adjusts the node structure in the following way: Within the current time window, the key attribute observations of the atomic work units belonging to the work node are statistically analyzed, and the average of the squares of the differences between each key attribute observation and the arithmetic mean of the key attribute observations is calculated to obtain the dispersion. When the dispersion is greater than the preset structural splitting threshold, the node splitting logic is triggered to split the job node into multiple child nodes. Based on the proportion of the total resource consumption of the atomic job units contained in each child node, the inbound and outbound weights of the split job node are allocated to the child nodes.

5. The cost-assistance accounting and risk prediction system for construction projects according to claim 4, characterized in that, The adaptive topology reconstructing module executes node fusion logic in the following way: The adjacent job node pairs with dispersion below the preset aggregation threshold are selected. The cross-correlation coefficient of the job attributes of the adjacent job node pairs within the historical time step is calculated. When the cross-correlation coefficient of the job attributes is greater than the preset correlation judgment threshold, the adjacent job node pairs are merged into a single aggregate node, and the attribute value of the aggregate node is set as the weighted sum of the attribute values ​​of the adjacent job node pairs.

6. The cost-assistance accounting and risk prediction system for construction projects according to claim 1, characterized in that, The cost sensitivity cascade analysis module calculates the cost sensitivity weight in the following way: The upstream and downstream work nodes in the updated directed weighted engineering topology network are identified. The partial derivative of the cost deviation of the downstream work node with respect to the schedule deviation of the upstream work node is obtained by linear regression analysis. The nonlinear coupling factor is calculated based on the ratio of the intersection and union of the resource sets occupied by the upstream and downstream work nodes. The partial derivative is multiplied by the coefficient after correction by the nonlinear coupling factor to obtain the cost sensitivity weight.

7. The cost-assistance accounting and risk prediction system for construction projects according to claim 1, characterized in that, The cost sensitivity cascade analysis module predicts the cumulative risk energy in the following way: The weighted risk transmission network is converted into a global risk transmission adjacency matrix. An initial risk disturbance vector is constructed, consisting of the actual schedule deviation and cost deviation values ​​of each work node. The product of the preset risk transmission attenuation coefficient and the global risk transmission adjacency matrix is ​​subtracted from the identity matrix. The matrix obtained by subtraction is inverted. The inverted matrix is ​​multiplied by the initial risk disturbance vector to obtain the cumulative risk energy distribution vector after the cascade effect stabilizes.

8. The cost-assistance accounting and risk prediction system for construction projects according to claim 1, characterized in that, The risk damping and reverse potential energy control module selects the optimal damping control path for reverse transmission of the virtual resource token in the following way: Based on the reverse connection relationship of the directed weighted engineering topology network, upstream transmission edges pointing to high-risk end nodes are identified. According to the cumulative risk energy of the starting node of the upstream transmission edge, the cost sensitivity weight of the connecting edge, the available risk buffer resources, and the estimated intervention cost, the damping intervention efficiency index for the upstream transmission edge is calculated. Paths with damping intervention efficiency index higher than the preset intervention initiation threshold are selected as the optimal damping control path.

9. A cost-aided accounting and risk prediction system for construction projects according to claim 8, characterized in that, The virtual resource token encapsulates a reverse potential energy value, and the calculation logic for the reverse potential energy value is as follows: Calculate the difference between the cumulative risk energy and the preset safe energy threshold of the upstream and downstream nodes of the optimal damping control path, calculate the ratio of the difference to the resource response sensitivity coefficient of the upstream node, and multiply the ratio by a negative exponential decay factor based on the topological path distance between the upstream and downstream nodes to obtain the reverse potential energy value.

10. A cost-aided accounting and risk prediction system for construction projects according to claim 9, characterized in that, The risk damping and reverse potential energy control module calculates the hardware lock duration of the physical access control based on the reverse potential energy value in the following way: Calculate the ratio of the reverse potential energy value to the preset forced circuit breaker threshold, perform a power operation on the ratio based on the risk response intensity index, subtract one from the power operation result, multiply by the preset benchmark cooling time unit, and add the fixed time cost of on-site safety inspection to obtain the hardware lock duration.

Images