A green wharf construction carbon emission AI monitoring and optimization system

Abstract

The application relates to the technical field of wisdom construction sites, and discloses a green wharf construction carbon emission AI monitoring and optimization system, which comprises an environment and cell modeling module, a carbon potential field dynamic evolution module, a cell adaptive decision module and a visual presentation module. The environment and cell modeling module voxelizes a construction site space, constructs a static carbon potential field, and abstracts a dynamic machine into a work cell. The carbon potential field dynamic evolution module iteratively calculates a total carbon potential field according to instantaneous potential energy and time attenuation effects generated by a real-time state of the work cell, and identifies potential congestion through a field-induced phase change mechanism. The cell adaptive decision module plans an optimal behavior path with the lowest comprehensive cost for the work cell based on the total carbon potential field. The visual presentation module renders the carbon potential field into a three-dimensional cloud chart for intuitive display. The application realizes dynamic quantification and prediction of construction carbon emission risks by constructing a virtual carbon potential field, provides intelligent early warning and decision support for supervisors, effectively guides equipment to avoid high emission risk areas, and significantly improves the management efficiency and emission reduction effect of green construction.

Description

Technical Field

[0001] This invention relates to the field of smart construction site technology, specifically to an AI monitoring and optimization system for carbon emissions from green wharf construction. Background Technology

[0002] Currently, carbon emission management at construction sites typically relies on the later-stage statistical analysis and calculation of data such as fuel consumption and working hours. This approach has significant time lags, making it impossible to intervene in real time when carbon emission events occur, and it lacks the ability to predict and mitigate high emission risks.

[0003] With the application of technologies such as the Internet of Things (IoT) and Building Information Modeling (BIM) in the engineering field, although real-time positioning and preliminary collection of operational status of construction machinery can be achieved, this raw data is often discrete and isolated. Even if managers can see the distribution of equipment on site, it is difficult to gain a macro-level and dynamic understanding of the invisible and potential congestion risks and emission hotspots formed by the interaction of multiple devices across the entire construction site. Existing management systems lack an effective model that can transform dispersed equipment behavior into a quantitative assessment and prediction of the overall site status. Therefore, equipment behavior decisions (such as path planning) often only consider single static factors such as the shortest distance, ignoring the real-time changing environmental costs due to the activities of other dynamic units on site. This leads to equipment frequently entering a state of inefficient operation or even idling, resulting in unnecessary energy waste and carbon emissions. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an AI-based monitoring and optimization system for carbon emissions from green port construction. This system solves the problem that existing construction carbon emission management relies on outdated statistics and lacks effective means to quantify, predict, and proactively avoid dynamic congestion and emission risks caused by the interaction of multiple devices in real time, resulting in suboptimal equipment operation decisions and serious energy waste.

[0005] To achieve the above objectives, the first aspect of this invention provides an AI monitoring and optimization system for carbon emissions from green wharf construction, comprising: The module includes an environment and cell modeling module, a carbon potential field dynamic evolution module, a cell adaptive decision-making module, and a visualization module.

[0006] The environment and cellular modeling module receives and processes 3D geographic information data and construction organization design schemes for the wharf construction site, discretizes the construction site space into a set of voxels, and assigns a static carbon potential value based on the planned operation type to each voxel. This module constructs a static carbon potential field; simultaneously, it abstracts dynamic units such as excavators, cranes, and transport trucks at the construction site into work cells. Each task cell carries a unique identifier, type, and real-time location obtained through an IoT terminal. and real-time working status Attributes such as...

[0007] The carbon potential field dynamic evolution module, connected to the environment and cell modeling module, is used to calculate the total carbon potential field covering the entire construction site. The total carbon potential field consists of the static carbon potential field and a dynamic carbon potential field. Composed of multiple layers: ; in: Representative voxels At any moment The total carbon potential; Representative voxels The static carbon potential; Representative voxels At any moment The dynamic carbon potential value.

[0008] The dynamic carbon potential field is calculated using a discrete-time iterative method, which in The value of time ,Depend on Dynamic carbon potential value at time t After decay over time and accumulation All moments The instantaneous potential energy generated by each working cell The iterative calculation follows the formula: ; in: Representative voxels In the previous moment The dynamic carbon potential value; Represents the discrete time step; This represents the preset time decay coefficient; The total number of cells representing the job; Representing the Each task cell at time voxels The instantaneous potential energy generated.

[0009] Each job cell exist For any voxel at any time The instantaneous potential energy generated , by an emission intensity function It consists of a spatial attenuation term: ; in: It depends on the job cell type and real-time working status The determined value of the basic potential energy intensity; Represents the preset spatial attenuation coefficient; Representative voxel v and working cell At any moment Location The spatial distance between them.

[0010] Furthermore, the carbon potential field dynamic evolution module is also used to monitor the total carbon potential value of any region, and when it is determined that the total carbon potential value of a certain region exceeds a preset critical duration... The internal potential energy remains consistently higher than the preset potential threshold. When this occurs, a field-induced phase transition is triggered in the region, leading to a viscous blockage state.

[0011] The cellular adaptive decision-making module, connected to the carbon potential field dynamic evolution module, is used to plan behavioral paths for the task-receiving cells. This cellular adaptive decision-making module solves the comprehensive cost function. The minimum value is used to determine the optimal behavior path. The comprehensive cost function Defined as: ; in: Represents a spatial path; Represents along the path Distance variable; Represents the distance from the starting point on the path. The point at the location; The representative cell moves to the path point. The predicted time; Represents the preset distance cost weight; Represents the preset carbon potential cost weight; Representing the job cell at the path point and corresponding time Total carbon potential at that time; Representing path points At the corresponding time The passage cost multiplier when the point is in the viscous congestion state. It is set to a default value greater than 1, otherwise it is 1.

[0012] The visualization module is connected to the carbon potential field dynamic evolution module. It is used to acquire the total carbon potential field data and perform three-dimensional visualization rendering in the form of a carbon potential cloud map, while highlighting the areas that have entered the viscous blockage state.

[0013] A second aspect of this invention provides an AI-based method for monitoring and optimizing carbon emissions from green wharf construction, the method comprising the following steps: Step 1: Construct the static carbon potential field of the wharf construction site, and abstract the dynamic units of the construction site into operational cells that carry real-time location and real-time working status.

[0014] Step 2: Based on the real-time position and real-time working status of each of the work cells, the total carbon potential field covering the entire construction site is obtained by iteratively calculating the total carbon potential field covering the entire construction site by performing time decay processing on the total carbon potential field of the previous moment and superimposing the instantaneous potential energy generated by all work cells at the current moment.

[0015] Step 3: Based on the total carbon potential field, the behavior path is planned for the job cell by solving for the minimum value of the comprehensive cost function.

[0016] Step 4: Dynamically visualize the total carbon potential field in the form of a carbon potential cloud map.

[0017] This invention provides an AI-based monitoring and optimization system for carbon emissions from green wharf construction. It offers the following advantages: 1. This invention achieves real-time and global monitoring of carbon emission risks across the entire construction site by constructing a dynamic carbon potential field driven by the real-time state of all working cells and possessing time decay characteristics. This mechanism not only reflects the current working conditions but also provides continuous risk warnings for areas where high emission events have just occurred by briefly retaining historical potential fields, thus overcoming the inherent data lag and low accuracy problems of traditional static accounting methods based on emission factors.

[0018] 2. By introducing a field-induced phase transition mechanism, this invention can identify and quantify nonlinear effects such as increased congestion caused by the aggregation of multiple devices. When a region enters a viscous congestion state, the system will drastically increase the travel cost of that region through a travel cost multiplier. This nonlinear penalty mechanism enables the working cells to proactively and in advance avoid the congestion hotspots that are about to form when making decisions, thus achieving effective prevention of major and continuous high emission events.

[0019] 3. This invention utilizes a comprehensive cost function based on the total carbon potential field to provide a global perspective for the decision-making of each job cell. Since the total carbon potential field is the superposition of the influence of all job cells in the field, when any cell plans its own optimal path, it indirectly considers the state and behavior of all other cells. This transforms isolated, local individual decisions into interconnected, globally collaborative optimizations, effectively reducing unnecessary idling and energy consumption caused by scheduling conflicts and disordered operations. Attached Figure Description

[0020] Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a functional diagram of the environment and cell modeling module of the present invention; Figure 4 This is a schematic diagram illustrating the path planning principle of the cellular adaptive decision-making module of the present invention; Figure 5 This is a functional diagram of the visualization module of the present invention.

[0021] Among them, 100 is the Environment and Cell Modeling Module; 200 is the Carbon Potential Field Dynamic Evolution Module; 300 is the Cell Adaptive Decision Module; and 400 is the Visualization Module. Detailed Implementation

[0022] The technical solutions of 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] Example: Please see the appendix Figure 1 - Appendix Figure 5 This invention provides an AI-based monitoring and optimization system for carbon emissions from green wharf construction, comprising: The Environment and Cellular Modeling Module 100 serves as the data foundation of the system. Its function is to statically digitize the physical construction environment and to perform data abstraction and real-time status acquisition of dynamic work units during the construction process.

[0024] Please see the appendix Figure 3 The specific implementation of the environment and cell modeling module 100 includes two parallel sub-processes: static modeling of the construction environment and dynamic modeling of the work cells.

[0025] In the static modeling section of the construction environment, the environment and cell modeling module 100 receives and processes three-dimensional geographic information data of the wharf construction site, such as building information model data or computer-aided design drawings. Based on the concept of finite element analysis, the processor inside the environment and cell modeling module 100 executes a spatial discretization algorithm, dividing the continuous three-dimensional space into a set of voxels of predetermined size (e.g., 1 meter × 1 meter × 1 meter). Each voxel... It has a unique three-dimensional coordinate system to enable the discretization of continuous fields.

[0026] Based on this, the environment and cellular modeling module 100 receives and parses the construction organization design plan file, which contains functional plans for different areas of the site, such as heavy machinery intensive operation areas, main material transportation routes, temporary storage areas, and ecological protection areas where entry is prohibited. According to this planning information, the module uses preset mapping rules or lookup tables to map each voxel... Assign a definite, time-invariant static carbon potential value. As a preferred setting, voxels located in areas with intensive heavy machinery operations are assigned higher base potential energy values, while voxels located in ecological protection areas are assigned extremely high values ​​that approach infinity, in order to form an insurmountable potential barrier in subsequent calculations. Once this process is completed, the static carbon potential field covering the entire construction site is constructed and stored.

[0027] In the dynamic modeling part of the work cell, this module abstracts the core dynamic units of the construction site, including each excavator, crane, transport truck, and construction team, into a data structure called a work cell. Each job cell Described by a set of attributes, which includes at least: a unique identifier Types used to distinguish different types of equipment This represents its current real-time position in three-dimensional coordinates. And its real-time operating status, indicating its current operating condition. .

[0028] To obtain real-time location and real-time working status Each key operational unit (mainly referring to high-energy-consuming construction equipment) is equipped with an IoT terminal. This IoT terminal integrates a Global Navigation Satellite System receiver to acquire high-precision latitude, longitude, and elevation data, and converts it into unified three-dimensional coordinates of the construction site as its real-time location. The output.

[0029] Meanwhile, the IoT terminal connects to the device's controller LAN via a data cable, directly reading parameters such as engine speed, torque, and throttle opening. The processor within the terminal uses these parameters, based on preset logical thresholds, to determine the device's current operating condition and maps it to one of predefined discrete states, such as {off, idling, moving, full load operation}, as its real-time operating status. For devices without a local area network connection, the output can be analyzed using external sensors, such as vibration sensors or motor current sensors, to determine the operating status. It's worth noting that, considering the potential discrepancies in sampling frequencies between GNSS and CAN bus data, this module internally maintains a time alignment buffer. A linear interpolation algorithm is used to unify the two data streams to the same timestamp 't', ensuring accurate location data. Operating conditions A strict correspondence in physical spacetime.

[0030] The environment and cell modeling module 100 transmits the completed static carbon potential field data, as well as the attribute data of all working cells collected and updated in real time, to the carbon potential field dynamic evolution module 200.

[0031] The carbon potential field dynamic evolution module 200 is connected to the environment and cell modeling module 100. It is used to receive static carbon potential field data and real-time attribute data of all working cells. Its function is to calculate and update the total carbon potential field that represents the spatiotemporal distribution of carbon emission risk of the entire construction site in real time.

[0032] The carbon potential field dynamic evolution module 200 is based on the principle of field superposition, and it receives the static carbon potential field. The dynamic carbon potential field calculated by the dynamic evolution module 200 of this carbon potential field Perform linear superposition to obtain any voxel At any moment Total carbon potential .

[0033] ; in: Representative voxels At any moment The total carbon potential value, whose physical meaning is the current comprehensive congestion risk and emission pressure at that spatial point; Representative voxels The static carbon potential; Representative voxels At any moment The dynamic carbon potential value.

[0034] Dynamic carbon potential field The calculation at discrete time steps The process proceeds iteratively. At each time step, the module first performs the calculation of instantaneous potential energy generation for the entire field. Each of the job cells The module calculates its time at time For any voxel The instantaneous potential energy generated The calculation follows the formula: ; in: It is an emission intensity function, implemented as a predefined two-dimensional lookup table, where the two dimensions of the lookup table are the types of the job cells. and real-time working status The values ​​stored in the table are the basic potential energy intensity values ​​of the corresponding types of equipment under the corresponding operating conditions. These values ​​are calibrated based on equipment emission standards or measured data, reflecting the instantaneous carbon emission capacity of different equipment under different operating conditions.

[0035] It is a preset spatial attenuation coefficient, a normal value, used to control the rate attenuation of instantaneous potential energy with spatial distance. Its value is set according to the openness of the construction site. The more open the site, the stronger the diffusion effect, and the smaller this value.

[0036] It is a voxel Center point coordinates and operation cell At any moment Real-time location The Euclidean distance between them, to ensure the stability of the calculation, when When it approaches 0, The function value is 1, which avoids the singularity problem in the traditional inverse proportional potential field function.

[0037] After calculating the instantaneous potential energy generated by all working cells for all voxels, the module performs a time evolution calculation of the dynamic carbon potential field, which takes the previous time step as an example. Dynamic carbon potential field Multiply by the time decay factor, and then superimpose it with the sum of the instantaneous potential energy generated by all working cells at the current moment to obtain the dynamic carbon potential field at the current moment. The calculation follows the formula: ; in: Representative voxels In the previous moment The dynamic carbon potential value; It is a preset time decay coefficient, which is a normal number, used to control the dissipation rate of historical potential energy, simulating the natural mitigation process of exhaust gas diffusion and congestion effects over time.

[0038] The total number of cells representing the job.

[0039] Furthermore, the carbon potential field dynamic evolution module 200 also includes the implementation of the field-induced phase transition mechanism, and the carbon potential field dynamic evolution module 200 provides each voxel with... The module maintains a time-series cache storing the total carbon potential over a past period. At each time step, the module calculates the time-series cache for a preset critical duration. The arithmetic mean within the range; if this arithmetic mean continuously exceeds a preset potential energy threshold... The carbon potential field dynamic evolution module 200 will then apply this to the voxel. The internal status flag is set to viscous blockage state.

[0040] Once voxels The status flag is set to a viscous blockage state. When the carbon potential field dynamic evolution module 200 outputs data, it will include the passage cost multiplier associated with that voxel. The value is updated from the default 1 to a preset value greater than 1. (For example, a value between 10 and 100) When the arithmetic mean of the total carbon potential in the time-series buffer of this voxel falls back to the potential threshold. After the preset release time, its status flag is reset, and the passage cost multiplier is adjusted. It also returns to 1.

[0041] Ultimately, the carbon potential field dynamic evolution module 200 will contain all voxels at the current moment. Total carbon potential and toll cost multiplier The data structure is transmitted to the cellular adaptive decision-making module 300 and the visualization module 400.

[0042] Cellular adaptive decision-making module 300 is connected to carbon potential field dynamic evolution module 200 and is used to receive real-time updated total carbon potential field data. When the system receives a scheduling task for a certain job cell, such as moving from starting point A to ending point B, the function of this module is to plan a behavior path with the best overall cost for the job cell.

[0043] Please see the appendix Figure 4 The core of this cellular adaptive decision-making module 300 is based on the comprehensive cost function. This function performs path optimization and is used to quantify any feasible path. The total cost is defined as follows: ; in: Represents a spatial path; Represents along the path Distance variable; Represents the distance from the starting point on the path. The point at the location; The representative cell moves to the path point under the assumption of uniform speed. The predicted time; This represents the preset distance cost weight, which physically represents the fixed cost per unit distance and is related to factors such as energy consumption and time. This represents a preset carbon potential cost weight, used to adjust the proportion of carbon emission risk in the total cost; Representing the job cell at the path point and corresponding prediction time The total carbon potential value at time is obtained by querying the carbon potential field dynamic evolution module 200 for the predicted potential energy at a specified spatiotemporal coordinate; Representing path points At the corresponding predicted time The passage cost multiplier, also provided by the carbon potential field dynamic evolution module 200, has a value greater than 1 when the voxel where the path point is located is in a viscous blockage state, and is 1 otherwise. This is achieved by introducing... This initiative implements soft-constraint penalties for highly congested and high-emission areas, guiding routes to proactively avoid them.

[0044] To solve for the comprehensive cost function Minimize the optimal path This embodiment employs a spatiotemporal A* search algorithm, which expands the search space from three-dimensional physical space to four-dimensional spatiotemporal space. Each node in the search graph contains not only three-dimensional coordinates but also a time dimension.

[0045] When the algorithm executes, it starts from an initial node containing the starting position and starting time. In each iteration, the algorithm selects the total cost prediction value from the set of nodes to be visited. smallest node Expanding the total cost forecast It consists of two parts: .

[0046] in, From the initial node to the current node The actual cumulative cost, for the parent node To the current node The cost increment generated by each step of movement is the discretized summation of the comprehensive cost function on the corresponding path segment.

[0047] From the current node The estimated cost to the target node, i.e., the heuristic function, in this embodiment, It is set as the Euclidean distance from the current node position to the target point position multiplied by the distance cost weight. ,because It is the minimum unit cost. This heuristic function satisfies the conditions of admissibility and consistency. It is an admissible heuristic function, which guarantees the optimality of the algorithm search.

[0048] In the extended node At that time, the algorithm generates its spatiotemporally adjacent successor nodes. For each successor node, the algorithm queries the carbon potential field dynamic evolution module 200 for the predicted total carbon potential value and passage cost multiplier of its corresponding spatiotemporal coordinates, and calculates its... The value is determined, and the process continues until the target node is accessed.

[0049] After the algorithm completes execution, it outputs an optimal behavioral path consisting of a series of spatiotemporal waypoints. The path data is then sent to the airborne control system or dispatch terminal of the corresponding job cell to guide its actual operation.

[0050] The visualization module 400, connected to the carbon potential field dynamic evolution module 200, is used to receive data containing all voxels at the current moment. Total carbon potential and toll cost multiplier The data structure. The function of this module is to convert these numerical data into a three-dimensional graphical representation for managers to observe and analyze.

[0051] Please see the appendix Figure 5 The core function of this visualization module 400 is the generation and display of carbon potential cloud maps. The module uses volume rendering technology and transmits the input scalar total carbon potential value through a preset transfer function. The mapping is applied to two visual attributes: opacity and color. For example, the transfer function maps potential values ​​below a first threshold to a low-opacity blue, potential values ​​between the first and second thresholds to a smooth transition color from green to yellow, and potential values ​​above the second threshold to a high-opacity red. By performing this rendering operation on the entire set of voxels, a three-dimensional cloud map that can intuitively reflect the spatial distribution and intensity gradient of the carbon potential field is finally generated.

[0052] When the data received by the visualization module 400 contains a toll cost multiplier... When the number of voxels is greater than 1, the module identifies the set of voxels in that state and overlays a preset, visually distinctive graphic marker on its spatial location. This marker can be a flashing wireframe of a specific color or a semi-transparent grid cube, used to clearly indicate the area that has entered the viscous blockage state.

[0053] The visualization module 400 also provides a user interface that allows users to pan, zoom, and rotate the rendered 3D scene using an input device to observe the carbon potential field from different perspectives. The interface also includes a timeline control bar, which users can drag to view a snapshot of the carbon potential field at any historical moment or to continuously play the evolution of the carbon potential field at a predetermined rate.

[0054] In addition, the visualization module 400 supports a data probe function. When the user specifies a spatial point in the rendering screen, the system will query and display the real-time total carbon potential value, status flag, and other numerical information of the voxel corresponding to the coordinates of that point. Finally, the graphic data stream after rendering and user interaction is sent to the display device for presentation.

[0055] See attached document Figure 2 , Figure 2 This invention relates to another embodiment of a green wharf construction carbon emission AI monitoring and optimization method, which includes the following steps: S601, Execution environment and cell modeling operations.

[0056] This step first receives and processes the three-dimensional geographic information data of the wharf construction site and the construction organization design plan, discretizes the construction site space into a set of voxels, and assigns each voxel a design based on the design plan. Assign a specific static carbon potential value To construct a static carbon potential field, and at the same time, to abstract dynamic units such as excavators, cranes, and transport trucks at the construction site into units carrying unique identifiers, types, and real-time locations. and real-time working status Job cells with equal attributes Real-time data acquisition is achieved through an IoT terminal configured for the job cell, which integrates a GNSS receiver and a status sensor and performs time synchronization operations.

[0057] S602, Perform dynamic evolution calculation of carbon potential field.

[0058] This step is performed at a preset discrete time step. The iteration continues. At each time step... Based on the real-time positions and working states of all job cells, by checking the previous time step... The total carbon potential field is subjected to time decay processing, and the instantaneous potential energy generated by all working cells at the current moment is superimposed to calculate the total carbon potential field covering the entire construction site at the current moment. .

[0059] In step S602, the total carbon potential field The calculation is based on the static carbon potential field. and dynamic carbon potential field Superimposed, dynamic carbon potential field The iterative calculation follows the formula: ; in: Representative voxels In the previous moment The dynamic carbon potential value; This represents the preset time decay coefficient; Represents the discrete time step; The total number of cells representing the job; Representing the Each task cell at time voxels The instantaneous potential energy generated.

[0060] Instantaneous potential energy The calculation follows the formula: ; in: It depends on the job cell type and real-time working status The basic potential energy intensity value is determined by querying a preset two-dimensional lookup table; Represents the preset spatial attenuation coefficient; Representative voxels With job cells At time t The spatial distance between them.

[0061] Furthermore, step S602 also includes performing a field-induced phase transition determination.

[0062] This judgment is made by continuously monitoring the total carbon potential value of any region, and determining when it exceeds a preset critical duration. The internal potential energy remains consistently higher than the preset potential threshold. If the state is not specified, the region's state is updated to a viscous congestion state, and its passage cost multiplier is adjusted. Set to a preset value greater than 1 to reflect the decrease in traffic efficiency in this area due to carbon potential overload.

[0063] S603, Perform cellular adaptive decision-making operation.

[0064] When a job cell receives a new job task, this step, based on the total carbon potential field output from step S602, solves the comprehensive cost function. The minimum value is used to plan an optimal behavioral path from the start point to the end point for the task cell. .

[0065] In step S603, the comprehensive cost function Defined as: ; in: Represents a spatial path; Represents along the path Distance variable; Represents the distance from the starting point on the path. The point at the location; The representative cell moves to the path point. The predicted time; Represents the preset distance cost weight; Represents the preset carbon potential cost weight; Representing path points In the predicted time Total carbon potential at that time; Representing path points In the predicted time The toll cost multiplier at that time.

[0066] S604, Perform visualization operation.

[0067] This step converts the total carbon potential field data calculated in step S602 into a three-dimensional carbon potential cloud map using rendering technology, highlights the regions in specific states, and finally outputs the rendering results to the display device.

Claims

1. An AI monitoring and optimization system for carbon emissions from green wharf construction, characterized in that, include: The Environment and Cell Modeling Module is used to receive and process the three-dimensional geographic information data and construction organization design scheme of the dock construction site. It discretizes the construction site space into a set of voxels and assigns a static carbon potential value based on the planned operation type to each voxel in order to construct a static carbon potential field. It also abstracts the dynamic units of the construction site into operation cells carrying real-time status. The carbon potential field dynamic evolution module is connected to the environment and cell modeling module. It is used to iteratively calculate the dynamic carbon potential field at discrete time steps based on the real-time state of each operation cell, and linearly superimpose the dynamic carbon potential field with the static carbon potential field to obtain the total carbon potential field covering the entire construction site. The cellular adaptive decision-making module, connected to the carbon potential field dynamic evolution module, is used to plan behavioral paths for the working cells based on the total carbon potential field using a comprehensive cost function. The comprehensive cost function quantifies the total cost of the behavioral path, and its calculation formula is as follows: ； in: Represents a spatial path; Represents along the path Distance variable; Represents the distance from the starting point on the path. The point at the location; The representative cell moves to the path point under the assumption of uniform speed. The predicted time; This represents the preset distance cost weight, which physically represents the fixed cost per unit distance and is related to energy consumption and time factors. Representing the job cell at the path point and corresponding prediction time The total carbon potential value at time is obtained by querying the predicted potential energy at a specified spatiotemporal coordinate from the carbon potential field dynamic evolution module; Representing path points At the corresponding predicted time The toll cost multiplier; This represents a preset carbon potential cost weight, used to adjust the proportion of carbon emission risk in the total cost; The visualization module is connected to the carbon potential field dynamic evolution module and is used to dynamically visualize the total carbon potential field in the form of a carbon potential cloud map. The carbon potential field dynamic evolution module is specifically used for: The dynamic carbon potential field of the previous moment is subjected to time decay processing to obtain the historical carbon potential field. Instantaneous potential energy is generated based on the real-time state of each of the said working cells; The historical carbon potential field is superimposed with the instantaneous potential energy generated by all the working cells to update the dynamic carbon potential field at the current moment. The instantaneous potential energy is the basic potential energy intensity value generated by the working cell according to its type and working state, and the basic potential energy intensity value is formed by exponentially decaying with spatial distance. Based on the superposition principle of fields, the received static carbon potential field is linearly superimposed with the dynamic carbon potential field calculated by the carbon potential field dynamic evolution module to obtain the total carbon potential value of any voxel at the current moment. The comprehensive cost function includes a weighted sum of the distance cost and carbon potential cost of the behavioral path; The carbon potential cost is the spatiotemporal integral of the total carbon potential field experienced by the working cell as it travels along the behavioral path.

2. The AI ​​monitoring and optimization system for carbon emissions from green wharf construction according to claim 1, characterized in that, The carbon potential field dynamic evolution module is also used for: Monitor the potential energy value of any region in the total carbon potential field; when it is determined that the potential energy value of a certain region is continuously higher than the potential energy threshold for a period of time exceeding the critical duration, trigger a field-induced phase transition in that region and enter a viscous blockage state.

3. The AI ​​monitoring and optimization system for carbon emissions from green wharf construction according to claim 2, characterized in that, The cell adaptive decision-making module is specifically used for: A preset passage cost multiplier is applied to areas entering the viscous congestion state; When planning the behavior path for the job cell, the travel cost multiplier is incorporated into the calculation of the comprehensive cost function.

4. The AI ​​monitoring and optimization system for carbon emissions from green wharf construction according to claim 1, characterized in that, The cellular adaptive decision-making module is configured to determine the behavioral path by solving for the minimum value of the comprehensive cost function.

5. The AI ​​monitoring and optimization system for carbon emissions from green wharf construction according to claim 3, characterized in that, When the behavioral path passes through the region entering the viscous blockage state, the calculation of the carbon potential cost also includes multiplying by a preset passage cost multiplier.

6. The AI ​​monitoring and optimization system for carbon emissions from green wharf construction according to claim 2, characterized in that, The visualization module is also used to highlight and mark areas that have entered the viscous blockage state on the carbon potential cloud map.

7. A method for monitoring and optimizing carbon emissions from green wharf construction using AI, and a system for monitoring and optimizing carbon emissions from green wharf construction using AI according to any one of claims 1-6, characterized in that, Includes the following steps: Construct a static carbon potential field for the wharf construction site and abstract the dynamic units of the construction site into operational cells carrying real-time status. Based on the real-time state of each of the work cells, the total carbon potential field covering the entire construction site is calculated iteratively. Based on the total carbon potential field, a behavior path is planned for the operation cell through a comprehensive cost function; The total carbon potential field is dynamically visualized in the form of a carbon potential cloud map.

Images