Simulation analysis method and system for electric power engineering cost based on digital twinning

Abstract

The invention discloses an electric power engineering cost simulation analysis method and system based on digital twinning, and relates to the technical field of simulation analysis, and the method comprises the steps: obtaining multiple engineering task multi-modal data and task dependency relationships for electric power engineering; processing the multi-modal data of the plurality of engineering tasks and the task dependency relationship based on a digital twinning technology so as to construct an electric power engineering virtual model; inputting the electric power engineering virtual model into a graph neural network so as to extract task node features corresponding to each engineering task virtual module through a message passing mechanism; and defining an input state of a reinforcement learning model according to each extracted task node feature so as to generate a task resource allocation action for each engineering task through reinforcement learning. Therefore, the digital twinborn technology, the graph theory feature extraction technology and the reinforcement learning technology are comprehensively applied, real-time monitoring and simulation analysis are carried out on cost and resource consumption of different links in the electric power engineering, and resource allocation is intelligently optimized.

Description

Technical Field

[0001] This application relates to the field of simulation analysis technology, and in particular, to a simulation analysis method and system for the cost of power engineering based on digital twin. Background Art

[0002] Due to the technical complexity and long construction period involved in power engineering projects, cost control has always been a key issue in project management. Traditional power engineering cost management methods rely on means such as static budget preparation, manual estimation, and historical data review. They usually make simple cost predictions and controls based on the initial setting of the project plan. Although these methods can provide certain references for cost prediction, they rely on a large number of assumptions and simplifications, often ignoring the complex interactions between various links or engineering tasks in power engineering.

[0003] In addition, the data of each link in power engineering is usually managed separately by different departments, lacking an effective data integration and sharing mechanism, resulting in the problem of "data islands". It is difficult to communicate information between different links, and even within the same link, it is difficult to effectively integrate and analyze data. For example, the complex relationships between engineering construction and equipment procurement, and between civil engineering tasks affecting mechanical and electrical installation tasks, etc., are often not fully reflected in the model. Therefore, in the face of a complex and dynamic engineering environment, the accuracy and efficiency of traditional cost optimization control strategies are greatly reduced.

[0004] In response to the above problems, the industry has not yet proposed a better technical solution. Summary of the Invention

[0005] This application provides a simulation analysis method, system, storage medium, computer program product, and electronic device for the cost of power engineering based on digital twin, so as to at least solve the problems such as relying on static assumptions and data islands in traditional power engineering project management.

[0006] In a first aspect, an embodiment of the present application provides a simulation analysis method for the cost of a power project based on digital twins, including: obtaining multi-modal data and task dependencies of multiple engineering tasks for the power project; the multi-modal data of the engineering tasks includes task progress data, resource consumption data, and engineering area meteorological data of the corresponding engineering tasks; the resource consumption data includes building material cost consumption information, engineering equipment cost usage information, and engineering labor cost usage information of the corresponding engineering tasks; processing the multi-modal data and task dependencies of the multiple engineering tasks based on digital twin technology to construct a virtual model of the power project; the virtual model of the power project includes multiple virtual modules of engineering tasks, each of the virtual modules of engineering tasks is modeled by the corresponding multi-modal data of the engineering tasks, and there are edge connections between the associated virtual modules of engineering tasks indicated by the task dependencies; inputting the virtual model of the power project into a graph neural network to extract the task node features corresponding to each of the virtual modules of engineering tasks through a message passing mechanism; defining the input state of a reinforcement learning model according to the extracted task node features to generate task resource allocation actions for each of the engineering tasks through reinforcement learning, the task resource allocation actions are used to indicate the resource allocation scheme for the corresponding engineering tasks, and the resource allocation scheme includes a building material allocation strategy, an engineering equipment allocation strategy, and a labor force allocation strategy; the reward function of the reinforcement learning model is defined according to the total project completion cost, the total project completion cycle, and the average resource usage efficiency of the engineering tasks of the power project.

[0007] Second aspect, an embodiment of the present application provides a simulation analysis system for the cost of a power project based on digital twins, including: a data acquisition unit, configured to acquire multi-modal data and task dependencies of multiple engineering tasks for the power project; the multi-modal data of the engineering tasks includes task progress data, resource consumption data, and engineering area meteorological data of the corresponding engineering tasks; the resource consumption data includes building material cost consumption information, engineering equipment cost usage information, and engineering labor cost usage information of the corresponding engineering tasks; a twin modeling unit, configured to process the multi-modal data and task dependencies of the multiple engineering tasks based on digital twin technology to construct a virtual model of the power project; the virtual model of the power project includes multiple virtual modules of engineering tasks, each of the virtual modules of the engineering tasks is modeled by the corresponding multi-modal data of the engineering tasks, and there are edge connections between the associated virtual modules of the engineering tasks indicated by the task dependencies; a feature extraction unit, configured to input the virtual model of the power project into a graph neural network to extract task node features corresponding to each of the virtual modules of the engineering tasks through a message passing mechanism; a reinforcement learning analysis unit, configured to define the input state of a reinforcement learning model according to the extracted task node features, so as to generate task resource allocation actions for each of the engineering tasks through reinforcement learning, the task resource allocation actions are used to indicate the resource allocation scheme for the corresponding engineering tasks, and the resource allocation scheme includes a building material allocation strategy, an engineering equipment allocation strategy, and a labor force allocation strategy; the reward function of the reinforcement learning model is defined according to the total cost of project completion, the total project completion period, and the average usage efficiency of engineering task resources of the power project.

[0008] Third aspect, there is provided an electronic device, including: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and when the instructions are executed by the at least one processor, the at least one processor is enabled to execute the steps of the simulation analysis method for the cost of a power project based on digital twins according to any embodiment of the present application.

[0009] Fourth aspect, an embodiment of the present application provides a storage medium, on which a computer program is stored, characterized in that when the program is executed by a processor, the steps of the simulation analysis method for the cost of a power project based on digital twins according to any embodiment of the present application are implemented.

[0010] Fifth aspect, an embodiment of the present application provides a computer program product, including computer programs / instructions, and when the computer programs / instructions are executed by a processor, the steps of the simulation analysis method for the cost of a power project based on digital twins according to any embodiment of the present application are implemented.

[0011] Through a simulation analysis method and system for the cost of power engineering based on digital twin provided by this application, the following technical effects can be achieved at least:

[0012] (1) By integrating multi-modal data of multiple engineering tasks (such as task progress, resource consumption, meteorological data, etc.) and establishing a virtual model of power engineering, the interaction relationship between each task can be reflected more accurately. The static budget and manual estimation of traditional methods are difficult to cope with the dynamically changing engineering environment, while the virtual model established through digital twin technology can simulate and track the project progress in real time, implement dynamic cost monitoring and resource scheduling, adjust the resource allocation plan according to the actual project progress, and effectively avoid the errors caused by assumption deviation and data simplification in traditional methods.

[0013] (2) Through digital twin technology, the data of each link are effectively integrated and aggregated in a unified virtual model for correlation modeling, which can share information in real time during the project management process, ensure the coordination and optimization of each link, improve the coordination efficiency between each link, and help to identify potential risks in a timely manner and make targeted adjustments.

[0014] (3) Using a graph neural network (Graph Neural Network, GNN) to process and capture the complex dependence relationships between virtual modules (i.e., nodes) of different engineering tasks in the virtual model of power engineering (i.e., graph structure), dynamically transmit information and optimize the resource allocation of each task virtual module, so as to improve the overall execution efficiency of the project and avoid the situation of resource waste and scheduling chaos caused by fuzzy or simplistic relationships between tasks in traditional methods.

[0015] (4) Through the reinforcement learning algorithm, integrating the real-time update of the task status of digital twin, it can self-learn and adjust between multiple engineering tasks, dynamically adjust the optimal resource allocation strategy, and support the implementation of a dynamically optimized decision support system. In addition, the reward function of reinforcement learning is based on the actual completion cost, total cycle and resource utilization efficiency of power engineering, which can adaptively adjust the resource allocation strategy, so that the power engineering project can achieve optimal execution within the budget. Thus, through intelligent optimization, the resource allocation strategy can be adjusted in a timely manner in the face of changing engineering environments, minimizing resource waste and improving the economy and sustainability of the project.

[0016] Through this technical solution, by comprehensively applying digital twin technology, graph theory feature extraction technology and reinforcement learning technology, the cost and resource consumption of different links in power engineering are monitored and simulated in real time, the resource allocation is intelligently optimized, the coordination efficiency between departments is enhanced and the unified management of engineering resources is realized, the resource utilization efficiency is improved, and the overall investment cost of power engineering can be effectively reduced. Brief Description of the Drawings

[0017] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0018] Figure 1 Shows a flowchart of an example of a simulation analysis method for the cost of a power project based on digital twins according to an embodiment of the present application;

[0019] Figure 2 Shows a schematic diagram of the principle of an example of the state transition action in the reinforcement learning model;

[0020] Figure 3 Shows according to Figure 1 An example of the operation flowchart of step S110 in

[0021] Figure 4 Shows an example of the operation flowchart of constructing a virtual model of a power project based on digital twin technology according to an embodiment of the present application;

[0022] Figure 5 Shows a schematic diagram of the structural connection of an example of the reinforcement learning model according to an embodiment of the present application;

[0023] Figure 6 Shows a structural block diagram of an example of a simulation analysis system for the cost of a power project based on digital twins according to an embodiment of the present application;

[0024] Figure 7 Is a schematic diagram of the structure of an embodiment of the electronic device of the present application. Specific embodiments

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0026] In the technical solutions of the present application, for the processing of the collection, storage, use, processing, transmission, provision, and disclosure of user personal information, etc., they all comply with the provisions of relevant laws and regulations and do not violate public order and good customs.

[0027] It should be noted that digital twin is a method of synchronizing the physical world and the virtual world through real-time data, and it can reflect the state and behavior of physical entities in real time through a highly accurate virtual model.

[0028] Figure 1 The flowchart of an example of the simulation analysis method for the cost of power engineering based on digital twin according to an embodiment of the present application is shown.

[0029] Regarding the execution subject of the method in the embodiment of the present application, it can be any controller or processor with computing or processing capabilities. Specifically, it can be implemented by the cost management cloud platform for power engineering. By integrating data from different departments and links, a unified digital twin model is constructed, breaking the information barrier between engineering tasks. The data between each link can be shared and interconnected, eliminating the "data island" problem, improving the efficiency of cooperation among departments, and ensuring the timely update and feedback of information. Furthermore, through simulation analysis and multi-dimensional data fusion, a comprehensive and accurate decision support tool is provided for engineering managers, helping them better evaluate cost risks, adjust resource allocation, and make reasonable cost optimization decisions during the engineering implementation process, thus significantly improving the intelligent management level of engineering projects.

[0030] In some examples, it can be integrated and configured in an electronic device or terminal in a software, hardware, or software-hardware combination manner, and the types of terminals or electronic devices can be diverse, such as mobile phones, tablets, or desktop computers, etc.

[0031] As Figure 1 shown, in step S110, a plurality of multi-modal data and task dependencies for power engineering tasks are obtained.

[0032] It should be understood that the acquisition methods of engineering task multi-modal data and task dependencies can be diverse, such as through channels such as sensors, intelligent devices, monitoring systems, and management terminal interactions, which are not limited here for the time being. In addition, the types of engineering tasks under power engineering can also be diverse and can be set or adjusted according to business requirements, such as civil engineering tasks, substation equipment installation tasks, communication system construction tasks, intelligent control system installation tasks, power equipment maintenance tasks, and power grid optimization tasks, etc. In some cases, the engineering tasks can also be designed in a multi-layer task structure, that is, the engineering tasks can also include multiple levels of subtasks. For example, the civil engineering task includes subtasks such as excavation of foundation pits, foundation construction, and reinforced concrete pouring.

[0033] It should be noted that there may be some potential correlation relationships between different tasks. As an example, task dependencies can be determined through project plans (such as Gantt charts, network diagrams) and construction process flows. In some embodiments, the system can extract the sequential relationships, parallel or serial execution constraint conditions between tasks based on the inputs in the project design phase, so as to clarify the dependency order and resource conflicts between tasks. For example, the substation equipment installation task can only be started after the completion of the civil engineering task, and the intelligent control system installation task can only be started after the completion of the communication system construction task, and so on.

[0034] Here, the engineering task multimodal data includes the task progress data, resource consumption data, and engineering area meteorological data of the corresponding engineering task. In some embodiments, the task progress data can include the current status of the engineering task (such as to be executed, in execution, completed, etc.), and the time nodes of the task (such as planned completion time, actual completion time, estimated delay time, etc.). The engineering area meteorological data can be various non-restrictive meteorological parameters that may affect the project progress and resource consumption, such as temperature, wind speed, and precipitation.

[0035] The resource consumption data includes the building material cost consumption information, engineering equipment cost usage information, and engineering labor cost usage information of the corresponding engineering task. The building material cost consumption information can be the consumption quantity and cost data of the building materials (such as steel, cement, cables, etc.) used in the engineering task. The engineering equipment cost usage information is used to record the usage duration, rental cost, maintenance cost, fuel consumption, etc. of the equipment (such as cranes, excavators, concrete mixers, etc.) during the implementation of the engineering task. The engineering labor cost usage information can record the number of laborers, types of work, salaries, and work-hour efficiency of the engineering task.

[0036] In step S120, multiple engineering task multimodal data and task dependencies are processed based on digital twin technology to construct a virtual model of the power project.

[0037] Here, the virtual model of the power project includes multiple virtual modules of engineering tasks. Each virtual module of an engineering task is respectively modeled from the corresponding engineering task multimodal data, and there are edge connections between the associated virtual modules of engineering tasks indicated by the task dependencies.

[0038] In some embodiments, based on multimodal data, the entire power engineering project is virtually modeled by using digital twin technology. Each engineering task is abstracted into a virtual module, and the attributes of the virtual module include task progress, resource consumption, meteorological data, etc. Exemplarily, various third-party digital twin platforms can be adopted, such as Siemens MindSphere, PTC ThingWorx, etc., which can provide real-time task status feedback for users by constructing a visual dashboard. In addition, when modeling, the dependencies between engineering tasks are represented by a graph structure, specifically manifested as a graph composed of nodes (engineering tasks) and edges (relationships between tasks). It should be noted that the tasks in power engineering often have complex dependencies, and using a graph structure can effectively express the interconnections between tasks, especially the sequential dependencies, parallel executions, and resource sharing relationships between tasks. In addition, each virtual module not only contains initial task data but also can be dynamically updated according to the real-time collected task progress and resource consumption data, so as to maintain real-time synchronization with the status of each task in the actual power engineering project.

[0039] In step S130, the virtual model of the power engineering is input into a graph neural network to extract the task node features corresponding to each virtual module of the engineering task through a message passing mechanism.

[0040] It should be noted that a graph neural network (GNN) is a deep learning model specifically used to process graph-structured data. A graph is a data structure composed of nodes (vertices) and edges (edges), where nodes represent entities and edges represent the relationships between entities. Through a graph neural network, information can be propagated and aggregated on the nodes and edges of the graph to learn the feature representation of the graph, capturing the complex topological structure information and interdependent relationships between nodes.

[0041] In this embodiment, each virtual module of the power engineering task is regarded as a node in the graph, and the dependencies between tasks (such as sequence and resource sharing) are used as edges to connect the nodes. The GNN propagates information through a message passing mechanism, transmitting the feature information between nodes, and then extracting the task features of each node (i.e., the engineering task), which includes information such as progress, resource consumption, and meteorological task impact. Through the iterative propagation of the graph neural network, each task node can more accurately transmit the key information between tasks, update the status of the task, and achieve accurate capture and extraction of the dynamic features of the engineering task.

[0042] In step S140, the input state of the reinforcement learning model is defined according to the extracted task node features, so as to generate task resource allocation actions for each engineering task through reinforcement learning.

[0043] It should be noted that reinforcement learning is a field in machine learning that emphasizes how an agent takes a series of actions in an environment to maximize cumulative rewards. In this process, the agent learns the optimal behavior strategy by continuously interacting with the environment. The agent selects an action based on the current state. After executing this action, the environment will feedback a reward signal, and at the same time, the environment state will also be updated. The goal of the agent is to learn a policy that maximizes the cumulative rewards obtained during the long-term interaction process.

[0044] Figure 2 The schematic diagram of the principle shows an example of the state transition action in the reinforcement learning model.

[0045] As Figure 2 shown, it involves the actions corresponding to the state transitions occurring in the state space of multiple states S1 to Sn. For example, a1 represents the action of the state transition from S1 to S2, a2 represents the action from S2 to S1, a3 represents the state transition from S1 to S3, and so on. Here, the corresponding state transitions can occur based on the policy, and each state transition policy can be used to occur different transitions respectively. Exemplarily, based on the transition policy for state S1, actions a2 or a3 can occur.

[0046] It should be noted that the range of another state (also known as the transferable state) that a state can transfer to may be restricted or conditional. For example, no state transition will occur between any of S1 to S3 and S4 to Sn, and the states that state S1 can transfer to are S2 and S3, and so on.

[0047] In some embodiments, each action has a corresponding action reward, and each action reward can be determined based on a preset reward function. Generally, if the transfer reward is larger, it can be considered that the action of this transfer is more valuable, and the system will preferentially choose to execute this action. Exemplarily, if the reward corresponding to action a1 is greater than the reward corresponding to a3, it means that the transfer action a1 is more valuable.

[0048] In the embodiments of the present application, the task resource allocation action is used to indicate the resource allocation plan for the corresponding engineering task. The resource allocation plan includes the building material allocation strategy, the engineering equipment allocation strategy, and the labor force allocation strategy. In this way, the decision-making process of task resource allocation is regarded as an "action selection" problem, and the reinforcement learning model is used to learn the best resource allocation strategy by continuously trial and error. The input state of the reinforcement learning model is composed of the extracted features of each task node, including information such as task progress, resource consumption, and external meteorological factors, and these information jointly affect the decision-making of task resource allocation.

[0049] In addition, the reward function of the reinforcement learning model is defined based on the total cost of project completion, the total project completion period, and the average resource utilization efficiency of engineering tasks in the power project. Considering multiple optimization objectives, the resource allocation strategy is optimized through a trial-and-error process to maximize the comprehensive reward, thereby guiding the resource scheduling for different tasks and ensuring the on-time completion and cost control of the power engineering project. Thus, through the adaptive adjustment ability of the reinforcement learning model, intelligent resource allocation can be performed according to the real-time task status and the actual engineering situation, ensuring the reasonable allocation of resources among tasks and avoiding over-allocation or under-allocation.

[0050] Through the embodiments of the present application, by comprehensively applying digital twins, graph neural networks, and reinforcement learning models, the complex dependencies between engineering tasks are fully captured, ensuring that the relevance of each task is effectively modeled. Through dynamic data updates and task feature extraction, and introducing reinforcement learning for intelligent decision-making, the resource allocation becomes more reasonable. Thus, in the face of a complex engineering environment, an intelligent analysis solution for timely and accurate cost control can be provided, significantly improving the management efficiency of power engineering projects.

[0051] In some examples of the embodiments of the present application, the reward function of the reinforcement learning model is:

[0052]

[0053] where R t represents the reward value obtained by the reinforcement learning model according to the action policy for the input state at time step t; C total represents the total cost of project completion, and C max represents the maximum cost budget of the power project, T total represents the total project completion period, and T max represents the maximum planned project duration. U avg represents the average resource utilization efficiency of engineering tasks, and the larger U avg is, the higher the resource utilization rate, indicating that the resource scheduling and allocation are more reasonable, and the higher the reward at this time. U max represents the maximum value of the resource utilization efficiency of engineering tasks; w 1 , w 2 , w 3 represent weighting coefficients, which are used to control the importance of cost, cycle, and resource utilization efficiency in the reward function respectively; is an indicator function, which takes the value of 1 when the condition {·} is satisfied, and 0 otherwise; λ 1 and λ 2 represent the cost overrun penalty coefficient and the project duration overrun penalty coefficient respectively. For example, as the total cost of project completion increases, the penalty coefficient λ 1will also increase; in addition, as the completion cycle extends, the penalty coefficient λ 2 will also increase.

[0054] It should be noted that the total project completion cycle, the total project completion cost, and the average utilization efficiency of project task resources are important feedback data during the training process of the reinforcement learning model. These data can be jointly obtained through the action execution process of the reinforcement learning model and the dynamic simulation of the project task environment. Specifically, the environment dynamically updates according to the task progress based on the resource allocation plan of the reinforcement learning model and the current task state, and takes the maximum value of the completion times of all tasks to estimate the total project completion cycle T total . Additionally, the environment dynamically calculates the resource consumption of each task according to the resource allocation plan generated by the reinforcement learning model and the current task state, thereby obtaining the corresponding total project completion cost C total . Furthermore, the task resource utilization efficiency is calculated by the ratio of the task progress increment to the resource allocation amount, and U avg is used to represent the average value of the task resource utilization efficiency of all project tasks. The environment generates an allocation plan according to the actions of the reinforcement learning model and calculates the task resource utilization efficiency at each time step through the dynamic change of the task progress.

[0055] Thus, by combining the interaction between the reinforcement learning model and the project simulation environment, accurately calculating the project duration, cost, and efficiency through the simulation environment, providing an optimization direction for the reinforcement learning model, updating the task state in real time, and dynamically adjusting the resource allocation to reduce the project duration and cost. Furthermore, through the guidance of the reward function, multi-objective optimization of the project duration, cost, and resource efficiency is achieved.

[0056] In the embodiments of the present application, the reward function encourages the reduction of the total project completion cost C total . When the total cost gradually approaches the budget upper limit C max , the reward will decrease, thereby guiding the reinforcement learning model to actively optimize the resource allocation strategy, being able to detect the risk of cost overrun in real time, and triggering the optimization of the project resource allocation strategy. Additionally, through the design of the penalty term, if the actual cost exceeds the budget upper limit, an indicator function will be triggered to generate a fixed penalty value λ 1 , thereby further suppressing the behavior of exceeding the budget.

[0057] Moreover, the reward function encourages the project completion cycle T total to tend towards the maximum project planned duration T max and below. The shorter the cycle, the higher the reward. In this way, it can effectively guide the reinforcement learning model to preferentially select strategies that can accelerate the project progress when optimizing the resource allocation, such as concentrating resources to complete key tasks. Through the design of the penalty term, when T total exceeds T maxWhen triggered, a penalty is imposed, further strengthening the constraint on the construction period, preventing the model from neglecting the time target when pursuing cost or resource utilization rate, and ensuring that the power project can be completed within the planned time.

[0058] In addition, in the reward function, rewarding actions where the resource utilization rate is close to the maximum value U max (e.g., 100%) enables the reinforcement learning model to dynamically allocate resources, avoiding resource idleness or waste, ensuring the efficient use of resources (such as labor, materials, and machinery), improving the overall execution efficiency of the project, and preventing resource waste caused by unreasonable scheduling.

[0059] Through the embodiments of this application, based on the adjustment of w 1 , w 2 , w 3 , when optimizing one objective, the reward function does not neglect other objectives, enabling the system to comprehensively consider cost, construction period, and resource utilization efficiency during dynamic decision-making, and preventing the problem of global imbalance caused by over-optimizing a single objective. In addition, it can also meet the personalized management requirements for different power engineering projects, such as cost-sensitive, construction period-sensitive, etc. Exemplarily, in a project with a tight construction period, the weight of w 2 can be increased to prioritize the optimization of the construction period; in a budget-sensitive project, the weight of w 1 can be increased to focus on cost control.

[0060] Figure 3 shows an operation flowchart of an example according to Figure 1 step S110 therein.

[0061] As Figure 3 shown, in step S310, task data collection instructions are periodically sent to each preset task management terminal, and each task management terminal is respectively used to maintain the data update of the corresponding engineering task.

[0062] In some embodiments, the task management terminal may refer to one or more specific terminal devices, or it may refer to one or more clients with specific user accounts. In some scenarios, a power engineering project involves the collaboration of multiple different departments, such as a civil engineering group, a cable laying group, an equipment installation group, a communication debugging group, etc. Each department uploads the task-related data of its corresponding department by maintaining the corresponding task management terminal, realizing the real-time synchronization and sharing of the task data of each engineering task within the global scope of the engineering project.

[0063] In some embodiments, the system periodically (e.g., daily or weekly) sends task data collection instructions to each task management terminal, enabling administrators of each department to upload multi-modal data of engineering tasks through the task management terminal after receiving the instructions, such as task progress, resource usage, meteorological conditions, etc.

[0064] In step S320, in response to the task data collection instructions, corresponding multi-modal data of engineering tasks are received from each task management terminal.

[0065] In some embodiments, the system stores the data received from the task management terminals. For example, the data from different task management terminals are integrated according to task IDs, timestamps, etc. to form a complete multi-modal data packet for each task, and then stored in a cloud database or a distributed database.

[0066] In step S330, the engineering task tree structure of the power project is obtained. The engineering task tree structure includes the decomposition level information and task plan information of each task.

[0067] Here, the decomposition level information includes the parent-child node relationship of the tasks, and the task plan information includes the planned start time, planned end time, and task resource requirements of the tasks.

[0068] It should be noted that the engineering task tree is a tree-like hierarchical structure, which reflects the decomposition relationship of engineering tasks. The root node represents the entire power engineering project, and each child node represents a specific engineering task or task group. The construction of the task tree is usually designed by project managers through project management tools (such as Microsoft Project, Primavera, etc.) in the project planning stage. Each task node contains sub-tasks, parent-tasks, and execution order information, enabling the system to effectively extract the task hierarchical relationship of the power engineering project and clarify the position of each task node in the entire project. In addition, each task node not only includes hierarchical relationship information, but also includes the planned start time, planned end time, required resources, etc. of the task, which can be preset according to the project plan arrangement.

[0069] In step S340, according to the engineering task tree structure, by quantifying the dependence strength between tasks, the task dependence relationship between each engineering task is determined.

[0070] It should be noted that the dependence relationship between different engineering tasks can be diversified, such as sequential dependence, parallel dependence, resource sharing, etc. By combining the parent-child task relationship, task plan information in the task tree structure, and constraints in project management (such as project delivery date, resource limitation, etc.), the dependence strength between every two tasks is quantified.

[0071] Exemplarily, the dependency strength between tasks can be quantitatively represented by the following formula:

[0072]

[0073] Where: Weight ij represents the dependency strength or weight between tasks v i and v j ; Level ij represents the association strength between tasks v i and v j at the task decomposition level; TimeGap ij is the time interval between tasks v i and v j , which is used to measure the tightness of task scheduling in terms of time.

[0074] For the association strength Level ij at the decomposition level, it can be defined by matching preset rules. Exemplarily, if v i is the direct parent task of v j , then Level ij = 1; if v i and v j are subtasks of the same parent task, then Level ij = 0.5; if v i and v j are on different decomposition paths and have no direct parent-child relationship, then Level ij = 0.

[0075] For TimeGap ij , exemplarily, if the start time of task v j is later than the end time of v i , then the time interval is positive; if the start time of task v j is earlier than or equal to the end time of v i , then the time interval is 0 (indicating no time delay or overlapping execution).

[0076] Through the embodiments of the present application, by combining the decomposition level and task schedule time information to quantify the dependency strength between tasks, it is ensured that the model accurately reflects the actual logical relationship of engineering tasks, providing key reference information for subsequent task scheduling optimization.

[0077] In some examples of the embodiments of the present application, when it is detected that the multimodal data of the first engineering task corresponding to the first engineering task is updated, based on the digital twin technology, the updated multimodal data of the first engineering task is re-modeled to trigger the re-generation of task resource allocation actions for each engineering task.

[0078] In the context of business application scenarios, when administrators of each department upload or update multi-modal engineering task data to the system platform daily or weekly, the digital twin technology is used to re-model the updated multi-modal engineering data, enabling each task and link of the project to be presented in real time in the virtual model. Managers can understand the current status and resource consumption of each task at any time, improving the transparency of project management. In addition, based on the re-modeled virtual model of the power project, the feature extraction of the graph neural network is triggered again, and the dynamic decision-making adjustment of resource allocation is re-performed through the reinforcement learning model. In this way, after the new resource allocation plan is generated, the system will feedback the optimized resource configuration to the task management terminal to guide the scheduling and allocation of various resources (such as manpower, equipment, materials) to ensure the completion of tasks on schedule. For example, if the system detects potential risks, it can avoid the expansion of project risks through real-time warning and resource adjustment and allocation.

[0079] Figure 4 The operation flowchart of an example of constructing a virtual model of a power project based on digital twin technology according to an embodiment of the present application is shown.

[0080] As Figure 4 shown, in step S410, the task status vectors corresponding to the multi-modal data of each engineering task are extracted. The task status vectors include the task completion percentage, the resource consumption deviation, and the impact degree of meteorological tasks.

[0081] Here, the resource consumption deviation defines the difference between the resource consumption data of the corresponding engineering task and the expected resource consumption data. If the actual resource consumption of the task is 10% higher than the expected, the deviation is +10%. The impact degree of meteorological tasks defines the degree of interference of the meteorological data in the project area on the corresponding engineering task. For example, too high wind speed may affect the safety of high-altitude operations, and too much precipitation may cause delays in civil engineering progress. The task completion percentage in the task status vector reflects the workload completed during the execution of the task.

[0082] In step S420, according to each task status vector, the corresponding virtual module of the engineering task is rendered in real time.

[0083] It should be noted that the virtual module is the digital representation of the engineering task in the digital twin system. Each task is rendered as a virtual module through its corresponding task status vector, thereby dynamically displaying the execution of each task. For example, the progress of the task will be adjusted in real time as the completion percentage is updated, the deviation of resource consumption will affect the consumption progress bar of the relevant resources in the virtual module, and the change in the impact degree of meteorological tasks will affect the status identifier of the task. In addition, during the rendering process, the system will update the virtual module in real time according to the data in the task status vector.

[0084] In step S430, edge connections are built between the indicated associated engineering task virtual modules according to the task dependencies, and task priorities are marked for each engineering task virtual module to build a virtual power engineering model.

[0085] Here, the edge weight corresponding to the edge connection is defined by the dependency strength between the engineering tasks indicated by the task dependencies. For the description of the dependency strength between tasks and the relevant calculation details, part of it can refer to the description in combination with other examples in the above text, and will not be elaborated here.

[0086] In some embodiments, the system automatically assigns priorities to each task virtual module according to the criticality, resource consumption, and dependencies of the tasks. A task with a higher priority usually means that it has a greater impact on the overall progress of the project, or its delay may cause other tasks to lag. Therefore, the task priority will directly affect resource allocation and task scheduling, and tasks with higher priorities will be given more resources by the system for guarantee.

[0087] Through the embodiments of the present application, the virtual power engineering model constructed based on the digital twin technology can reflect the status of tasks, resource consumption, and changes in the external environment in real time during the task execution process. The real-time rendering of task virtual modules and the accurate modeling of task dependencies enable the system to flexibly respond to changes in task progress, optimize resource allocation, and provide clear decision-making support for project managers. Ultimately, the real-time feedback and automated optimization of the system will greatly improve the efficiency of project management and the completion degree of the project.

[0088] Regarding the details of the marking of task priorities, in some examples of the embodiments of the present application, the path cumulative impact scores and task dependency path depths corresponding to each engineering task virtual module are obtained, and task priorities are marked for the corresponding engineering task virtual modules according to the path cumulative impact scores and task dependency path depths.

[0089] Exemplarily, the calculation formula for task priority is:

[0090]

[0091] In the formula, pr i represents the task priority of the i-th engineering task virtual module, All_Dts(i) represents the set of all subsequent engineering task virtual modules of the engineering task virtual module i. Depth(j) is the task dependency path depth of the subsequent engineering task virtual module j relative to the engineering task virtual module i, which is represented by the virtual module hierarchical distance, and the path impact decays according to the hierarchical depth Depth(j) between tasks, and the contribution of deep-level tasks to i is smaller. Denotes the edge weight between the k-th virtual module of the engineering task and the l-th virtual module of the engineering task. Is the cumulative impact score of the path corresponding to i, representing the product of all edge weights on the path from i to j. For all subsequent j of i, the impact strength of the path from i to j is evaluated. By calculating the transmission impact using the product of path weights, the chain effect of dependencies can be reflected.

[0092] Through the cumulative path impact score, the chain transmission impact of task i on all its subsequent tasks j is captured. The more complex the path and the lower the edge weight, the weaker the cumulative impact. Thus, it avoids the interference of over - focusing on deep - level tasks on the priority ranking and makes the priority score more reasonable.

[0093] By the depth of the task - dependency path, a depth - attenuation term is constructed. The greater the depth, the smaller the transmission impact, indicating that the dependence of deep - level tasks on upstream tasks gradually weakens. When the dependency paths of some tasks become more complex or the risk of delay propagation of some tasks increases, the priority calculation will be dynamically adjusted to reflect the change in task importance.

[0094] Through the embodiments of the present application, by combining the cumulative path weight and the path depth, the complex dependency relationship between tasks can be dynamically evaluated. High - priority tasks can be quickly identified due to their large cumulative impact scores and low path depths. Therefore, by dynamically considering the task - dependency structure and depth impact for priority annotation, the task ranking no longer only focuses on a single task, but considers its global impact in the entire task network, which helps to achieve the global optimization of the engineering goal (such as achieving the optimal cost without delaying the project duration).

[0095] In some examples of the embodiments of the present application, the graph neural network adopts a message - passing mechanism based on a time - dynamic graph. Specifically, according to the task - status vectors of each task virtual module, the node features of each task virtual module are initialized:

[0096]

[0097] In the formula, Represents the initial node feature of the virtual module i of the engineering task at the corresponding time step t, Represents the task - completion percentage of i at the corresponding time step t, Represents the deviation of resource consumption of i at the corresponding time step t, Represents the impact degree of the meteorological task of i at the corresponding time step t, T i current Represents the current time progress of the task of i, T i expected Represents the expected end time of the task of i.

[0098] Dynamically update the edge weights of edge connections in the following manner:

[0099]

[0100] In the formula, represents the initial edge weight between the virtual module i of the engineering task and the virtual module j of the subsequent engineering task, pr j represents the task priority of the virtual module j of the engineering task, and respectively represent the updated edge weight, resource sharing intensity factor, and delay propagation risk factor between i and j at the corresponding time step t; γ 1 and γ 2 respectively represent the adjustment weight parameters for the influence of resource sharing intensity and delay propagation risk on the edge weight, and β represents the priority sensitivity parameter; represents the amount of shared resources provided by the engineering task corresponding to i to the engineering task corresponding to j at time step t; represents the total resource usage of the engineering task corresponding to i at time step t; represents the delay time of the engineering task corresponding to i at time step t, Plan_Duration i represents the planned duration of the engineering task corresponding to i.

[0101] It should be noted that the resource sharing intensity factor and the delay propagation risk factor are the most crucial parts in the dynamic edge weight formula. They directly affect the dependence strength between task nodes and are dynamically changing.

[0102] Specifically, the resource sharing intensity factor represents the degree of resource support provided by i to j at time step t. When the amount of shared resources provided by i to j approaches its total resource usage, approaches 1, indicating a very high resource support intensity; when i does not provide any resources to j Thus, by reflecting the actual dependence degree in terms of resources through the resource sharing intensity factor, it can ensure that under limited resources, the dependence relationship can accurately reflect the actual resource support degree between tasks, thereby optimizing task scheduling and resource allocation.

[0103] The delay propagation risk factor represents the degree of negative impact of the task delay of i on the task of j. In engineering tasks, the delay of the previous task often causes the subsequent task to not start on time, and even leads to a global project duration delay. When there is no task delay for i indicates that the delay risk is zero. When the task delay time of i is longer, The larger it is, the higher the risk of delay. Thus, the risk factor of delay propagation quantifies the direct impact of task delay on subsequent tasks, enabling the model to accurately capture the project duration risk, optimize task dependencies, and reduce the risk of overall project delay.

[0104] In addition, by incorporating the task priority difference |Pr i -Pr j | into the dynamic edge weight formula, the model can explicitly consider the difference in urgency between tasks. Between tasks with a large priority difference, the strength of the dependency is appropriately weakened, which helps to ensure the priority of high-priority tasks in resource scheduling.

[0105] Define the message received by each virtual module of the engineering task from the virtual module of the subsequent engineering task at time step t as:

[0106]

[0107] In the formula, represents the dynamic attention weight between i and j at the corresponding time step t, ‖ represents the vector concatenation operation, represents the initial node feature of j at time step t, a represents the learnable attention vector, represents the message aggregated to i from the virtual module of the subsequent engineering task at the corresponding time step t; is the normalization term to ensure

[0108] Here, the message passing mechanism weights the messages of neighbor nodes through the dynamic attention mechanism, combining task features and edge weights, enabling each task node to prioritize the neighbor nodes that have the greatest impact on it during feature update. In the dynamic attention mechanism, numerical stability is ensured through normalization operations, while explicitly reflecting the feature contribution strength of neighbor nodes, improving the accuracy of message passing and the reliability of task node feature updates.

[0109] Capture the historical temporal features of nodes through the LSTM module, and update the node features by fusing the aggregation results of message passing:

[0110]

[0111] In the formula, W s represents the learnable weight matrix, σ represents the activation function, represents the historical temporal feature of i at the corresponding time step t, △T represents the preset time window length, represents the temporal feature analysis of each historical feature of i within the time window through the LSTM module; represents the task node feature of the i-th virtual module of the engineering task at the corresponding time step t.

[0112] It should be noted that in the embodiments of the present application, the dynamic graph network adopts a structure modeled by a temporal graph, and maintains the corresponding task dependency graph G at different time steps t =(V, E t ), and introduces temporal connections between time steps to model the temporal dynamic evolution process of task node features. Calculate the dynamic state of the current task through the features of the direct neighbors of the task node to achieve local dependency aggregation; at the same time, through time series modeling based on the LSTM module, capture the historical features of the task node to achieve time series feature fusion.

[0113] Through the embodiments of the present application, by introducing a dynamic edge weight update mechanism, the initial static task dependency relationship is combined with dynamic adjustment factors (task priority, resource sharing intensity, and delay propagation risk) to more accurately quantify the dependency relationship between task nodes, and the dynamic update of the dependency relationship can help decision-makers quickly identify critical task links. By introducing the LSTM module to model the historical features of the nodes, the historical state of the task node and the current dependency relationship are fused to dynamically optimize the connection strength between tasks, capture the state evolution law of the task node at multiple time steps, and make up for the limitations of the state at a single time point.

[0114] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of actions combined. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present application. In the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0115] Figure 5 Shows a schematic structural connection diagram of an example of a reinforcement learning model according to an embodiment of the present application.

[0116] As Figure 5 shown, the reinforcement learning model 500 includes a resource multi-head allocation network 510, a policy constraint adjustment network 520, and a task multi-head output network 530. Each head (511, 513... 51n) of the resource multi-head allocation network 510 corresponds to a unique resource type respectively, and each head (531, 533... 53m) of the task multi-head output network 530 corresponds to a unique engineering task respectively.

[0117] Each head (511, 513…51n) of the resource multi-head allocation network 510 is used to determine the initial allocation strategy of each engineering task for the corresponding resource type; the resource types include construction material resources, engineering equipment resources, and labor human resources.

[0118] Exemplarily, for the three main types of resources (construction materials, engineering equipment, and labor) in a power engineering task, three independent allocation heads are designed, and each allocation head generates an initial allocation strategy for the corresponding resource type. It should be understood that the number of resource types can also be increased or adjusted according to the requirements of the engineering project. Through the multi-head design of resource types, it is possible to independently model the characteristics of different resource types and avoid interference between resource allocation strategies. Thus, for each task, an initial allocation ratio is generated according to its characteristics and the global resource situation, which can ensure the logical rationality of resource allocation. For example, high-priority tasks and tasks with large resource requirements should receive more allocations.

[0119] The policy constraint adjustment network 520 is used to count the total amount of policy resource allocation corresponding to the initial allocation strategy of each engineering task for the corresponding resource type, and when it is detected that the total amount of the first policy resource allocation corresponding to the first resource type exceeds the threshold of the total amount of available policy resources, the first initial allocation strategy for the first resource type is adjusted according to the task priorities corresponding to each engineering task.

[0120] It should be noted that the main function of the policy constraint adjustment network is to detect whether the initial allocation strategy exceeds the global resource limit and dynamically adjust the allocation strategy to meet the constraint conditions when it exceeds the limit. For example, the allocation ratio is adjusted according to the task priority (e.g., Pr i ) so that high-priority tasks can obtain resource allocations first. Thus, it is ensured that the final resource allocation plan meets the global resource constraints and avoids resource overload. In addition, in the case of resource constraints, key tasks are supported first to improve the overall efficiency of the project.

[0121] Each head (531, 533…53m) of the task multi-head output network 530 is respectively used to fuse the respective allocation strategies of the corresponding engineering tasks to output the corresponding task resource allocation actions, and the task resource allocation actions are used to indicate the resource allocation plan for the corresponding engineering task.

[0122] It should be noted that the main function of the task multi-head output network is to fuse the allocation strategies of each task for different resource types. For example, the allocation strategies of each task for the three types of resources are fused to generate the final resource allocation action to guide the actual engineering resource allocation.

[0123] Through the reinforcement learning model structure provided by the embodiments of the present application, the targeted and flexible resource allocation is ensured based on the multi-head network design. The network is adjusted in real time to respond to resource limitations by using policy constraints to achieve global optimization. The allocation logic driven by task priorities improves the completion efficiency of key tasks, thereby ensuring the overall project duration and cost optimization. Finally, the resource allocation plan output based on the model can provide refined and intelligent support for project management.

[0124] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a combination of a series of actions. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present application. In the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0125] Figure 6 A structural block diagram of an example of a simulation analysis system for the cost of a power project based on digital twins according to an embodiment of the present application is shown.

[0126] As Figure 6 shown, the simulation analysis system 600 for the cost of a power project based on digital twins includes a data acquisition unit 610, a twin modeling unit 620, a feature extraction unit 630, and a reinforcement learning analysis unit 640.

[0127] The data acquisition unit 610 is used to acquire multi-modal data of multiple engineering tasks for the power project and task dependencies; the multi-modal data of the engineering tasks includes task progress data, resource consumption data, and engineering area meteorological data of the corresponding engineering tasks; the resource consumption data includes building material cost consumption information, engineering equipment cost usage information, and engineering labor cost usage information of the corresponding engineering tasks.

[0128] The twin modeling unit 620 is used to process the multi-modal data of the multiple engineering tasks and task dependencies based on digital twin technology to construct a virtual model of the power project; the virtual model of the power project includes multiple virtual modules of engineering tasks, and each virtual module of the engineering task is respectively modeled by the corresponding multi-modal data of the engineering task, and there are edge connections between the associated virtual modules of the engineering tasks indicated by the task dependencies.

[0129] The feature extraction unit 630 is used to input the virtual model of the power project into a graph neural network to extract the task node features corresponding to each virtual module of the engineering task through a message passing mechanism.

[0130] The reinforcement learning analysis unit 640 is used to define the input state of the reinforcement learning model according to the extracted task node features of each item, so as to generate task resource allocation actions for each engineering task through reinforcement learning. The task resource allocation actions are used to indicate the resource allocation scheme for the corresponding engineering task, and the resource allocation scheme includes a building material allocation strategy, an engineering equipment allocation strategy, and a labor force allocation strategy; the reward function of the reinforcement learning model is defined according to the total project completion cost, the total project completion period, and the average utilization efficiency of engineering task resources of the power engineering project.

[0131] In some embodiments, the embodiments of the present application provide a non-volatile computer-readable storage medium, in which one or more programs including execution instructions are stored. The execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to be used for executing the steps of any one of the above-mentioned simulation analysis methods for the cost of power engineering projects based on digital twins.

[0132] In some embodiments, the embodiments of the present application further provide a computer program product. The computer program product includes a computer program stored on a non-volatile computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer is enabled to execute the steps of any one of the above-mentioned simulation analysis methods for the cost of power engineering projects based on digital twins.

[0133] In some embodiments, the embodiments of the present application further provide an electronic device, which includes: at least one processor, and a memory communicatively connected to the at least one processor. Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the steps of the simulation analysis method for the cost of power engineering projects based on digital twins.

[0134] Figure 7 It is a schematic diagram of the hardware structure of an electronic device for executing the simulation analysis method for the cost of power engineering projects based on digital twins provided by another embodiment of the present application. As Figure 7 shown, the device includes:

[0135] One or more processors 710 and a memory 720, Figure 7 Taking one processor 710 as an example.

[0136] The device for executing the simulation analysis method for the cost of power engineering projects based on digital twins may further include: an input device 730 and an output device 740.

[0137] The processor 710, the memory 720, the input device 730, and the output device 740 can be connected via a bus or other means. Figure 7 Take the connection via the bus as an example.

[0138] As a non-volatile computer-readable storage medium, the memory 720 can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the simulation analysis method for the cost of power engineering based on digital twins in the embodiments of the present application. By running the non-volatile software programs, instructions, and modules stored in the memory 720, the processor 710 executes various functional applications and data processing of the server, that is, implements the simulation analysis method for the cost of power engineering based on digital twins in the above method embodiments.

[0139] The memory 720 can include a program storage area and a data storage area. Among them, the program storage area can store an operating system and application programs required for at least one function; the data storage area can store data created according to the use of the electronic device. In addition, the memory 720 can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the memory 720 can optionally include a memory remotely set relative to the processor 710, and these remote memories can be connected to the electronic device through a network. Examples of the above networks include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0140] The input device 730 can receive input digital or character information, and generate signals related to the user settings and function control of the electronic device. The output device 740 can include a display device such as a display screen.

[0141] The one or more modules are stored in the memory 720, and when executed by the one or more processors 710, execute the simulation analysis method for the cost of power engineering based on digital twins in any of the above method embodiments.

[0142] The above product can execute the method provided in the embodiments of the present application, and has the corresponding functional modules and beneficial effects of the executed method. For technical details not described in detail in this embodiment, reference can be made to the method provided in the embodiments of the present application.

[0143] The electronic device in the embodiments of the present application exists in various forms, including but not limited to:

[0144] (1) Mobile communication devices: These devices are characterized by having mobile communication functions and mainly aiming to provide voice and data communication. Such terminals include: smart phones, multimedia phones, functional phones, and low-end phones, etc.

[0145] (2) Ultra-mobile personal computer devices: Such devices fall within the category of personal computers, have computing and processing capabilities, and generally also possess the characteristic of mobile Internet access. Such terminals include: PDA, MID, and UMPC devices, etc.

[0146] (3) Portable entertainment devices: Such devices can display and play multimedia content. Such devices include: audio and video players, handheld game consoles, e-books, as well as smart toys and portable in-vehicle navigation devices.

[0147] (4) Other on-board electronic devices with data interaction functions, such as in-vehicle device installed on a vehicle.

[0148] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0149] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0150] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

Claims

1. A simulation analysis method for power engineering cost based on digital twin, comprising: Obtain multimodal data and task dependencies for multiple engineering tasks for power engineering; The engineering task multimodal data includes task progress data, resource consumption data and engineering area meteorological data of the corresponding engineering task; the resource consumption data includes construction material cost consumption information, engineering equipment cost usage information and engineering manpower cost usage information of the corresponding engineering task; Processing the plurality of engineering task multimodal data and task dependency relationships based on digital twin technology to construct a power engineering virtual model; the power engineering virtual model comprises a plurality of engineering task virtual modules, each of the engineering task virtual modules is modeled by corresponding engineering task multimodal data, and there are edge connections between the associated engineering task virtual modules indicated by the task dependency relationships; Inputting the electric power engineering virtual model into the graph neural network to extract the task node features corresponding to each of the engineering task virtual modules through a message passing mechanism; Defining the input state of the reinforcement learning model according to the extracted features of each of the task nodes, so as to generate a task resource allocation action for each of the engineering tasks through reinforcement learning, wherein the task resource allocation action is used to indicate a resource allocation plan for the corresponding engineering task, and the resource allocation plan includes a construction material allocation strategy, an engineering equipment allocation strategy, and a labor allocation strategy; The reward function of the reinforcement learning model is defined based on the total cost of completing the project, the total cycle of completing the project, and the average utilization efficiency of project task resources of the power project.

2. The method according to claim 1, wherein: The reward function of the reinforcement learning model is: In the formula, R t represents the reward value obtained by the reinforcement learning model according to the action strategy for the input state at time step t; C total represents the total cost of the completed project, C max represents the maximum cost budget of the power project, T total represents the total period of project completion, T max Indicates the maximum duration of the project plan, U avg represents the average utilization efficiency of engineering task resources, U max represents the maximum value of resource utilization efficiency of engineering tasks; w1, w2, w3 represent weighting coefficients, which are used to control the importance of cost, cycle and resource utilization efficiency in the reward function respectively; is an indicator function. When the condition {·} is met, the indicator function takes the value of 1, otherwise it takes the value of 0. λ1 and λ2 represent the penalty coefficient for cost overrun and the penalty coefficient for construction period overrun, respectively.

3. The method according to claim 1, wherein: The obtaining of multimodal data and task dependencies of multiple engineering tasks for electric power engineering includes: Periodically sending task data collection instructions to each preset task management terminal; each of the task management terminals is used to maintain data updates of corresponding engineering tasks; In response to the task data collection instruction, receiving corresponding engineering task multimodal data from each of the task management terminals; and Acquire the engineering task tree structure of the power engineering project, wherein the engineering task tree structure includes decomposition level information and task plan information of each task; the decomposition level information includes the parent-child node relationship of the task, and the task plan information includes the planned start time, planned end time and task resource requirements of the task; According to the engineering task tree structure, the task dependency relationship between each of the engineering tasks is determined by quantifying the dependency strength between tasks.

4. The method according to claim 3, further comprising: When it is detected that the first engineering task multimodal data corresponding to the first engineering task is updated, the updated first engineering task multimodal data is remodeled based on the digital twin technology to trigger the regeneration of task resource allocation actions for each of the engineering tasks.

5. The method according to claim 4, wherein: The method of processing the multimodal data of the plurality of engineering tasks and the task dependencies based on the digital twin technology to construct a virtual model of the power engineering project includes: Extracting the task state vector corresponding to each of the multimodal data of the engineering tasks, the task state vector includes the task completion percentage, the resource consumption deviation and the meteorological task influence; the resource consumption deviation defines the difference between the resource consumption data of the corresponding engineering task and the expected resource consumption data, and the meteorological task influence defines the interference degree of the meteorological data in the engineering area on the corresponding engineering task; According to each of the task state vectors, respectively render the corresponding engineering task virtual modules in real time; According to the task dependency relationship, edge connections are constructed between the indicated associated engineering task virtual modules, and task priorities are marked for each engineering task virtual module to construct a power engineering virtual model; wherein the edge weight corresponding to the edge connection is defined by the dependency strength between the engineering tasks indicated by the task dependency relationship.

6. The method according to claim 5, wherein: The step of marking the task priority for each engineering task virtual module includes: Obtaining the path cumulative impact score and the task dependency path depth corresponding to each engineering task virtual module, and marking the task priority for the corresponding engineering task virtual module according to the path cumulative impact score and the task dependency path depth; The calculation formula of task priority is: In the formula, pr i represents the task priority of the i-th engineering task virtual module, All_Dts(i) represents the set of all subsequent engineering task virtual modules of engineering task virtual module i; Depth(j) is the task dependency path depth of subsequent engineering task virtual module j relative to engineering task virtual module i, which is represented by the virtual module level distance; represents the edge weight between the kth engineering task virtual module and the lth engineering task virtual module; The cumulative influence score of the path corresponding to i, which represents the product of all edge weights on the path from i to j.

7. The method according to claim 6, wherein: The graph neural network adopts a message passing mechanism based on a time dynamic graph; According to the task state vector of each task virtual module, initialize the node characteristics of each task virtual module: In the formula, represents the initialization node characteristics of the engineering task virtual module i corresponding to time step t, represents the completion percentage of task i at time step t, represents the resource consumption deviation of i at time step t, represents the impact of meteorological task i at time step t, T i current Indicates the current time progress of task i, T i expected represents the expected end time of task i; The edge weights of edge connections are updated dynamically in the following way: In the formula, represents the initial edge weight between the engineering task virtual module i and the subsequent engineering task virtual module j, pr j represents the task priority of the engineering task virtual module j, and denote the updated edge weight, resource sharing intensity factor and delay propagation risk factor between i and j at the corresponding time step t, respectively; γ1 and γ2 denote the adjustment weight parameters of the influence of resource sharing intensity and delay propagation risk on edge weight, respectively; β denotes the priority sensitivity parameter; represents the amount of shared resources provided by the engineering task corresponding to i to the engineering task corresponding to j at time step t; represents the total resource usage of the engineering task corresponding to time step t,i; Plan_Duration represents the delay time of the engineering task corresponding to i at time step t. i represents the planned duration of the engineering task corresponding to i; The message received by each engineering task virtual module from the subsequent engineering task virtual module at time step t is defined as: In the formula, represents the dynamic attention weight between i and j at the corresponding time step t, ‖ represents the vector concatenation operation, represents the initialization node feature of j at time step t, a represents the learnable attention vector, Represents the message corresponding to time step t aggregated from the subsequent engineering task virtual module to i; is a normalization term to ensure The historical time series features of the nodes are captured through the LSTM module, and the node features are updated by integrating the aggregation results of message passing: Where W s represents the learnable weight matrix, σ represents the activation function, represents the historical time series characteristics of i corresponding to time step t, △T represents the preset time window length, Indicates that the LSTM module is used to analyze the time series characteristics of each historical feature of i in the time window; Represents the extracted task node features of the i-th engineering task virtual module corresponding to time step t.

8. The method according to claim 6, wherein: The reinforcement learning model includes a resource multi-head allocation network, a strategy constraint adjustment network and a task multi-head output network; each head of the resource multi-head allocation network corresponds to a unique resource type, and each head of the task multi-head output network corresponds to a unique engineering task; Each head of the resource multi-head allocation network is used to determine the initial allocation strategy of each engineering task in the corresponding resource type according to the input state; the resource types include building material resources, engineering equipment resources and labor human resources; The policy constraint adjustment network is used to count the total amount of policy resource allocation corresponding to the initial allocation policy of each engineering task under the corresponding resource type, and when it is detected that the total amount of first policy resource allocation corresponding to the first resource type exceeds the total amount of policy resources available for allocation, adjust the first initial allocation policy under the first resource type according to the task priority corresponding to each engineering task; Each head of the task multi-head output network is used to fuse various allocation strategies of the corresponding engineering tasks to output the corresponding task resource allocation action; the task resource allocation action is used to indicate the resource allocation plan for the corresponding engineering task.

9. A simulation analysis system for power engineering cost based on digital twin, comprising: A data acquisition unit, used to acquire multimodal data and task dependencies of multiple engineering tasks for power engineering; The engineering task multimodal data includes task progress data, resource consumption data and engineering area meteorological data of the corresponding engineering task; the resource consumption data includes construction material cost consumption information, engineering equipment cost usage information and engineering manpower cost usage information of the corresponding engineering task; A twin modeling unit, configured to process the plurality of engineering task multimodal data and task dependencies based on digital twin technology to construct a power engineering virtual model; the power engineering virtual model comprises a plurality of engineering task virtual modules, each of the engineering task virtual modules is modeled by corresponding engineering task multimodal data, and there are edge connections between the associated engineering task virtual modules indicated by the task dependencies; A feature extraction unit, used for inputting the electric power engineering virtual model into a graph neural network, so as to extract the task node features corresponding to each of the engineering task virtual modules through a message passing mechanism; A reinforcement learning analysis unit, used to define the input state of the reinforcement learning model according to the extracted features of each of the task nodes, so as to generate a task resource allocation action for each of the engineering tasks through reinforcement learning, wherein the task resource allocation action is used to indicate a resource allocation plan for the corresponding engineering task, and the resource allocation plan includes a construction material allocation strategy, an engineering equipment allocation strategy, and a labor allocation strategy; The reward function of the reinforcement learning model is defined based on the total cost of completing the project, the total cycle of completing the project, and the average utilization efficiency of project task resources of the power project.

Images