Multi-modal simulation method and system based on real-time data component

Abstract

The invention discloses a multi-modal simulation method and system based on a real-time data component, and relates to the technical field of power system simulation. Through construction of a multi-modal simulation application framework, data-driven cross-modal interactive modeling is realized, full-process configuration operation from visual design to lightweight deployment is supported, collaboration with an embedded development system of a unified research and development platform is ensured, a real-time collection and fusion system of multi-source heterogeneous data is established, and the real-time collection and fusion system of the multi-source heterogeneous data is realized. Filtering conversion and interactive visualization of data streams are achieved, balance of data real-time performance and display performance is ensured, state collaboration and self-adaptive adjustment of multi-modal components are achieved through linkage collaboration of a hardware simulation test environment and components of the platform, a linkage rule system conforming to power grid service logic is formed, equipment configuration service of the platform is achieved, and the service performance of the platform is improved. A cross-platform simulation operation environment is constructed, lightweight deployment and continuous optimization of applications are realized, the consistency of a simulation result and physical equipment is ensured, and the simulation efficiency and reliability are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of power system simulation technology, and in particular to a multimodal simulation method and system based on real-time data components. Background Technology

[0002] Simulation technology is playing an increasingly important role in power system equipment testing, system optimization, and training simulation. For example, patent application CN109190305B discloses a panoramic real-time simulation method for power systems, including: constructing primary and secondary system models of a substation respectively; writing the simulation data for each step of the primary system model into a real-time database; after the primary system model has calculated the simulation data for any step and written it into the real-time database, the secondary system model collects the data from the real-time database; simultaneously, the primary system model transmits the simulation data to a signal conversion device; if the primary system model receives the action information sent by the secondary system model and simultaneously receives the information returned by the signal conversion device to the primary system model, then simulation is performed based on the current topology. This patent coordinates and synchronizes the operation of primary system simulation, secondary equipment simulation, and some real secondary equipment, achieving interoperability and thus forming a complete simulation testing system that ensures seamless integration between the simulation platform and real secondary equipment and virtual digital protection.

[0003] However, although the aforementioned patents have achieved co-simulation of the primary and secondary systems, the following problems still exist:

[0004] Existing technologies lack flexible multimodal interaction design capabilities, making it difficult to quickly customize visual interfaces and interaction logic according to different business needs. They cannot meet the diverse simulation scenario requirements. In terms of simulation operation and deployment, they lack cross-platform adaptive optimization mechanisms, cannot dynamically adjust resource allocation according to system load, and are difficult to achieve lightweight deployment and continuous optimization. The simulation efficiency and reliability in complex power grid business scenarios urgently need to be improved. Summary of the Invention

[0005] The purpose of this invention is to provide a multimodal simulation method and system based on real-time data components. By utilizing real-time data and component linkage mechanisms, it simulates the dynamic changes of actual systems, improves the reliability of simulation results, supports multiple simulation models and scenarios, and allows users to customize components, thereby enhancing the flexibility and scalability of simulation and meeting the personalized needs of different users, thus solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] Multimodal simulation methods based on real-time data components include:

[0008] Build a multimodal application design environment: Build a multimodal simulation application framework, and manage the design panel, real-time data component library, and data binding and configuration based on the multimodal application designer;

[0009] Component linkage and collaboration: Based on data binding and configuration management, an event-driven linkage mechanism is built between components, allowing for the editing of component properties and the setting of linkage rules between components to enable multi-component collaborative work;

[0010] Simulation runtime optimization and deployment: Based on the inter-component linkage rule system, a cross-platform simulation runtime environment is built. According to the simulation objectives, different simulation scenarios are set to carry out lightweight deployment and continuous optimization of applications.

[0011] Furthermore, a multimodal application design environment is constructed, specifically including:

[0012] Design panel management: Based on the constructed multimodal simulation application framework, the page structure is decomposed into a basic layout layer, a component display layer, and an interaction logic layer. Panels are deployed in combination with the components of each component. The component display layer interfaces with a multi-source heterogeneous data acquisition and fusion system through standardized data interfaces.

[0013] Real-time data component library management: Establish a clear and scalable real-time data component library, which stores various preset components, determines the basic attributes of each preset component, and divides each preset component into component families based on the functions corresponding to the basic attributes. Each component family is set with corresponding category tags and keywords, and the preset components are classified and managed based on the component family category.

[0014] Data binding and configuration management: Determine the mapping relationship between the data source of each real-time data and the basic attributes of each preset component. Based on the mapping relationship, perform multi-source data fusion and mapping of real-time data from each data source to dynamically display the real-time data.

[0015] Furthermore, redundancy is assessed on the collected real-time data, and a real-time data acquisition anomaly alarm is triggered when redundancy is abnormal, including:

[0016] Retrieve real-time data collected from the data source;

[0017] Determine the standard deviation of the data volume collected in all windows based on the amount of data collected in each sliding window of the data source.

[0018] Retrieve the number of data collections for each sliding window;

[0019] The data deviation index is determined by using the standard deviation of the data volume corresponding to all window data collections corresponding to the data source and the number of data collections corresponding to each sliding window.

[0020] The data deviation index is compared with a preset data deviation index threshold.

[0021] When the data deviation index exceeds the preset data deviation index threshold, redundancy is determined, and a real-time data acquisition anomaly alarm is triggered when redundancy is abnormal.

[0022] Furthermore, when the data deviation index exceeds a preset data deviation index threshold, a redundancy determination is performed, including:

[0023] When the data deviation index exceeds the preset data deviation index threshold, the data deviation index is retrieved.

[0024] Retrieve the resource and energy consumption of data acquisition for each sliding window of the current data source;

[0025] The standard deviation of resource energy consumption is obtained by using the resource energy consumption of each sliding window corresponding to the current data acquisition from the data source.

[0026] The standard deviation of resource energy consumption corresponding to the data source is normalized to obtain the normalized standard deviation of resource energy consumption.

[0027] The resource energy consumption deviation is obtained by combining the normalized standard deviation of the resource energy consumption corresponding to the data source with the standard deviation of the collected data of m completed sliding windows, with the total amount of collected data corresponding to each sliding window as the benchmark.

[0028] A redundancy quantification function is obtained by utilizing the data deviation index and resource energy consumption deviation of the data source;

[0029] The value corresponding to the redundancy quantification function is compared with a preset function threshold;

[0030] When the value corresponding to the redundancy quantification function exceeds the preset function threshold, the redundancy level of the real-time data is determined to be abnormal.

[0031] Furthermore, data binding and configuration management also includes establishing a real-time acquisition and fusion system for multi-source heterogeneous data:

[0032] Real-time data is collected from various data sources, and the sampling interval of each data source is automatically adjusted according to the data change rate.

[0033] The collected real-time data is cleaned, transformed, and standardized to remove invalid data, correct outliers, and integrate to generate multi-source heterogeneous data.

[0034] Multi-source heterogeneous data is input into the entity semantic model to establish a semantic mapping of cross-source data. Multi-source data of the same entity is weighted and fused to generate a standardized data entity and determine the mapping relationship of fields from different data sources.

[0035] Based on the mapping relationship of different data source fields, determine the mapping rules between data sources and component properties, and based on the mapping rules between data sources and component properties, determine the mapping relationship between the data source of each real-time data and the basic properties of each preset component.

[0036] Furthermore, the multimodal application designer also includes:

[0037] Component attribute editing: Set the component visual attribute configuration interface, parametrically edit the basic visual attributes of the real-time data component, and set the interactive attributes of the real-time data component. The size of the visual attribute configuration interface is automatically adapted under the grid layout constraints, and the display parameters are recalculated when the component position is adjusted.

[0038] Component composition and decomposition: Combine multiple time-sequence, state-sequence, or spatial components into a composite component, generate a component hierarchy tree structure, and establish data interaction channels and event transmission links between sub-components during the composition process. At the same time, recursively decompose the composite component, obtain the attribute configuration and interaction logic mapping relationship before decomposition, and edit each component individually.

[0039] Furthermore, component linkage and collaboration specifically include:

[0040] Establish a standardized event description system. Based on the standard event metadata structure, semantically bind the rule triggering conditions with the type definitions in the event metadata. When a component triggers an event, the event is automatically distributed to the corresponding component.

[0041] Construct a rule knowledge graph, formalize the linkage rules into a graph structure, and predict the state change trend of components based on historical interaction data and linkage rules;

[0042] The system matches the corresponding linkage rules and strategies based on the prediction results, adjusts the weights of the linkage rules based on historical execution data, generates the optimal linkage rules based on the runtime environment, and evaluates the execution effect of the optimal linkage rules in real time and generates optimization suggestions.

[0043] Furthermore, the simulation operation optimization deployment also includes:

[0044] An adaptive resource allocation algorithm based on load prediction is adopted to dynamically adjust the computing resource allocation strategy by monitoring simulation scenario indicators.

[0045] Construct a consistency verification mechanism for a cross-platform simulation environment to achieve integrated encapsulation of the compilation toolchain, dependency libraries, and component runtime;

[0046] Design an interface adaptation layer for the hardware simulation test environment and real-time data components to realize the mapping and verification of communication protocols and electrical parameters between the simulation model and the IoT terminal.

[0047] This invention provides another technical solution: a multimodal simulation system based on real-time data components, comprising:

[0048] The real-time data acquisition module is configured to acquire real-time data from various data sources in the power system, including electrical and non-electrical quantities.

[0049] The data processing and analysis module is configured to perform standardized processing, semantic fusion, and trend prediction on the collected real-time data, generating standardized data entities that conform to the component interface specifications.

[0050] The multimodal simulation module is configured to construct simulation scenarios based on standardized data entities and combined with power system models to perform multimodal simulations of dynamic operating conditions and simulate the dynamic behavior of the power system.

[0051] The visualization module is configured to integrate time-series, state-based, and spatial component libraries. Through data binding and configuration management, it drives standardized data entities to visualization components and displays simulation results in a visual format.

[0052] Furthermore, the multimodal simulation module includes a multimodal application designer, a component linkage engine, and a cross-platform deployment engine.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] By constructing a multimodal simulation application framework, data-driven cross-modal interactive modeling is achieved, supporting full-process configuration operations from visual design to lightweight deployment. This ensures collaboration with the embedded development system of the unified R&D platform, establishes a real-time acquisition and fusion system for multi-source heterogeneous data, realizes data stream filtering and transformation and interactive visualization, ensures a balance between data real-time performance and display performance, connects to the platform's hardware simulation testing environment, enables component linkage and collaboration to achieve state coordination and adaptive adjustment of multimodal components, forms a linkage rule system that conforms to the power grid business logic, connects to the platform's equipment configuration service, constructs a cross-platform simulation operation environment, realizes lightweight deployment and continuous optimization of applications, and ensures consistency between simulation results and physical equipment. Attached Figure Description

[0055] Figure 1 This is a flowchart of the multimodal simulation method based on real-time data components of the present invention. Detailed Implementation

[0056] 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.

[0057] To address the technical problems of rigid design, inefficient data processing, and poor deployment adaptability inherent in traditional simulation methods, please refer to [link / reference]. Figure 1 This embodiment provides the following technical solution:

[0058] Multimodal simulation methods based on real-time data components include:

[0059] Build a multimodal application design environment: Construct a multimodal simulation application framework with hierarchical scalability, and manage design panels, real-time data component libraries, and data binding and configuration based on the multimodal application designer;

[0060] Component linkage and collaboration: Based on data binding and configuration management, an event-driven linkage mechanism is built between components, allowing for the editing of component properties and the setting of linkage rules between components to enable multi-component collaborative work;

[0061] Simulation operation optimization and deployment: Based on the linkage rule system between components, a cross-platform simulation operation environment is built. According to the simulation objectives, different simulation scenarios are set, such as equipment failure simulation, system load simulation, and environmental change simulation, to carry out lightweight deployment and continuous optimization of applications.

[0062] In this embodiment, by constructing a multimodal simulation application framework, data-driven cross-modal interactive modeling is achieved, supporting full-process configuration operations from visual design to lightweight deployment. This ensures collaboration with the embedded development system of the unified R&D platform, establishes a real-time acquisition and fusion system for multi-source heterogeneous data, realizes data stream filtering and interactive visualization, ensures a balance between data real-time performance and display performance, connects to the platform's hardware simulation testing environment, and enables component linkage and collaboration to achieve state coordination and adaptive adjustment of multimodal components, forming a linkage rule system that conforms to the power grid business logic. It also connects to the platform's equipment configuration service, constructs a cross-platform simulation operation environment, realizes lightweight deployment and continuous optimization of applications, ensures consistency between simulation results and physical equipment, and significantly improves simulation efficiency and reliability.

[0063] In this embodiment, constructing a multimodal application design environment specifically includes:

[0064] Design panel management: Based on the construction of a scalable multimodal simulation application framework, the page structure is decomposed into a basic layout layer, a component display layer, and an interaction logic layer. Panels are deployed in combination with the components of each component. The component display layer interfaces with a multi-source heterogeneous data acquisition and fusion system through standardized data interfaces.

[0065] Real-time data component library management: Establish a clearly categorized and scalable real-time data component library, which stores various preset components. Define the basic attributes of each preset component, and classify the preset components into component families such as time series (curves / bar charts), status (dashboards / switches), and spatial (3D topology / heat maps) based on the functions corresponding to the basic attributes. Each component family is set with corresponding category tags and keywords, and the preset components are classified and managed based on the component family category.

[0066] Data binding and configuration management: Determine the mapping relationship between the data source of each real-time data and the basic attributes of each preset component. Based on the mapping relationship, perform multi-source data fusion and mapping of real-time data from each data source to dynamically display the real-time data.

[0067] Specifically, redundancy is assessed in the collected real-time data, and a real-time data acquisition anomaly alarm is triggered when redundancy is abnormal, including:

[0068] Retrieve real-time data collected from the data source;

[0069] Determine the standard deviation of the data volume collected in all windows based on the amount of data collected in each sliding window of the data source.

[0070] Retrieve the number of data collections for each sliding window;

[0071] The data deviation index is determined by using the standard deviation of the data volume corresponding to all window data collections corresponding to the data source and the number of data collections corresponding to each sliding window.

[0072] The data deviation index is compared with a preset data deviation index threshold, wherein the data deviation index threshold is determined based on experience;

[0073] When the data deviation index exceeds the preset data deviation index threshold, redundancy is determined, and a real-time data acquisition anomaly alarm is triggered when redundancy is abnormal.

[0074] The technical effects of the above solution are as follows: Based on a sliding window mechanism, real-time data streams are dynamically analyzed in time segments, enabling the immediate capture of abnormal fluctuations during data acquisition and avoiding cumulative errors caused by delayed detection. Combining the standard deviation with the deviation index of the number of acquisitions, it can quickly identify situations where data volume fluctuations exceed the normal range, ensuring that abnormal states are detected immediately. By measuring the dispersion of data volume within the sliding window using the standard deviation, redundant features such as "abnormally concentrated or sparse data volume" (e.g., duplicate acquisition, missed acquisition) can be effectively identified. Introducing the number of acquisitions as a correction factor allows the deviation index to simultaneously reflect "data volume fluctuations" and "acquisition frequency stability," avoiding misjudgments based on a single dimension and improving the accuracy of redundancy detection. The entire process, from data retrieval and index calculation to threshold comparison, is automated, reducing manual intervention costs and the risk of human error, making it suitable for high-concurrency, high-frequency data acquisition scenarios. When the deviation index exceeds the threshold, a real-time alarm is triggered, which can be linked with the data acquisition system to achieve closed-loop control of "monitoring-judgment-response," ensuring the reliability of the data acquisition chain. Early identification of redundant data anomalies can prevent invalid data from occupying storage resources and consuming computing power, improving data quality from the source.

[0075] Specifically, when the data deviation index exceeds a preset data deviation index threshold, a redundancy determination is performed, including:

[0076] When the data deviation index exceeds a preset data deviation index threshold, the data deviation index is retrieved, wherein the data deviation index is obtained by the following formula:

[0077]

[0078] Where R represents the data deviation index; σ c σ represents the standard deviation of the collected data volume of m sliding windows, based on the total amount of data collected for each sliding window; n represents the number of data collections for each sliding window; σ i This represents the standard deviation of the data volume corresponding to n data acquisitions within the i-th sliding window; specifically, The global standard deviation σ c By averaging over the number of windows m, we obtain the "theoretical fluctuation benchmark that each window should have" (ideally, global and local fluctuations should be consistent). Local standard deviation σ i The average fluctuation of each window is obtained by averaging the number of data collections n; R represents the absolute difference between the theoretical fluctuation benchmark and the actual average fluctuation, which is used to reflect the degree of deviation of the global-local data collection stability.

[0079] Retrieve the resource consumption (such as memory utilization, bandwidth utilization and / or CPU utilization) of each sliding window in the current data source for data collection.

[0080] The standard deviation of resource energy consumption is obtained by using the resource energy consumption of each sliding window corresponding to the current data acquisition from the data source.

[0081] The standard deviation of resource energy consumption corresponding to the data source is normalized to obtain the normalized standard deviation of resource energy consumption.

[0082] The resource energy consumption deviation is obtained by combining the normalized standard deviation of the resource energy consumption corresponding to the data source with the standard deviation of the collected data of m completed sliding windows, with the total amount of collected data corresponding to each sliding window as the benchmark.

[0083] The resource energy consumption deviation is obtained by the following formula:

[0084] Y = log 10 (1+|σ e -σ gc |)

[0085] Where Y represents the deviation of resource energy consumption; σ e σ represents the standard deviation of resource energy consumption corresponding to the data source after normalization. gc This represents the standard deviation of the data collected by the m sliding windows corresponding to the normalized data source; specifically, |σ e -σ gc | Used to reflect the "coordination between data collection behavior and resource consumption" (ideally, resource consumption should also fluctuate synchronously when data volume fluctuates greatly); log 10 (1+|σ e -σ gc |) Using logarithmic functions to "compress extreme values ​​and amplify small differences" can both prevent outliers from dominating the results and sensitively capture "small collaborative anomalies" (such as no change in data volume, but a sudden increase in resource consumption, which may indicate redundant data collection).

[0086] A redundancy quantification function is obtained by utilizing the data deviation index and resource energy consumption deviation of the data source;

[0087] The redundancy metric function is obtained through the following formula:

[0088]

[0089] Where J represents the redundancy quantification function; λ represents the preset time sensitivity factor, with a value range of (0, 1); I t (τ) represents the time-shifted mutual information entropy of the data source; specifically, By taking the geometric mean of "data deviation" and "resource coordination anomaly", the anomalies in the two dimensions "reinforce each other" (the redundancy judgment is more significant when data anomalies and resource anomalies exist at the same time). The exponential function is used to introduce a negative adjustment for "time series dependence"; J comprehensively reflects the interrelationship between "severity of data anomalies" (R), "co-abnormality of resource consumption and data" (Y) and "reasonableness of time series" (It(τ)), and finally outputs a dynamic quantitative value to determine "whether the degree of redundancy is abnormal".

[0090] The value corresponding to the redundancy quantification function is compared with a preset function threshold, wherein the function threshold is determined based on experience;

[0091] When the value corresponding to the redundancy quantification function exceeds the preset function threshold, the redundancy level of the real-time data is determined to be abnormal.

[0092] The technical effects of the above solution are as follows: The above solution constructs a multi-dimensional correlation model of "data deviation - resource energy consumption - time-shift mutual information entropy". First, using a sliding window as the unit, the data deviation index (R) is calculated through the standard deviation of data volume and the number of collections to characterize the abnormal fluctuations in data collection volume. Then, the resource energy consumption dimension is introduced, and through standard deviation calculation, normalization, and deviation analysis (Y) from the standard deviation of data volume, the synergistic anomalies between data collection behavior and resource consumption are correlated. Finally, the time-sensitive factor (λ) and time-shift mutual information entropy I are fused. t (τ) Construct a redundancy quantification function (J) to integrate data fluctuations, resource consumption, and time series dependencies into a unified model, thereby achieving accurate quantitative judgment and anomaly alarm for the degree of redundancy. Compared to existing redundancy determination methods, which often rely solely on simple dimensions such as data volume repetition and timestamp intervals, resulting in insufficient dimensional coverage and weak scenario adaptability, the aforementioned technical solution breaks through the limitations of a single dimension. It employs a multi-indicator hierarchical fusion approach—using data deviation indicators to capture statistical anomalies in collected data, using resource energy consumption deviation to correlate data collection with system resource coordination anomalies, and introducing time-shifted mutual information entropy to incorporate the dependency characteristics of time series. Furthermore, it uses time-sensitive factors to adapt to the time sensitivity requirements of different scenarios. The constructed redundancy quantification function (J) more comprehensively and dynamically reflects the essential characteristics of data collection redundancy, improving the accuracy of redundancy determination and scenario adaptability. In addition, the entire process, from data fluctuation monitoring to resource correlation analysis and multi-factor fusion modeling, forms a complete logical closed loop from anomaly identification to redundancy quantification. This allows for more precise location of the root cause when redundancy anomalies occur, providing richer decision-making basis for subsequent data collection optimization and resource scheduling adjustments. This effectively compensates for the shortcomings of existing technologies in low redundancy determination accuracy and insufficient correlation analysis in complex scenarios.

[0093] In this embodiment, data binding and configuration management also includes establishing a real-time acquisition and fusion system for multi-source heterogeneous data:

[0094] Real-time data is collected from various data sources such as sensors, SCADA systems, and databases, and the sampling interval of each data source is automatically adjusted according to the data change rate.

[0095] The collected real-time data is cleaned, transformed, and standardized to remove invalid data, correct outliers, and integrate to generate multi-source heterogeneous data.

[0096] Multi-source heterogeneous data is input into the entity semantic model to establish a semantic mapping of cross-source data. Multi-source data of the same entity is weighted and fused to generate a standardized data entity, which can be dynamically called by time-series and state-type components in the real-time data component library to determine the mapping relationship of fields from different data sources.

[0097] Based on the mapping relationship of different data source fields, determine the mapping rules between data sources and component properties, and based on the mapping rules between data sources and component properties, determine the mapping relationship between the data source of each real-time data and the basic properties of each preset component.

[0098] In this embodiment, the basic layout layer provides various layout templates such as grid and flow, and supports custom grid parameters and responsive adaptation rules;

[0099] In this embodiment, the component display layer establishes standardized component mounting points and uses a coordinate mapping algorithm to achieve component positioning and dynamic size adjustment;

[0100] In this embodiment, the interaction logic layer reserves an event listening interface to support event mapping for input modes such as mouse and touch.

[0101] In this embodiment, the layered architecture design breaks the traditional rigid simulation interface mode, and the scalable framework and standardized interface enable flexible connection of multi-source data; the visual layout and component classification management greatly reduce the design threshold and improve the development efficiency of simulation applications; based on the dynamic sampling of data change rate and the entity semantic model fusion mechanism, it breaks through the traditional single data processing; and realizes real-time information collection, heterogeneous data fusion, data supply and analysis, and data link monitoring, providing real-time data support for simulation applications.

[0102] In this embodiment, the multimodal application designer further includes:

[0103] Component attribute editing: Set the component visual attribute configuration interface to parametrically edit the basic visual attributes of the real-time data component, such as color, size, position, and transparency, and set interactive attributes of the real-time data component, such as click events and hover tooltips; the visual attribute configuration interface automatically adapts to the size under grid layout constraints, and triggers a coordinate mapping algorithm to recalculate display parameters when the component position is adjusted;

[0104] Component composition and decomposition: Combine multiple time-sequence, state-sequence, or spatial components into a composite component, generate a component hierarchy tree structure, and establish data interaction channels and event transmission links between sub-components during the composition process. At the same time, recursively decompose the composite component, obtain the attribute configuration and interaction logic mapping relationship before decomposition, and edit each component individually.

[0105] In this embodiment, component attribute editing enables fine-grained control of real-time data component visualization and interactivity. Automatic grid layout adaptation and coordinate mapping algorithms ensure the display effect and interactive stability of components under different layouts. Component combination and splitting functions empower users to flexibly construct complex component structures. The hierarchical tree structure and data interaction channel design support efficient management and personalized customization of composite components, significantly improving the design flexibility and user experience of multimodal simulation applications. Furthermore, it deeply integrates the parametric editing of visual and interactive attributes and achieves layout adaptation through algorithms. The component combination and splitting mechanism constructs a unique hierarchical tree and data link system, forming a novel component management mode that effectively improves the development efficiency and functional scalability of multimodal simulation applications, making it suitable for applications in multiple scenarios such as power systems.

[0106] In this embodiment, the component linkage and collaboration specifically includes:

[0107] Establish a standardized event description system based on a standard event metadata structure, including elements such as event type, triggering conditions, and propagation rules. Semantically bind the rule triggering conditions with the type definition in the event metadata. When a component triggers an event, the event is automatically distributed to the corresponding component.

[0108] Construct a rule knowledge graph, formalize the linkage rules into a graph structure, and predict the state change trend of components based on historical interaction data and linkage rules;

[0109] Based on the prediction results, the corresponding linkage rule strategy is matched, and the weight of the linkage rule is adjusted according to historical execution data. The optimal linkage rule is generated based on the runtime environment to realize a dynamic load balancing mechanism. At the same time, the execution effect of the optimal linkage rule is evaluated in real time and optimization suggestions are generated.

[0110] In this embodiment, technologies such as semantic binding of event metadata, construction of rule knowledge graphs, and dynamic weight adjustment are integrated to form an innovative component linkage framework. Unlike traditional fixed linkage logic, it provides a new paradigm for multimodal simulation component collaboration by mining component relationships and dynamically optimizing rules through knowledge graph mining. In practical applications such as power systems, it can quickly respond to events such as equipment status changes and load fluctuations, realize intelligent component linkage, generate optimal rules and load balancing, effectively cope with high-concurrency scenarios, and improve system resource utilization. The real-time evaluation and feedback mechanism helps to continuously optimize linkage strategies, enhancing the practicality and business adaptability of the simulation system.

[0111] In this embodiment, the simulation operation optimization deployment also includes:

[0112] An adaptive resource allocation algorithm based on load prediction is adopted. By monitoring indicators such as CPU utilization, memory usage and data throughput in the simulation scenario, the allocation strategy of computing resources such as physical servers and Docker containers is dynamically adjusted, and hardware acceleration resources are given priority in high-load scenarios such as equipment failure simulation.

[0113] A consistency verification mechanism for cross-platform simulation environments is constructed. For different system compilation environments such as Elektronix, Android, and CentOS, a virtual machine image packaging tool is used to achieve integrated encapsulation of the compilation toolchain, dependency libraries, and component runtime, ensuring instruction-level compatibility of simulation applications on different platforms.

[0114] Design an interface adaptation layer for the hardware simulation test environment and real-time data components. Based on the hardware development checklist (schematic / PCB design specifications) of the unified R&D platform, realize the mapping and verification of communication protocols and electrical parameters between the simulation model and IoT terminals such as power distribution smart gateways and video acquisition equipment.

[0115] In this embodiment, the adaptive resource allocation algorithm based on load prediction dynamically allocates computing resources by accurately monitoring system operation indicators, ensuring simulation smoothness under high load scenarios and significantly improving resource utilization efficiency. The cross-platform consistency verification mechanism enables seamless migration of simulation applications across multiple systems, breaking down platform barriers. The interface adaptation layer opens up the connection channel between hardware and simulation components, ensuring accurate mapping between simulation models and actual IoT terminal parameters, greatly enhancing the consistency between simulation results and the physical world. In power system simulation, resources can be intelligently allocated according to real-time load to avoid system lag. The cross-platform feature supports rapid deployment of simulation applications across different devices and systems. The hardware interface adaptation layer facilitates efficient docking between simulation models and actual equipment, effectively reducing development costs and improving the feasibility and reliability of simulation technology applications.

[0116] To better demonstrate the implementation of a multimodal simulation method based on real-time data components, this invention provides a multimodal simulation system based on real-time data components, comprising:

[0117] The real-time data acquisition module is configured to acquire real-time data from various data sources in the power system, including electrical quantities such as voltage, current, and power, as well as non-electrical quantities such as equipment status and environmental parameters.

[0118] The data processing and analysis module is configured to perform standardized processing, semantic fusion, and trend prediction on the collected real-time data, generating standardized data entities that conform to the component interface specifications.

[0119] The multimodal simulation module is configured to construct simulation scenarios based on standardized data entities and combined with power system models to perform multimodal simulations of dynamic operating conditions such as equipment failure and system load, thereby simulating the dynamic behavior of the power system.

[0120] The visualization module is configured to integrate time-series, state-based, and spatial component libraries. Through data binding and configuration management, it drives standardized data entities to visualization components, and displays simulation results in an intuitive and easy-to-understand way, facilitating user analysis and decision-making.

[0121] In this embodiment, the multimodal simulation module includes a multimodal application designer, which comprises a design panel management unit, a real-time data component library unit, and a data binding configuration unit.

[0122] The component linkage engine includes an event-driven mechanism unit, a linkage rule setting unit, and an adaptive adjustment unit.

[0123] The cross-platform deployment engine includes a multi-system environment adaptation unit, an adaptive resource allocation unit, and a hardware simulation interface unit.

[0124] In this embodiment, a modular design achieves a closed-loop simulation process. The real-time data acquisition module covers both electrical and non-electrical quantities, ensuring data integrity. The standardization and semantic fusion of the data processing module enhance data usability, while trend prediction provides forward-looking support for the simulation. The multimodal simulation module combines a power system model to construct dynamic scenarios, accurately simulating actual operating conditions. The visualization module uses a multi-type component library to intuitively present simulation results, significantly improving decision-making efficiency. The multimodal application designer supports flexible configuration, the component linkage engine enables intelligent collaboration, and the cross-platform deployment engine ensures multi-environment adaptability. Overall, this significantly improves simulation accuracy and engineering practicality, effectively reduces testing costs, and enhances the engineering feasibility and business adaptability of simulation technology.

[0125] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A multimodal simulation method based on real-time data components, characterized in that, include: Build a multimodal application design environment: Build a multimodal simulation application framework, and manage the design panel, real-time data component library, and data binding and configuration based on the multimodal application designer; Component linkage and collaboration: Based on data binding and configuration management, an event-driven linkage mechanism is built between components, allowing for the editing of component properties and the setting of linkage rules between components to enable multi-component collaborative work; Simulation runtime optimization and deployment: Based on the inter-component linkage rule system, a cross-platform simulation runtime environment is built. According to the simulation objectives, different simulation scenarios are set to carry out lightweight deployment and continuous optimization of applications.

2. The multimodal simulation method based on real-time data components as described in claim 1, characterized in that, Building a multimodal application design environment specifically includes: Design panel management: Based on the constructed multimodal simulation application framework, the page structure is decomposed into a basic layout layer, a component display layer, and an interaction logic layer, and the panel is deployed in combination with the components of each component; Real-time data component library management: Establish a real-time data component library, which stores various preset components, determine the basic attributes of each preset component, and divide each preset component into component families based on the functions corresponding to the basic attributes. Each component family is set with corresponding category tags and keywords, and the preset components are classified and managed based on the component family category. Data binding and configuration management: Determine the mapping relationship between the data source of each real-time data and the basic attributes of each preset component. Based on the mapping relationship, perform multi-source data fusion and mapping of real-time data from each data source to dynamically display the real-time data.

3. The multimodal simulation method based on real-time data components as described in claim 2, characterized in that, Data binding and configuration management also includes establishing a real-time acquisition and fusion system for multi-source heterogeneous data: Real-time data is collected from various data sources, and the sampling interval of each data source is automatically adjusted according to the data change rate. The collected real-time data is cleaned, transformed, and standardized to remove invalid data, correct outliers, and integrate to generate multi-source heterogeneous data. Multi-source heterogeneous data is input into the entity semantic model to establish a semantic mapping of cross-source data. Multi-source data of the same entity is weighted and fused to generate a standardized data entity and determine the mapping relationship of fields from different data sources. Based on the mapping relationship of different data source fields, determine the mapping rules between data sources and component properties, and based on the mapping rules between data sources and component properties, determine the mapping relationship between the data source of each real-time data and the basic properties of each preset component.

4. The multimodal simulation method based on real-time data components as described in claim 3, characterized in that, Redundancy is assessed in the collected real-time data, and a real-time data acquisition anomaly alarm is triggered when redundancy is abnormal, including: Retrieve real-time data collected from the data source; Determine the standard deviation of the data volume collected in all windows based on the amount of data collected in each sliding window of the data source. Retrieve the number of data collections for each sliding window; The data deviation index is determined by using the standard deviation of the data volume corresponding to all window data collections corresponding to the data source and the number of data collections corresponding to each sliding window. The data deviation index is compared with a preset data deviation index threshold. When the data deviation index exceeds the preset data deviation index threshold, redundancy is determined, and a real-time data acquisition anomaly alarm is triggered when redundancy is abnormal.

5. The multimodal simulation method based on real-time data components as described in claim 4, characterized in that, When the data deviation index exceeds a preset data deviation index threshold, a redundancy determination is performed, including: When the data deviation index exceeds the preset data deviation index threshold, the data deviation index is retrieved. Retrieve the resource and energy consumption consumed by data acquisition for each sliding window of the current data source; The standard deviation of resource energy consumption is obtained by using the resource energy consumption of each sliding window corresponding to the current data acquisition from the data source. The standard deviation of resource energy consumption corresponding to the data source is normalized to obtain the normalized standard deviation of resource energy consumption. The resource energy consumption deviation is obtained by combining the normalized standard deviation of the resource energy consumption corresponding to the data source with the standard deviation of the collected data of m completed sliding windows, with the total amount of collected data corresponding to each sliding window as the benchmark. A redundancy quantification function is obtained by utilizing the data deviation index and resource energy consumption deviation of the data source; The value corresponding to the redundancy quantification function is compared with a preset function threshold; When the value corresponding to the redundancy quantification function exceeds the preset function threshold, the redundancy level of the real-time data is determined to be abnormal.

6. The multimodal simulation method based on real-time data components as described in claim 3, characterized in that, The multimodal application designer also includes: Component attribute editing: Set the component visual attribute configuration interface, parametrically edit the basic visual attributes of the real-time data component, and set the interactive attributes of the real-time data component. The size of the visual attribute configuration interface is automatically adapted under the grid layout constraints, and the display parameters are recalculated when the component position is adjusted. Component composition and decomposition: Combine multiple time-sequence, state-sequence, or spatial components into a composite component, generate a component hierarchy tree structure, and establish data interaction channels and event transmission links between sub-components during the composition process. At the same time, recursively decompose the composite component, obtain the attribute configuration and interaction logic mapping relationship before decomposition, and edit each component individually.

7. The multimodal simulation method based on real-time data components as described in claim 6, characterized in that, Component linkage and collaboration, specifically including: Establish a standardized event description system. Based on the standard event metadata structure, semantically bind the rule triggering conditions with the type definitions in the event metadata. When a component triggers an event, the event is automatically distributed to the corresponding component. Construct a rule knowledge graph, formalize the linkage rules into a graph structure, and predict the state change trend of components based on historical interaction data and linkage rules; The system matches the corresponding linkage rules and strategies based on the prediction results, adjusts the weights of the linkage rules based on historical execution data, generates the optimal linkage rules based on the runtime environment, and evaluates the execution effect of the optimal linkage rules in real time and generates optimization suggestions.

8. The multimodal simulation method based on real-time data components as described in claim 7, characterized in that, Simulation operation optimization deployment also includes: An adaptive resource allocation algorithm based on load prediction is adopted to dynamically adjust the computing resource allocation strategy by monitoring simulation scenario indicators. Construct a consistency verification mechanism for a cross-platform simulation environment to achieve integrated encapsulation of the compilation toolchain, dependency libraries, and component runtime; Design an interface adaptation layer for the hardware simulation test environment and real-time data components to realize the mapping and verification of communication protocols and electrical parameters between the simulation model and the IoT terminal.

9. A multimodal simulation system based on real-time data components, employing the multimodal simulation method based on real-time data components as described in claim 1, characterized in that, include: The real-time data acquisition module is configured to acquire real-time data from various data sources in the power system, including electrical and non-electrical quantities. The data processing and analysis module is configured to perform standardized processing, semantic fusion, and trend prediction on the collected real-time data, generating standardized data entities that conform to the component interface specifications. The multimodal simulation module is configured to construct simulation scenarios based on standardized data entities and combined with power system models to perform multimodal simulations of dynamic operating conditions and simulate the dynamic behavior of the power system. The visualization module is configured to integrate time-series, state-based, and spatial component libraries. Through data binding and configuration management, it drives standardized data entities to visualization components and displays simulation results in a visual format.

10. The multimodal simulation method based on real-time data components as described in claim 9, characterized in that, The multimodal simulation module includes a multimodal application designer, a component linkage engine, and a cross-platform deployment engine.

Images