Smart park visual management method based on digital twinning

Abstract

The invention relates to the technical field of data management, in particular to a digital twinning-based smart park visual management method, which comprises the following steps of S1, acquiring real-time data through a multi-source sensor, formulating a uniform data format specification and an interface protocol, performing data access conversion, and outputting uniform data; s2, an integrated three-dimensional digital twinborn model is established according to the unified data, and a simulation engine and an event linkage rule engine are introduced for real-time deduction. During use, by establishing the park integrated three-dimensional digital twinborn model, when an emergency situation such as a fire alarm occurs, real-time deduction can be carried out; the method comprises the following steps: automatically deducing simulation animations of entrance guard opening and personnel evacuation in a digital twin interface, synchronously turning off ventilation systems of related floors at the same time to form a linkage reaction, further predicting the fire spreading direction and time through a simulation engine, and intuitively presenting a future change trend through a visual means, so as to improve the fire spreading effect. And the cognition and response speed of a manager to potential risks and resource demand changes can be improved.

Description

Technical Field

[0001] The present invention relates to the field of data management technology, and in particular to a visual management method for a smart park based on digital twins. Background Art

[0002] Digital twin-based smart park visual management is an innovative management model that leverages advanced technologies such as the Internet of Things (IoT), big data, cloud computing, and artificial intelligence (AI) to create a virtual representation of the physical park—a digital twin model. This model allows managers to monitor, analyze, and optimize park facilities and services in real time within a virtual environment, enabling more efficient and intelligent management.

[0003] The patent publication number is CN202411109426.0, which states in its specification that “the present invention relates to the field of visual management technology, specifically to a visual management system and method for a smart park based on digital twin technology, the management method comprising the following steps: determining the monitoring range of each monitor in the smart park, dividing the smart park according to the distribution of the monitoring area image; performing periodic rotation monitoring on the monitors of the adjacent areas where there are missing areas, and determining the monitoring range of the monitors in each missing area; calculating the monitoring time allocated to the effective area; adjusting the monitoring time allocated to the effective area and the unit monitoring cycle of the monitor according to the difference between the average flow of people; judging whether there are overlapping areas, and analyzing the monitoring order of each monitor on the overlapping areas; obtaining the real-time monitoring coverage of the smart park every unit detection cycle. Coverage rate, analyze the range of changes in the expected monitoring coverage of the smart park, and determine whether to issue abnormal reminders for the monitoring of the smart park." Although the above technology combines digital twins with dynamic monitoring strategies and optimizes the spatiotemporal allocation of monitoring resources through mathematical models to achieve the advantages of low-cost, high-coverage smart park security management, some current smart park systems can only achieve static visualization of basic data on buildings, equipment, and personnel, and the update frequency is low. There is a lack of dynamic change simulation and future status prediction based on real-time IoT data. Managers often only see the current situation and cannot make decisions based on trends or prevent potential problems. At the same time, most smart park management systems are isolated subsystems coexisting. For example, the security system, energy consumption management system, parking system, and environmental monitoring system are independent and display their own data respectively, without unified scheduling and integrated presentation.

[0004] To sum up, developing a visual management method for smart parks based on digital twins is still a key issue that needs to be urgently addressed in the field of data management technology. Summary of the Invention

[0005] The purpose of this invention is to solve the problem in the prior art that although the above-mentioned technology combines digital twins with dynamic monitoring strategies, optimizes the spatiotemporal allocation of monitoring resources through mathematical models, and achieves the advantages of low-cost, high-coverage smart park security management, some current smart park systems can only achieve static display and visualization of basic data of buildings, equipment, and personnel, and the update frequency is low. There is a lack of dynamic change simulation and future state prediction based on real-time Internet of Things data. Managers can often only see the current situation and cannot make decisions based on trends or prevent potential problems. At the same time, most smart park management systems are isolated subsystems coexisting, such as security systems, energy consumption management, parking systems, and environmental monitoring systems. Each system operates independently and displays its own data, without unified scheduling and integrated presentation.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention provides a visual management method for a smart park based on digital twins, comprising the following steps: S1, collecting real-time data through multi-source sensors, formulating a unified data format specification and interface protocol, performing data access conversion, and outputting unified data;

[0008] S2. Build an integrated 3D digital twin model based on unified data, introduce a simulation engine and event linkage rule engine for real-time deduction, and output evolution data;

[0009] S3. Modeling is performed based on the time series prediction model according to the evolution data, and then the prediction results are output and visualized in the integrated digital twin interface;

[0010] S4. By setting an abnormal threshold, performing comparative analysis based on the prediction results, and automatically pushing early warning data;

[0011] S5. Based on the warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated to recommend a linked response plan.

[0012] Furthermore, in step S1, real-time data is collected through multi-source sensors, and a unified data format specification and interface protocol are formulated to perform data access conversion and output unified data as follows:

[0013] For different physical objects such as buildings, roads, equipment, and environmental factors in the park, multiple types of sensors are deployed to collect raw data in real time. The data source set collected by each sensor is set as: Q = {q i (t,p,a)|i=1,2,...,N}, where q irepresents the observation value of the i-th sensor at time t, spatial position p, and attribute dimension a. Q is a collection of real-time data. The unified data format specification and interface protocol are formulated, and the standard data format is defined as a triple: Where W is the standard data format, ID is the unique identifier of the device / sensor, Ttp is the standardized timestamp, E∈R k It is the collected attribute value vector, k is the number of standard attributes, and the data access conversion completes the conversion of original data to standard format through mapping function, satisfies data consistency constraints, formulates unified data format specifications, and then defines the interface protocol set. The specific protocol adaptation and message encapsulation are realized through the transformation function family. The standard intermediate data stream is output through heterogeneous protocol fusion mapping and normalization to obtain a standard data set. Based on the standard data set, a park data bus is constructed to output unified data Y(t).

[0014] Furthermore, in step S2, an integrated three-dimensional digital twin model is established based on the unified data, and a simulation engine and an event linkage rule engine are introduced for real-time deduction. The method for outputting the evolution data is as follows:

[0015] According to the unified data Y(t), all physical objects in the park are represented by the entity modeling set U, which is expressed as follows: U = {u i |i=1,2,...,M}, where u i Each entity contains static geometric features and dynamic state attributes, defining u i The structure is: i =(I i (x,y,z),o i (t)), where I i (x, y, z) is the spatial geometric description function, o i (t) is the dynamic state vector, and the integrated three-dimensional digital twin model is: Where P(t) is the set at time t, Indicates that the union is performed for each subsystem i from 1 to M, I i (x, y, z) is the spatial geometric description function, o i (t) is a dynamic state vector, and a high-speed state synchronization interface set is constructed, and the synchronization operation is defined as a state flow mapping, and the real-time constraints are met.

[0016] Furthermore, in step S2, an integrated three-dimensional digital twin model is established based on the unified data, and a simulation engine and an event linkage rule engine are introduced for real-time deduction. The method for outputting the evolution data is as follows:

[0017] The simulation engine is based on differential modeling of physical state changes and defines various change deduction processes, including device state deduction, personnel movement deduction, and environmental change deduction. The event linkage rule engine sets the cross-system event linkage rule set as A, and each cross-system event linkage rule is defined as: Among them S k is the trigger condition set, D k It's an action set. represents the state of the i-th event at time t, α j (t) represents the state of the j-th system at time t, β m (t) represents the value of the mth parameter at time t, ∧ represents a logical AND operation, → represents logical implication, Trigger represents a response action, and the event linkage engine outputs the evolution data ε(t) in real time.

[0018] Furthermore, in step S3, a method for modeling based on a time series prediction model according to the evolution data, outputting prediction results, and visually presenting them in an integrated digital twin interface is as follows:

[0019] According to the evolution data ε(t), the state sequences of different subsystems are extracted and standardized, and the subsystem set is defined as G = {g i |g1,g2,...,g N}, each subsystem g i Associate a set of original indicator time series, use wavelet transform to remove high-frequency noise, smooth and denoise the original indicator time series, and model the model based on the time series prediction model. Use a multi-layer stacked time series depth prediction network to define the prediction model in a generalized autoregressive form. Among them F i (·) is the prediction model, φ i is the feature encoder, is a time series aggregator, γ i is the decoder, expressed as:

[0020] in is the predicted state of the ith subsystem at time t+χ, F i (·) is the prediction model, H i (t) is the state of the ith subsystem at the current time t, H i (t-1) is the state of the i-th subsystem at the current time t-1, H i (t-L+1) is the state of the i-th subsystem at the current time t-L+1. The residual learning in the prediction process is: in is the predicted state of the ith subsystem at time t+χ, H i (t) is the state of the i-th subsystem at the current time t, Δ i (H i (t)) represents the state H of the i-th subsystem at the current time t i The change in (t).

[0021] Furthermore, in step S3, a method for modeling based on a time series prediction model according to the evolution data, outputting prediction results, and visually presenting them in an integrated digital twin interface is as follows:

[0022] All subsystems g i The prediction results Perform the collection to obtain the prediction matrix, which is expressed as: in represents the predicted state set of all subsystems within a period of time x in the future at time point t+χ, i=1,...,N is the index range indicating that there are N subsystems in total. It represents the predicted state of the i-th subsystem at time t+χ, and further defines the multi-index joint prediction scenario, which is expressed as follows: Where K is the joint risk assessment and trend extraction function, Z(χ) represents the input The calculated prediction output is, Represents the predicted state set of all subsystems in the future period of time x at the time point t+χ. The visualization presents the predicted results Mapped to the 3D interface through the digital twin rendering engine, the equipment load and energy consumption curves are dynamically drawn in the form of smooth broken lines.

[0023] Furthermore, in step S4, by setting an abnormality threshold and performing comparative analysis based on the prediction results, the method for automatically pushing warning data is as follows:

[0024] The abnormal threshold is set according to each subsystem g i Set static thresholds for different monitoring indicators and the dynamic threshold correction term η i′,j′ (t), constitutes the final application threshold, expressed as: where η i′,j′ (t) is the total state of the i′th subsystem to the j′th subsystem at time t, i′ represents the subsystem number, j′ represents the corresponding indicator number, is the static threshold, η i′,j′(t) is the dynamic threshold correction term, which is dynamically adjusted according to the real-time environmental status (seasonal changes, special events), and is expressed as: Δη i′,j′ (t) = ι i′,j′ ·sin(λ i′,j′ t+μ i′,j′ ), where ι, λ, μ represent the adjustment amplitude, change frequency, and phase offset of the i′th subsystem to the j′th subsystem, respectively, and Δη i′,j′ (t) is the total state deviation change of the i′th subsystem to the j′th subsystem at time t, and sin(·) is the sine function.

[0025] Furthermore, in step S4, by setting an abnormality threshold and performing comparative analysis based on the prediction results, the method for automatically pushing warning data is as follows:

[0026] The comparative analysis is performed based on the prediction results. For each prediction time step χ, the difference between the predicted value and the corresponding threshold is calculated and the abnormality degree function is defined. At the same time, the abnormal cumulative integral is introduced to take into account the trend abnormality. The expression formula is: where Δ i′,j′ (χ) is the state deviation caused by the i′th subsystem to the j′th subsystem at the prediction time step χ, is the predicted state of subsystem j′ generated by subsystem i′ at time t+χ, η i′,j′ (t+χ) is the total state of the i′th subsystem with respect to the j′th subsystem at time t+χ, is the linkage effect value of the i′th subsystem on the j′th subsystem within the prediction time step χ, max(·) is the maximum value function, θ i′,j′ is the sensitivity coefficient of the influence of the i′th subsystem on the j′th subsystem, C i′,j′ (χ) represents the cumulative linkage influence between the i′th subsystem and the j′th system within the prediction time step χ, represents the integration of the time interval from t to t+χ, represents the instantaneous linkage effect function imposed by the i′th subsystem on the j′th subsystem within the prediction time step χ, is the integral variable, ds is the small time interval, χ is the prediction time step, the automatic push warning data generates the warning data structure V i′ (t), triggered by the event bus, and pushed to the front-end of the integrated digital twin interface for display.

[0027] Furthermore, in step S5, based on the early warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated. The method for recommending a linked response solution is as follows:

[0028] According to the warning data Prediction result V i′ (t), define the subsystem state vector as: Among them B h (t) is the overall state vector of the h-th subsystem at time t, is a triple, is the real-time physical monitoring data of the hth subsystem at time t, is the predicted future trend vector of the hth subsystem at time t, is the warning level and abnormal index of the hth subsystem at time t, the dynamic update scenario defines the twin scenario state field The state of each twin space position point (x, y, z) is updated with time t, expressed as: in Indicates the scene status The partial derivative with respect to time t is, Represents the overall scene state at the three-dimensional spatial position (x, y, z) and time t, represents the driving force of overall scene change calculated based on the current status of all subsystems, is the state of (energy consumption system, security system, traffic system) at time t, represents the local disturbance term occurring at spatial position (x, y, z) and time t, is the scene evolution state at the k+1th moment, is the scene evolution state at the kth moment, Δt is the state from the current time t k Go to the next step k+1 The time interval between Indicates that based on the current subsystem status The calculated scene change rate or amount, is the state of the subsystem at time t, and the recommended linkage response plan defines the response plan space as the set Each response plan It is a set of executable actions, establishing a matching scoring function, expressing the formula: in Indicates the response plan In the current subsystem state The comprehensive rating of The outer summation symbol represents the sum from 1 to n for the i-th subsystem. The inner summation symbol represents the sum of the jth attribute in the i-th subsystem from 1 to li , σ i,j is the weight coefficient indicating the importance of the j-th attribute of the i-th subsystem in the overall score, Indicates the response plan With the current subsystem status The matching degree between them, τ is a regulating factor indicating the weight of the urgency score, Indicates the alarm event corresponding to the i-th subsystem at the current time t

[0029] The urgency score is calculated by maximizing the matching score and recommending a response plan. The expression formula is: in Indicates the final selected response plan, argmax indicates the parameter that maximizes it, Represents the set of all optional response options One of the solutions Indicates the response plan In the current subsystem state The comprehensive score under the above conditions will be pushed to the front end of the integrated digital twin interface for display.

[0030] Furthermore, in step S5, based on the early warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated. The method for recommending a linked response solution is as follows:

[0031] In the selected response plan Then, execute the response plan in sequence The action chain and synchronously update the scene state to the response effect, expressed as:

[0032] in Indicates to execute the recommended solution The actions contained in it are Action1, Action2, Action d′ Indicates that each element in the collection is a specific action, represents the final selected response plan, The state of the scene after executing the response action is a four-dimensional function. It represents the state of the twin scene before executing the response action, and ψ(·) is the transformation mapping of the action to the scene. Finally, the changes of each subsystem (evacuation guidance, re-planning of pedestrian flow lines, energy load adjustment animation) are dynamically displayed on the front end of the integrated digital twin interface, and the current linkage execution status is marked through a pop-up window.

[0033] Beneficial effects

[0034] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0035] When in use, the present invention establishes an integrated three-dimensional digital twin model of the campus. In the event of an emergency such as a fire, it automatically performs simulation animations of access control opening and personnel evacuation in the digital twin interface, and simultaneously shuts down the ventilation systems of relevant floors, forming a chain reaction. Furthermore, the direction and time of fire spread are predicted through the simulation engine, and future changes are intuitively presented through visualization means, which is conducive to improving managers' awareness and response speed to potential risks and changes in resource requirements.

[0036] When the present invention is in use, the method will comprehensively analyze the current status of all subsystems, evaluate the adaptability of each response plan to the current scenario status, automatically select the optimal response plan for execution, and synchronously update the executed status to the three-dimensional digital twin interface in real time, dynamically and visually display the change process of each subsystem, which is conducive to comprehensive perception of the park system status, intuitively presenting the evolution process of multi-system linkage, improving the overall perception and control accuracy, and avoiding the problem of isolated subsystems coexisting and acting independently in the park. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a flow chart of a digital twin-based smart park visualization management method of the present invention. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0039] It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the numbers used in this manner are interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or apparatuses.

[0040] The present invention is described in further detail below with reference to the accompanying drawings:

[0041] Example:

[0042] like Figure 1 As shown, the present invention provides a smart park visualization management method based on digital twins, comprising the following steps: S1, collecting real-time data through multi-source sensors, formulating a unified data format specification and interface protocol, performing data access conversion, and outputting unified data;

[0043] Furthermore, in step S1, real-time data is collected through multi-source sensors, and a unified data format specification and interface protocol are formulated to perform data access conversion and output unified data as follows:

[0044] For different physical objects such as buildings, roads, equipment, and environmental factors in the park, multiple types of sensors are deployed to collect raw data in real time. The data source set collected by each sensor is set as: Q = {q i (t,p,a)|i=1,2,...,N}, where q i represents the observation value of the i-th sensor at time t, spatial position p, and attribute dimension a. Q is a collection of real-time data. The unified data format specification and interface protocol are formulated, and the standard data format is defined as a triple: Where W is the standard data format, ID is the unique identifier of the device / sensor, Ttp is the standardized timestamp, E∈R k It is the collected attribute value vector, k is the number of standard attributes, and the data access conversion completes the conversion of original data to standard format through mapping function, satisfies data consistency constraints, formulates unified data format specifications, and then defines the interface protocol set. The specific protocol adaptation and message encapsulation are realized through the transformation function family. The standard intermediate data stream is output through heterogeneous protocol fusion mapping and normalization to obtain a standard data set. Based on the standard data set, a park data bus is constructed to output unified data Y(t).

[0045] In this embodiment, this method collects data from various physical objects such as buildings, roads, equipment, environmental factors, etc. in the park in real time through multi-source sensors (such as temperature and humidity sensors, access control card readers, road traffic detectors, energy metering meters, etc.), formulates a set of interface protocols, and designs mapping functions and transformation function families to achieve automatic conversion of raw data to standard formats and protocol adaptation, integrate heterogeneous data, and finally output a unified standardized intermediate data stream to form a unified data set, which is conducive to improving data interoperability and compatibility between various subsystems in the park, reducing the technical complexity of later system expansion and new equipment access, and at the same time providing a high-quality, standardized data foundation for subsequent data analysis, intelligent decision-making, and event linkage.

[0046] S2. Build an integrated 3D digital twin model based on unified data, introduce a simulation engine and event linkage rule engine for real-time deduction, and output evolution data;

[0047] Furthermore, in step S2, an integrated three-dimensional digital twin model is established based on the unified data, and a simulation engine and an event linkage rule engine are introduced for real-time deduction. The method for outputting the evolution data is as follows:

[0048] According to the unified data Y(t), all physical objects in the park are represented by the entity modeling set U, which is expressed as follows: U = {u i |i=1,2,...,M}, where u i Each entity contains static geometric features and dynamic state attributes, defining u i The structure is: i =(I i (x,y,z),o i (t)), where I i (x, y, z) is the spatial geometric description function, o i (t) is the dynamic state vector, and the integrated three-dimensional digital twin model is: Where P(t) is the set at time t, Indicates that the union is performed for each subsystem i from 1 to M, I i (x, y, z) is the spatial geometric description function, o i (t) is a dynamic state vector, and a high-speed state synchronization interface set is constructed, and the synchronization operation is defined as a state flow mapping, and the real-time constraints are met.

[0049] Furthermore, in step S2, an integrated three-dimensional digital twin model is established based on the unified data, and a simulation engine and an event linkage rule engine are introduced for real-time deduction. The method for outputting the evolution data is as follows:

[0050] The simulation engine is based on differential modeling of physical state changes and defines various change deduction processes, including device state deduction, personnel movement deduction, and environmental change deduction. The event linkage rule engine sets the cross-system event linkage rule set as A, and each cross-system event linkage rule is defined as: Among them S k is the trigger condition set, D k It's an action set. represents the state of the i-th event at time t, α j (t) represents the state of the j-th system at time t, β m(t) represents the value of the mth parameter at time t, ∧ represents a logical AND operation, → represents logical implication, Trigger represents a response action, and the event linkage engine outputs the evolution data ε(t) in real time.

[0051] In this embodiment, based on standardized, unified data, an integrated three-dimensional digital twin model of the campus can be established. This model is synchronized with real-time data in the physical world with a latency of less than one second via high-speed synchronization interfaces (MQTT, OPC UA, and WebSocket). For example, in a large office building, when a fire alarm occurs, not only is the floor immediately displayed, but the digital twin interface also automatically displays simulated animations of access control opening and personnel evacuation, while simultaneously shutting down the ventilation systems on the relevant floors, creating a chain reaction. Furthermore, the simulation engine can predict the direction and time of fire spread, providing intuitive and dynamic support for emergency decision-making. This facilitates real-time state synchronization between the physical campus and the digital model, greatly enhancing the visualization and controllability of the campus situation.

[0052] S3. Modeling is performed based on the time series prediction model according to the evolution data, and then the prediction results are output and visualized in the integrated digital twin interface;

[0053] Furthermore, in step S3, a method for modeling based on a time series prediction model according to the evolution data, outputting prediction results, and visually presenting them in an integrated digital twin interface is as follows:

[0054] According to the evolution data ε(t), the state sequences of different subsystems are extracted and standardized, and the subsystem set is defined as G = {g i |g1,g2,...,g N}, each subsystem g i Associate a set of original indicator time series, use wavelet transform to remove high-frequency noise, smooth and denoise the original indicator time series, and model the model based on the time series prediction model. Use a multi-layer stacked time series depth prediction network to define the prediction model in a generalized autoregressive form. Among them F i (·) is the prediction model, φ i is the feature encoder, is a time series aggregator, γ i is the decoder, expressed as:

[0055] in is the predicted state of the ith subsystem at time t+χ, F i (·) is the prediction model, H i(t) is the state of the ith subsystem at the current time t, H i (t-1) is the state of the i-th subsystem at the current time t-1, H i (t-L+1) is the state of the i-th subsystem at the current time t-L+1. The residual learning in the prediction process is: in is the predicted state of the ith subsystem at time t+χ, H i (t) is the state of the i-th subsystem at the current time t, Δ i (H i (t)) represents the state H of the i-th subsystem at the current time t i The change in (t).

[0056] Furthermore, in step S3, a method for modeling based on a time series prediction model according to the evolution data, outputting prediction results, and visually presenting them in an integrated digital twin interface is as follows:

[0057] All subsystems g i The prediction results Perform the collection to obtain the prediction matrix, which is expressed as: in represents the predicted state set of all subsystems within a period of time x in the future at time point t+χ, i=1,...,N is the index range indicating that there are N subsystems in total. It represents the predicted state of the i-th subsystem at time t+χ, and further defines the multi-index joint prediction scenario, which is expressed as follows: Where K is the joint risk assessment and trend extraction function, Z(χ) represents the input The calculated prediction output is, Represents the predicted state set of all subsystems in the future period of time x at the time point t+χ. The visualization presents the predicted results Mapped to the 3D interface through the digital twin rendering engine, the equipment load and energy consumption curves are dynamically drawn in the form of smooth broken lines.

[0058] In this embodiment, accurate time series predictions are used to achieve early perception of the operating trends of key systems in the park (such as energy consumption, transportation, and security). For example, in a large commercial complex, through such a prediction model, managers can see on the digital twin interface that elevator usage will increase sharply and power load will approach the warning line during the evening shopping peak on a certain day in the future. The system automatically issues an early warning and prompts that temporary personnel may need to be dispatched to maintain order or optimize energy consumption scheduling strategies. The future change trends are intuitively presented through visualization, which helps to improve managers' cognition and response speed to potential risks and changes in resource demand.

[0059] S4. By setting an abnormal threshold, performing comparative analysis based on the prediction results, and automatically pushing early warning data;

[0060] Furthermore, in step S4, by setting an abnormality threshold and performing comparative analysis based on the prediction results, the method for automatically pushing early warning data is:

[0061] The abnormal threshold is set according to each subsystem g i Set static thresholds for different monitoring indicators and the dynamic threshold correction term η i′,j′ (t), constitutes the final application threshold, expressed as: where η i′,j′ (t) is the total state of the i′th subsystem to the j′th subsystem at time t, i′ represents the subsystem number, j′ represents the corresponding indicator number, is the static threshold, η i′,j′ (t) is the dynamic threshold correction term, which is dynamically adjusted according to the real-time environmental status (seasonal changes, special events), and is expressed as: Δη i′,j′ (t) = ι i′,j′ ·sin(λ i′,j′ t+μ i′,j′ ), where ι, λ, μ represent the adjustment amplitude, change frequency, and phase offset of the i′th subsystem to the j′th subsystem, respectively, and Δη i′,j′ (t) is the total state deviation change of the i′th subsystem to the j′th subsystem at time t, and sin(·) is the sine function.

[0062] Furthermore, in step S4, by setting an abnormality threshold and performing comparative analysis based on the prediction results, the method for automatically pushing early warning data is:

[0063] The comparative analysis is performed based on the prediction results. For each prediction time step χ, the difference between the predicted value and the corresponding threshold is calculated and the abnormality degree function is defined. At the same time, the abnormal cumulative integral is introduced to take into account the trend abnormality. The expression formula is: where Δ i′,j′ (χ) is the state deviation caused by the i′th subsystem to the j′th subsystem at the prediction time step χ, is the predicted state of subsystem j′ generated by subsystem i′ at time t+χ, η i′,j′ (t+χ) is the total state of the i′th subsystem with respect to the j′th subsystem at time t+χ, is the linkage effect value of the i′th subsystem on the j′th subsystem within the prediction time step χ, max(·) is the maximum value function, θ i′,j′ is the sensitivity coefficient of the influence of the i′th subsystem on the j′th subsystem, C i′,j′ (χ) represents the cumulative linkage influence between the i′th subsystem and the j′th system within the prediction time step χ, represents the integration of the time interval from t to t+χ, represents the instantaneous linkage effect function imposed by the i′th subsystem on the j′th subsystem within the prediction time step χ, is the integral variable, ds is the small time interval, χ is the prediction time step, the automatic push warning data generates the warning data structure V i′ (t), triggered by the event bus, and pushed to the front-end of the integrated digital twin interface for display.

[0064] In this embodiment, the use of static and dynamic dual thresholds is conducive to improving the accuracy and flexibility of anomaly detection and avoiding misjudgments due to seasonality or special events. For example, when the power system of a large shopping mall predicts continuous high temperatures in the next few days and the air-conditioning load continues to exceed the normal threshold, the system can push early warning notifications to management personnel in advance based on the dynamically corrected standards, so that they can prepare for capacity increase, zoned cooling or load transfer. After detecting an anomaly, an early warning data structure will be automatically generated, triggering the event bus, and pushing the early warning information to the visualization terminal of the digital twin interface, displaying the anomaly location, risk level and possible impact range in real time, supporting managers to quickly perceive and make decisions, and through the abnormal accumulation point method, not only sudden failures can be captured, but also progressive risks can be identified, thereby improving the long-term health protection capabilities of the system.

[0065] S5. Based on the warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated to recommend a linked response plan;

[0066] Furthermore, in step S5, based on the early warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated. The method for recommending a linked response solution is as follows:

[0067] According to the warning data Prediction result Vi′ (t), define the subsystem state vector as: Among them B h (t) is the overall state vector of the h-th subsystem at time t, is a triple, is the real-time physical monitoring data of the hth subsystem at time t, is the predicted future trend vector of the hth subsystem at time t, is the warning level and abnormal index of the hth subsystem at time t, the dynamic update scenario defines the twin scenario state field The state of each twin space position point (x, y, z) is updated with time t, expressed as: in Indicates the scene status The partial derivative with respect to time t is, Represents the overall scene state at the three-dimensional spatial position (x, y, z) and time t, represents the driving force of overall scene change calculated based on the current status of all subsystems, is the state of (energy consumption system, security system, traffic system) at time t, represents the local disturbance term occurring at spatial position (x, y, z) and time t, is the scene evolution state at the k+1th moment, is the scene evolution state at the kth moment, Δt is the state from the current time t k Go to the next step k+1 The time interval between Indicates that based on the current subsystem status The calculated scene change rate or amount, is the state of the subsystem at time t, and the recommended linkage response plan defines the response plan space as the set Each response plan It is a set of executable actions, establishing a matching scoring function, expressing the formula: in Indicates the response plan In the current subsystem state The comprehensive rating of The outer summation symbol represents the sum from 1 to n for the i-th subsystem. The inner summation symbol represents the sum of the jth attribute in the i-th subsystem from 1 to l i , σ i,jis the weight coefficient indicating the importance of the j-th attribute of the i-th subsystem in the overall score, Indicates the response plan With the current subsystem status The matching degree between them, τ is a regulating factor indicating the weight of the urgency score, Indicates the alarm event corresponding to the i-th subsystem at the current time t The urgency score is calculated by maximizing the matching score and recommending a response plan. The expression formula is: in Indicates the final selected response plan, argmax indicates the parameter that maximizes it, Represents the set of all optional response options One of the solutions Indicates the response plan In the current subsystem state The comprehensive score under the above conditions will be pushed to the front end of the integrated digital twin interface for display.

[0068] Furthermore, in step S5, based on the early warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated. The method for recommending a linked response solution is as follows:

[0069] In the selected response plan Then, execute the response plan in sequence The action chain and synchronously update the scene state to the response effect, expressed as:

[0070] in Indicates to execute the recommended solution The actions contained in it are Action1, Action2, Action d′ Indicates that each element in the collection is a specific action, represents the final selected response plan, The state of the scene after executing the response action is a four-dimensional function. It represents the state of the twin scene before executing the response action, and ψ(·) is the transformation mapping of the action to the scene. Finally, the changes of each subsystem (evacuation guidance, re-planning of pedestrian flow lines, energy load adjustment animation) are dynamically displayed on the front end of the integrated digital twin interface, and the current linkage execution status is marked through a pop-up window.

[0071] In this embodiment, this method displays the real-time status of all subsystems in a linked manner based on the early warning data and prediction results, dynamically updates the overall scenario, and recommends the optimal linkage response plan. This method defines a comprehensive state vector for each subsystem that includes real-time monitoring data, future prediction trends, and current warning levels. Based on this, it constructs the state field of the entire twin scenario. Through dynamic calculations of mathematical models, it updates the scene changes of each spatial location in real time, such as energy consumption density, pedestrian flow distribution, or changes in security risk points. This method comprehensively analyzes the current status of all subsystems, establishes a response plan space, and evaluates each response plan (such as The method automatically selects the optimal response plan by considering factors such as the degree of urgency and the weight of each indicator, and after selecting the response plan, it executes each operation in sequence according to the action chain, and synchronously updates the executed status to the three-dimensional digital twin interface in real time, dynamically and visually displaying the change process of each subsystem (such as evacuation path animation, crowd flow change, energy consumption load curve adjustment, etc.). At the same time, a prompt window pops up to mark the currently executed linkage operation steps, which is conducive to comprehensive perception of the park system status, intuitively presenting the evolution process of multi-system linkage, and improving the overall perception and control accuracy.

[0072] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A visual management method for a smart park based on digital twins, characterized in that: The following steps are involved: S1. Collect real-time data through multi-source sensors, formulate unified data format specifications and interface protocols, perform data access conversion, and output unified data; S2. Build an integrated 3D digital twin model based on unified data, introduce a simulation engine and event linkage rule engine for real-time deduction, and output evolution data; S3. Modeling is performed based on the time series prediction model according to the evolution data, and then the prediction results are output and visualized in the integrated digital twin interface; S4. By setting an abnormal threshold, performing comparative analysis based on the prediction results, and automatically pushing early warning data; S5. Based on the warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated to recommend a linked response plan.

2. A digital twin-based smart park visualization management method according to claim 1, characterized in that: In step S1, real-time data is collected through multi-source sensors, and a unified data format specification and interface protocol are formulated to perform data access conversion and output unified data as follows: For different physical objects such as buildings, roads, equipment, and environmental factors in the park, multiple types of sensors are deployed to collect raw data in real time. The data source set collected by each sensor is set as: Q = {q i (t,p,a)|i=1,2,...,N}, where q i represents the observation value of the i-th sensor at time t, spatial position p, and attribute dimension a. Q is a collection of real-time data. The unified data format specification and interface protocol are formulated, and the standard data format is defined as a triple: Where W is the standard data format, ID is the unique identifier of the device / sensor, Ttp is the standardized timestamp, E∈R k It is the collected attribute value vector, k is the number of standard attributes, and the data access conversion completes the conversion of original data to standard format through mapping function, satisfies data consistency constraints, formulates unified data format specifications, and then defines the interface protocol set. The specific protocol adaptation and message encapsulation are realized through the transformation function family. The standard intermediate data stream is output through heterogeneous protocol fusion mapping and normalization to obtain a standard data set. Based on the standard data set, a park data bus is constructed to output unified data Y(t).

3. A digital twin-based smart park visualization management method according to claim 2, characterized in that: In step S2, an integrated three-dimensional digital twin model is established based on the unified data, and a simulation engine and an event linkage rule engine are introduced for real-time deduction. The method for outputting the evolution data is as follows: According to the unified data Y(t), all physical objects in the park are represented by the entity modeling set U, which is expressed as follows: U = {u i |i=1,2,...,M}, where u i Each entity contains static geometric features and dynamic state attributes, defining u i The structure is: i =(I i (x,y,z),o i (t)), where I i (x, y, z) is the spatial geometric description function, o i (t) is the dynamic state vector, and the integrated three-dimensional digital twin model is: Where P(t) is the set at time t, Indicates that the union is performed for each subsystem i from 1 to M, I i (x, y, z) is the spatial geometric description function, o i (t) is a dynamic state vector, and a high-speed state synchronization interface set is constructed, and the synchronization operation is defined as a state flow mapping, and the real-time constraints are met.

4. A digital twin-based smart park visualization management method according to claim 3, characterized in that: In step S2, an integrated three-dimensional digital twin model is established based on the unified data, and a simulation engine and an event linkage rule engine are introduced for real-time deduction. The method for outputting the evolution data is as follows: The simulation engine is based on differential modeling of physical state changes and defines various change deduction processes, including device state deduction, personnel movement deduction, and environmental change deduction. The event linkage rule engine sets the cross-system event linkage rule set as A, and each cross-system event linkage rule is defined as: Among them S k is the trigger condition set, D k It's an action set. represents the state of the i-th event at time t, α j (t) represents the state of the j-th system at time t, β m (t) represents the value of the mth parameter at time t, ∧ represents a logical AND operation, → represents logical implication, Trigger represents a response action, and the event linkage engine outputs the evolution data ε(t) in real time.

5. A digital twin-based smart park visualization management method according to claim 4, characterized in that: In step S3, a method for modeling based on the time series prediction model according to the evolution data, outputting the prediction results, and visually presenting them in the integrated digital twin interface is as follows: According to the evolution data ε(t), the state sequences of different subsystems are extracted and standardized, and the subsystem set is defined as G = {g i |g1,g2,...,g N }, each subsystem g i Associate a set of original indicator time series, use wavelet transform to remove high-frequency noise, smooth and denoise the original indicator time series, and model the model based on the time series prediction model. Use a multi-layer stacked time series depth prediction network to define the prediction model in a generalized autoregressive form. Among them F i (·) is the prediction model, φ i is the feature encoder, is a time series aggregator, γ i is the decoder, expressed as: in is the predicted state of the ith subsystem at time t+χ, F i (·) is the prediction model, H i (t) is the state of the ith subsystem at the current time t, H i (t-1) is the state of the i-th subsystem at the current time t-1, H i (t-L+1) is the state of the i-th subsystem at the current time t-L+1. The residual learning in the prediction process is: in is the predicted state of the ith subsystem at time t+χ, H i (t) is the state of the i-th subsystem at the current time t, Δ i (H i (t)) represents the state H of the i-th subsystem at the current time t i The change in (t).

6. A digital twin-based smart park visualization management method according to claim 5, characterized in that: In step S3, a method for modeling based on the time series prediction model according to the evolution data, outputting the prediction results, and visually presenting them in the integrated digital twin interface is as follows: All subsystems g i The prediction results Perform the collection to obtain the prediction matrix, which is expressed as: in represents the predicted state set of all subsystems within a period of time x in the future at time point t+χ, i=1,...,N is the index range indicating that there are N subsystems in total. It represents the predicted state of the i-th subsystem at time t+χ, and further defines the multi-index joint prediction scenario, which is expressed as follows: Where K is the joint risk assessment and trend extraction function, Z(χ) represents the input The calculated prediction output is, Represents the predicted state set of all subsystems in the future period of time x at the time point t+χ. The visualization presents the predicted results Mapped to the 3D interface through the digital twin rendering engine, the equipment load and energy consumption curves are dynamically drawn in the form of smooth broken lines.

7. A digital twin-based smart park visualization management method according to claim 6, characterized in that: In step S4, by setting an abnormality threshold and performing comparative analysis based on the prediction results, the method for automatically pushing warning data is as follows: The abnormal threshold is set according to each subsystem g i Set static thresholds for different monitoring indicators and the dynamic threshold correction term η i′,j′ (t), constitutes the final application threshold, expressed as: where η i′,j′ (t) is the total state of the i′th subsystem to the j′th subsystem at time t, i′ represents the subsystem number, j′ represents the corresponding indicator number, is the static threshold, η i′,j′ (t) is the dynamic threshold correction term, which is dynamically adjusted according to the real-time environmental status (seasonal changes, special events), and is expressed as: Δη i′,j′ (t) = ι i′,j′ ·sin(λ i′,j′ t+μ i′,j′ ), where ι, λ, μ represent the adjustment amplitude, change frequency, and phase offset of the i′th subsystem to the j′th subsystem, respectively, and Δη i′,j′ (t) is the total state deviation change of the i′th subsystem to the j′th subsystem at time t, and sin(·) is the sine function.

8. The method for visual management of a smart park based on digital twins according to claim 7 is characterized in that: In step S4, by setting an abnormality threshold and performing comparative analysis based on the prediction results, the method for automatically pushing warning data is as follows: The comparative analysis is performed based on the prediction results. For each prediction time step χ, the difference between the predicted value and the corresponding threshold is calculated and the abnormality degree function is defined. At the same time, the abnormal cumulative integral is introduced to take into account the trend abnormality. The expression formula is: where Δ i′,j′ (χ) is the state deviation caused by the i′th subsystem to the j′th subsystem at the prediction time step χ, is the predicted state of subsystem j′ generated by subsystem i′ at time t+χ, η i′,j′ (t+χ) is the total state of the i′th subsystem with respect to the j′th subsystem at time t+χ, is the linkage effect value of the i′th subsystem on the j′th subsystem within the prediction time step χ, max(·) is the maximum value function, θ i′,j′ is the sensitivity coefficient of the influence of the i′th subsystem on the j′th subsystem, C i′,j′ (χ) represents the cumulative linkage influence between the i′th subsystem and the j′th system within the prediction time step χ, ∫ t t+χ represents the integration of the time interval from t to t+χ, represents the instantaneous linkage effect function imposed by the i′th subsystem on the j′th subsystem within the prediction time step χ, is the integral variable, ds is the small time interval, χ is the prediction time step, the automatic push warning data generates the warning data structure V i′ (t), triggered by the event bus, and pushed to the front-end of the integrated digital twin interface for display.

9. The method for visual management of a smart park based on digital twins according to claim 8 is characterized in that: In step S5, based on the early warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated. The method for recommending a linked response solution is as follows: According to the warning data Prediction result V i′ (t), define the subsystem state vector as: Among them B h (t) is the overall state vector of the h-th subsystem at time t, is a triple, is the real-time physical monitoring data of the hth subsystem at time t, is the predicted future trend vector of the hth subsystem at time t, It is the warning level and abnormal index of the h-th subsystem at time t. The dynamic update scenario defines the twin scenario state field θ(x, y, z, t), where the state of each twin space position point (x, y, z) is updated with time t, expressed as: in represents the partial derivative of the scene state θ with respect to time t, θ(x,y,z,t) represents the overall scene state at the three-dimensional spatial position (x,y,z) and time t, represents the driving force of overall scene change calculated based on the current status of all subsystems, is the state of (energy consumption system, security system, traffic system) at time t, represents the local disturbance term occurring at spatial position (x, y, z) and time t, θ k+1 is the scene evolution state at the k+1th moment, θ k is the scene evolution state at the kth moment, Δt is the state from the current time t k Go to the next step k+1 The time interval between Indicates that based on the current subsystem status The calculated scene change rate or amount, is the state of the subsystem at time t, and the recommended linkage response plan defines the response plan space as the set Each response plan It is a set of executable actions, establishing a matching scoring function, expressing the formula: in Indicates the response plan In the current subsystem state The comprehensive rating of The outer summation symbol represents the sum from 1 to n for the i-th subsystem. The inner summation symbol represents the sum of the jth attribute in the i-th subsystem from 1 to l i ,σ i,j is the weight coefficient indicating the importance of the j-th attribute of the i-th subsystem in the overall score, Indicates the response plan With the current subsystem status The matching degree between them, τ is a regulating factor indicating the weight of the urgency score, Indicates the alarm event corresponding to the i-th subsystem at the current time t The urgency score is calculated by maximizing the matching score and recommending a response plan. The expression formula is: in Indicates the final selected response plan, argmax indicates the parameter that maximizes it, Represents the set of all optional response options One of the solutions Indicates the response plan In the current subsystem state The comprehensive score under the above conditions will be pushed to the front end of the integrated digital twin interface for display.

10. The method for visual management of a smart park based on digital twins according to claim 8, characterized in that: In step S5, based on the early warning data and prediction results, the real-time status of all subsystems is displayed in a linked manner, and the scenario is dynamically updated. The method for recommending a linked response solution is as follows: In the selected response plan Then, execute the response plan in sequence The action chain and synchronously update the scene state to the response effect, expressed as: in Indicates to execute the recommended solution The actions contained in it are Action1, Action2, Action d′ Indicates that each element in the collection is a specific action, represents the final selected response plan, θ ar (x, y, z, t) represents the state of the scene after executing the response action, which is a four-dimensional function. θ(x, y, z, t) represents the state of the twin scene before executing the response action. ψ(·) is the transformation mapping of the action to the scene. Finally, the changes of each subsystem (evacuation guidance, re-planning of pedestrian flow lines, energy consumption load adjustment animation) are dynamically displayed on the front end of the integrated digital twin interface, and the current linkage execution status is marked through a pop-up window.

Images