City-level digital twin cloud management platform and method supporting multi-scenario access

Abstract

The application provides a city-level digital twin cloud management platform and method supporting multi-scene access, relates to the technical field of digital twin, and comprises the following: a multi-dimensional modeling module for constructing a city digital twin model; a risk analysis module for constructing a multi-scene risk atlas; an operation and maintenance decision module for obtaining a multi-scene operation and maintenance space; an evaluation optimization module for obtaining a multi-scene operation and maintenance population; a global optimization module for determining a multi-scene operation and maintenance optimization result; and a cloud management module for performing risk adaptive cloud management of the city according to the multi-scene operation and maintenance optimization result. The application solves the technical problems of difficulty in multi-scene cooperation in existing digital twin city management, one-sided and lagging risk analysis, and low management efficiency, and achieves the technical effects of improving city operation and maintenance efficiency and enhancing risk response capability by constructing a multi-scene risk atlas and combining an evolutionary optimization algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital twinning, in particular to a city-level digital twinning cloud management platform and method supporting multi-scene access. BACKGROUND

[0002] With the accelerated advancement of global urbanization, the scale of cities is expanding and the complexity of functions is rising. Modern cities have gradually evolved into complex giant systems deeply intertwined by infrastructure, transportation networks, energy systems, public safety and other multi-scene scenarios. Under this background, city management is facing unprecedented challenges: on the one hand, data from different scenarios is scattered in different departments and systems, forming "data islands", making it difficult to achieve cross-scene information fusion and correlation analysis; on the other hand, traditional city management methods focus on single-scene risk warning and operation and maintenance decision-making, lacking the ability to analyze the dynamic coupling relationship of multi-scene risks, resulting in risk response often being limited to local optimization, making it difficult to develop scientific and coordinated operation and maintenance strategies from a global perspective. In addition, the real-time perception and dynamic simulation of city operation status are increasingly urgent, but existing digital twinning technologies focus on single-scene modeling, lacking the ability to support multi-scene data unified access, multi-source risk correlation deduction and global operation and maintenance decision optimization, making it difficult to meet the fine and intelligent management needs of city-level complex systems. SUMMARY

[0003] The present application provides a city-level digital twinning cloud management platform and method supporting multi-scene access, aiming to solve the technical problems of multi-scene coordination difficulty in existing digital twinning city management, leading to one-sided and lagging risk analysis, and low management efficiency.

[0004] The first aspect of the present application provides a city-level digital twinning cloud management platform supporting multi-scene access, which comprises: a multi-dimensional modeling module for multi-dimensional modeling according to multi-scene characteristic data of a city to construct a city digital twinning model; a risk analysis module for multi-scene risk analysis according to the city digital twinning model to construct a multi-scene risk atlas; an operation and maintenance decision module for operation and maintenance decision-making of the city according to the multi-scene risk atlas to obtain a multi-scene operation and maintenance space; an evaluation and optimization module for evaluation and optimization of the multi-scene operation and maintenance space through a multi-scene guarantee analysis model to obtain a multi-scene operation and maintenance population; a global optimization module for global evolutionary optimization of the multi-scene operation and maintenance population according to a predetermined evolution number to determine a multi-scene operation and maintenance optimization result; and a cloud management module for executing risk adaptive cloud management of the city according to the multi-scene operation and maintenance optimization result.

[0005] In a second aspect, the application discloses a city-level digital twin cloud management method supporting multi-scene access, which is implemented by the city-level digital twin cloud management platform supporting multi-scene access. The method comprises the following steps: performing multi-dimensional modeling according to multi-scene characteristic data of a city to construct a city digital twin model; performing multi-scene risk analysis according to the city digital twin model to construct a multi-scene risk atlas; performing operation and maintenance decision-making on the city according to the multi-scene risk atlas to obtain a multi-scene operation and maintenance space; evaluating and optimizing the multi-scene operation and maintenance space through a multi-scene guarantee analysis model to obtain a multi-scene operation and maintenance population; performing global evolution optimization on the multi-scene operation and maintenance population according to a predetermined evolution number to determine a multi-scene operation and maintenance optimization result; and performing risk adaptive cloud management on the city according to the multi-scene operation and maintenance optimization result.

[0006] The one or more technical solutions provided in the application have at least the following technical effects or advantages:

[0007] The method comprises the following steps: performing multi-dimensional modeling according to multi-scene characteristic data of a city to construct a city digital twin model; performing multi-scene risk analysis according to the city digital twin model to construct a multi-scene risk atlas; performing operation and maintenance decision-making on the city according to the multi-scene risk atlas to obtain a multi-scene operation and maintenance space; evaluating and optimizing the multi-scene operation and maintenance space through a multi-scene guarantee analysis model to obtain a multi-scene operation and maintenance population; performing global evolution optimization on the multi-scene operation and maintenance population according to a predetermined evolution number to determine a multi-scene operation and maintenance optimization result; and performing risk adaptive cloud management on the city according to the multi-scene operation and maintenance optimization result.

[0008] The above description is only a summary of the technical solutions of the application. In order to make the technical solutions of the application more clearly understood, the following detailed description is made in accordance with the content of the description, and in order to make the above and other purposes, characteristics and advantages of the application more obvious and easy to understand, the following detailed description is made. BRIEF DESCRIPTION OF DRAWINGS

[0009] Figure 1 A city-level digital twin cloud management platform structure diagram supporting multi-scene access is provided for the embodiments of the application.

[0010] Figure 2 A city-level digital twin cloud management method flow diagram supporting multi-scene access is provided for the embodiments of the application.

[0011] Explanation of reference signs: multi-dimensional modeling module 11, risk analysis module 12, operation and maintenance decision module 13, evaluation and optimization module 14, global optimization module 15, cloud management module 16. DETAILED DESCRIPTION

[0012] The embodiment of the present application provides a city-level digital twin cloud management platform supporting multi-scene access and a method thereof, which solves the technical problems of difficulty in multi-scene cooperation in the existing digital twin city management, one-sided and lagging risk analysis, and low management efficiency.

[0013] After introducing the basic principle of the present application, the various non-limiting embodiments of the present application will be specifically introduced in combination with the drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0014] Embodiment one, as shown in the figure, the embodiment of the present application provides a city-level digital twin cloud management platform supporting multi-scene access, which comprises: Figure 1

[0015] The multi-dimensional modeling module 11 is used for multi-dimensional modeling according to the multi-scene feature data of the city, and constructing a city digital twin model.

[0016] Specifically, in the multi-dimensional modeling module 11, first, the multi-scene monitoring data set covering various key fields of city operation is obtained from data sources such as Internet of Things sensing devices, traffic monitoring systems, energy management systems, and city infrastructure operation and maintenance platforms, which includes multi-dimensional information such as infrastructure monitoring data, traffic monitoring data, and energy monitoring data. Then, the collected multi-scene monitoring data set is cleaned, including outlier removal, missing value completion, and duplicate value removal, to ensure the accuracy of the input data. After that, the multi-scene feature data obtained after cleaning is input into a three-dimensional simulation software and injected into a city model framework preset by digital twin technology. The city model framework is usually designed by integrating building information model (BIM), geographic information system (GIS), and real-time three-dimensional engine technology, wherein the BIM model accurately depicts the internal structure and attributes of important single facilities, the GIS data provides macro geographic spatial relationship and topography, and the three-dimensional engine is responsible for integration and rendering to form a visual city appearance. The finally formed city digital twin model can not only intuitively reflect the physical entity structure of the city, but also dynamically simulate the running state and evolution trend under different scenes, thereby providing a unified digital support environment for subsequent risk analysis, operation and maintenance decision, and optimization analysis.

[0017] Further, the multi-dimensional modeling module 11 comprises:

[0018] ​obtain a multi-scene monitoring data set of the city, the multi-scene monitoring data set comprising infrastructure monitoring data, traffic monitoring data, and energy monitoring data; perform data cleaning on the multi-scene monitoring data set to generate the multi-scene feature data.

[0019] Preferably, in the multi-dimensional modeling module 11, the obtained multi-scene monitoring data set comprises infrastructure monitoring data, traffic monitoring data, and energy monitoring data of the city, wherein the infrastructure monitoring data comprises building structure, water supply pipe network pressure, drainage pipe flow, power line operation state, and communication facility operation condition, etc.; the traffic monitoring data comprises road traffic flow, vehicle speed, road congestion condition, traffic signal state, and public transportation operation condition, etc.; and the energy monitoring data comprises power load, gas consumption, renewable energy supply, and energy supply and demand balance state, etc. Subsequently, data cleaning is performed on these data, for possible abnormal values, threshold filtering and statistical models (such as Z-Score and IQR method) are used to automatically identify and eliminate abnormal values obviously beyond the physical meaning; for possible missing values, linear interpolation method is used to fill the missing values for continuous data (such as temperature and power consumption), and mode is used to fill for classification data (such as device state), and for missing rate exceeding 30%, it is directly eliminated; for possible duplicate values, it is de-duplicated based on the composite primary key of timestamp and sensor ID. After completing the data cleaning of the multi-scene monitoring data set, a set of high-quality and high-credibility multi-scene feature data can be obtained, which provides a reliable data basis for subsequent multi-dimensional modeling.

[0020] The risk analysis module 12 is configured to perform multi-scene risk analysis based on the city digital twin model, and construct a multi-scene risk map.

[0021] Specifically, in the risk analysis module 12, after the city digital twin model is constructed, abnormal detection and deduction are performed on infrastructure risk, traffic risk, and energy risk in three different scenes, and the corresponding analysis network is used to analyze the deduction result, and a multi-scene risk map is constructed, which reflects the risk distribution and severity of each single field, and provides a complete and visual risk panoramic map for city managers, which is used for subsequent operation decision and optimization of security.

[0022] Further, the risk analysis module 12 comprises:

[0023] According to the city digital twin model, infrastructure risk analysis is performed to obtain an infrastructure risk map; according to the city digital twin model, traffic risk analysis is performed to obtain a traffic risk map; according to the city digital twin model, energy risk analysis is performed to obtain an energy risk map; and the infrastructure risk map, the traffic risk map, and the energy risk map are integrated to obtain the multi-scene risk map.

[0024] Preferably, in the risk analysis module 12, according to the urban digital twin model, infrastructure risk analysis, traffic risk analysis and energy risk analysis will be carried out respectively. In terms of infrastructure risk analysis, multi-point monitoring data such as buildings, bridges, pipe networks and power lines accessed in the urban digital twin model are used for anomaly detection to identify possible structural damage, equipment failure and other abnormalities. Combined with historical accident data, accident deduction is carried out to predict the occurrence path and impact range of potential risk events. The deduction results are input into the trained infrastructure risk analysis network for infrastructure risk analysis to generate an infrastructure risk map reflecting the safety status of the facility. In terms of traffic risk analysis, multi-point monitoring data such as road traffic, vehicle speed, road congestion, traffic signal status and public transportation operation accessed in the urban digital twin model are used for anomaly detection to identify possible traffic bottlenecks and potential congestion trends. Combined with historical accident data, accident deduction is carried out to predict the occurrence path and impact range of potential risk events. The deduction results are input into the trained traffic risk analysis network for traffic risk analysis to generate a traffic risk map reflecting the traffic safety status. In terms of energy risk analysis, multi-point monitoring data such as power load, gas consumption and renewable energy supply accessed in the urban digital twin model are used for anomaly detection to identify possible energy supply and demand imbalance, equipment overload and energy supply interruption. Combined with historical accident data, accident deduction is carried out to predict the occurrence path and impact range of potential risk events. The deduction results are input into the trained energy risk analysis network for energy risk analysis to generate an energy risk map reflecting the energy safety status. Finally, the infrastructure risk map, traffic risk map and energy risk map are integrated, i.e. the infrastructure risk, traffic risk and energy risk in the same area are mapped to construct a final multi-scenario risk map. This multi-scenario risk map not only shows the risk distribution characteristics in each field, but also reveals the coupling effect and linkage mechanism between infrastructure, traffic and energy risks, providing a comprehensive, intuitive and dynamic risk support environment for urban operation and decision-making.

[0025] Further, the risk analysis module 12 comprises:

[0026] According to the urban digital twin model, multi-point facility anomaly detection is performed to obtain facility anomaly detection results. Accident deduction is performed on the facility anomaly detection results through the urban digital twin model to obtain facility accident deduction paths. A facility risk analysis network is trained according to the facility accident risk record set of the city. The facility risk analysis network is used to analyze the facility accident deduction paths to generate the infrastructure risk map.

[0027] Optionally, in the risk analysis module 12, according to the historical monitoring data of the multi-point site facility, the mean and standard deviation of each parameter are calculated, and the mean plus or minus multiple standard deviations (usually 2, which can be set according to actual needs) are used to construct the abnormal marking rule of each parameter. For example, when calculating the mean and standard deviation of the crack propagation rate, the abnormal marking rule constructed is that the crack propagation rate should not deviate from the mean plus or minus two standard deviations of the crack propagation rate. Then, the real-time synchronized infrastructure monitoring data in the city digital twin model is compared with the corresponding abnormal marking rule. If it is found that the monitoring value deviates from the abnormal marking rule and the duration exceeds multiple window periods (usually 3), an abnormal mark is triggered, and each facility anomaly detection result is generated, each of which includes facility ID, abnormal index, abnormal occurrence time, abnormal period data, etc. Then, for the detected abnormal situation, the accident deduction function of the city digital twin model is called to simulate the expansion of the facility anomaly in the time dimension and the influence in the space dimension. Specifically, representative historical cases (such as gas pipeline leakage and bridge structure collapse) are extracted from the facility accident risk record set, the trigger conditions (such as pipeline pressure mutation exceeding 1.2 times of the design value and bridge load overrun causing resonance) and evolution paths (such as gas diffusion path after leakage and failure sequence of key components during collapse) of each case are marked, forming reusable case templates, and a virtual accident simulation environment is constructed based on the city digital twin model. The multi-point site facility anomaly detection result is input into the model as the initial disturbance, and an accident evolution rule library is defined to quantify the risk transmission logic. The rules in the accident evolution rule library can be "when the pipeline pressure exceeds 1.2 times of the design value, the leakage probability increases exponentially with time" and "when the bridge vibration frequency enters the resonance band, the structure damage rate increases to 3 times of the normal state", etc., thereby providing rule support for dynamic deduction of the virtual scene. Then, the uncertain parameters (such as rainfall, traffic load, and material fatigue coefficient) in the accident evolution are modeled as probability distributions (such as rainfall following a normal distribution and traffic flow following a Poisson distribution), and the Latin hypercube sampling (LHS) method is used to generate 1000 groups of random scenarios, covering different combinations of environmental conditions and facility states. Then, relying on a GPU cluster, a parallel deduction engine is built to evolve the accident in each scenario, calculate the accident consequence indicators (such as gas leakage volume, bridge collapse range, and secondary disaster affected area), and record the complete path of the accident deduction, including key nodes (such as the initial leakage point and secondary damage location) and time series (such as the evolution stages of "leakage occurs → gas diffusion → triggers surrounding road collapse"). Finally, by integrating the deduction results of multiple scenarios, the accident deduction path of each facility can be generated, providing data support for risk analysis and emergency decision-making.Further, the historical accident deduction path, historical accident consequence index and historical anomaly detection result (such as risk level, propagation probability, main factor, etc.) are further obtained from the city facility accident risk record set to construct a facility accident risk record set, which is used as a training sample to train the facility risk analysis network. The facility risk analysis network can adopt a deep learning structure such as a graph neural network, a deep neural network, and a long short-term memory network. By inputting the training sample into the deep learning structure, the steps of forward propagation, loss calculation (such as the weighted sum of cross-entropy and mean square error), back propagation, parameter optimization (such as Adam optimizer, learning rate initialized to 1e-4, and decayed to 0.9 times of the original value every 10 rounds) are iteratively trained until the maximum number of iterations or the loss function converges. After training, the data not used for training is used for verification. If the verification result shows that the accuracy rate meets the preset accuracy rate, the current facility risk analysis network is output. Otherwise, the network structure parameters (such as the number of layers, the number of neurons, the type of activation function, the regularization coefficient, or the optimizer hyperparameters (such as learning rate, weight decay)) are adjusted, and the iterative training is performed again until the verification accuracy rate meets the preset accuracy rate, so that the facility risk analysis network has the ability to evaluate and analyze the risk of complex accident evolution path. Finally, each facility accident deduction path is input into the trained facility risk analysis network, and the anomaly detection result of each path at different evolution stages is output, and the result is presented in a graphed way to generate an infrastructure risk graph. This infrastructure risk graph not only can visually show the risk distribution of urban infrastructure, but also can reveal the risk transmission relationship between different facilities, thereby providing data support and decision basis for subsequent operation and maintenance decision and emergency response.

[0028] The operation and maintenance decision module 13 is configured to make operation and maintenance decisions for the city according to the multi-scenario risk graph, and obtain a multi-scenario operation and maintenance space.

[0029] Specifically, in the operation and maintenance decision module 13, after the multi-scenario risk graph is constructed, the potential risks of urban infrastructure, transportation system and energy system are comprehensively analyzed according to the multi-scenario risk graph to determine the key operation and maintenance needs of various facilities and regions. In the aspect of infrastructure, by performing infrastructure operation and maintenance decision on the data of the infrastructure risk part in the multi-scenario risk graph, multiple operation and maintenance strategies for buildings, bridges, pipe networks, etc. are generated, such as inspection frequency adjustment, fault prevention measure arrangement, maintenance and reinforcement plan formulation, etc., thereby forming an infrastructure operation and maintenance space. In the aspect of transportation, by performing transportation operation and maintenance decision on the data of the transportation risk part in the multi-scenario risk graph, multiple operation and maintenance strategies for congestion-prone road sections, accident-prone areas, etc. are generated, such as traffic control scheme, signal optimization strategy, emergency passage scheme, etc., thereby forming a transportation operation and maintenance space. In the aspect of energy, by performing energy operation and maintenance decision on the data of the energy risk part in the multi-scenario risk graph, multiple operation and maintenance strategies for load abnormalities, equipment overloads, etc. are generated, such as power, gas, renewable energy scheduling schemes, etc., thereby forming an energy operation and maintenance space. Finally, the infrastructure operation and maintenance space, the transportation operation and maintenance space and the energy operation and maintenance space are integrated to uniformly represent the operation and maintenance strategies, resource allocation and emergency response schemes of the city under various scenarios, to provide data basis and decision support for subsequent operation and maintenance evaluation and optimization, to realize scientific scheduling of urban resources and effective control of risks.

[0030] Further, the operation and maintenance decision module 13 includes:

[0031] According to the infrastructure risk graph, performing infrastructure operation and maintenance decision on the city to obtain an infrastructure operation and maintenance space; according to the transportation risk graph, performing transportation operation and maintenance decision on the city to obtain a transportation operation and maintenance space; according to the energy risk graph, performing energy operation and maintenance decision on the city to obtain an energy operation and maintenance space; integrating the infrastructure operation and maintenance space, the transportation operation and maintenance space and the energy operation and maintenance space to generate the multi-scenario operation and maintenance space.

[0032] Preferably, in the operation and maintenance decision module 13, according to the infrastructure risk graph in the multi-scene risk graph, the abnormal detection result of each region is obtained, for the region with abnormality, that is, the region with non-empty abnormal detection result, the risk level, the propagation probability and the main factor in the region abnormal detection result are extracted from the infrastructure operation and maintenance strategy rule library to extract all infrastructure operation and maintenance schemes that meet the three requirements, such as possible inspection plan, preventive maintenance measure, fault repair strategy and emergency response plan, and the infrastructure operation and maintenance schemes are summarized as an infrastructure operation and maintenance space, wherein the infrastructure operation and maintenance scheme includes the past operation and maintenance scheme and the corresponding risk scene, and the expert pre-set operation and maintenance scheme and the corresponding preset risk scene. Similarly, according to the traffic risk graph and the energy risk graph in the multi-scene risk graph, the traffic abnormal detection result and the energy abnormal detection result of each region are obtained, and then the traffic abnormal detection result and the energy abnormal detection result are extracted from the traffic operation and maintenance strategy rule library and the energy operation and maintenance strategy rule library to extract all traffic facility operation and maintenance schemes and energy operation and maintenance strategy rule libraries that meet the three requirements, such as possible signal optimization strategy, road control measure, dynamic scheduling scheme, emergency dredging scheme, energy scheduling optimization strategy, standby energy starting scheme, equipment maintenance plan and supply-demand balance adjustment scheme, so as to construct a traffic operation and maintenance space and an energy operation and maintenance space. Finally, the infrastructure operation and maintenance space, the traffic operation and maintenance space and the energy operation and maintenance space are integrated according to the corresponding positions, so as to uniformly represent and associate the operation and maintenance schemes of each scene, and form a complete multi-scene operation and maintenance space. The multi-scene operation and maintenance space not only contains the specific content and implementation path of each type of operation and maintenance scheme, but also embodies the resource coordination and strategy linkage between different scenes, and provides unified decision support and data basis for subsequent operation and maintenance evaluation, optimization and risk adaptive management.

[0033] The evaluation and optimization module 14 is used for evaluating and optimizing the multi-scene operation and maintenance space by the multi-scene guarantee analysis model to obtain a multi-scene operation and maintenance population.

[0034] Specifically, in the evaluation and optimization module 14, after obtaining the multi-scenario operation and maintenance space, a multi-scenario guarantee analysis model is activated, which includes an infrastructure guarantee evaluation model, a traffic guarantee evaluation model, and an energy guarantee evaluation model. Each sub-model corresponds to the operation and maintenance guarantee requirements and evaluation indexes of different scenarios. For the infrastructure operation and maintenance space in the multi-scenario operation and maintenance space, each operation and maintenance scheme in the infrastructure operation and maintenance space is simulated and operated, simulation data is collected, and the simulation data is input into the infrastructure guarantee evaluation model in the multi-scenario guarantee analysis model. The scheme that meets the infrastructure guarantee threshold is obtained and added to the multi-scenario operation and maintenance population. Similarly, each operation and maintenance scheme in the traffic operation and maintenance space and the energy operation and maintenance space is also simulated and operated, traffic simulation data and energy simulation data are collected, and the obtained traffic simulation data and energy simulation data are input into the traffic guarantee evaluation model and the energy guarantee evaluation model in the multi-scenario guarantee analysis model, respectively. The scheme that meets the traffic guarantee threshold and the energy guarantee threshold is obtained and added to the multi-scenario operation and maintenance population. The final multi-scenario operation and maintenance population will contain high-quality operation and maintenance schemes that have been evaluated, screened, and optimized in various scenarios, providing a basis for subsequent global evolutionary optimization to achieve efficient scheduling and risk control of urban resources.

[0035] Further, the evaluation and optimization module 14 includes:

[0036] The infrastructure guarantee evaluation model embedded in the multi-scenario guarantee analysis model is activated; the city digital twin model is simulated and operated according to each infrastructure operation and maintenance scheme in the infrastructure operation and maintenance space, and a plurality of facility operation and maintenance simulation data is obtained; the plurality of facility operation and maintenance simulation data is input into the infrastructure guarantee evaluation model, and an infrastructure guarantee evaluation sequence is obtained; based on the infrastructure guarantee evaluation sequence, the infrastructure operation and maintenance space is optimized and screened according to the infrastructure guarantee threshold, an infrastructure operation and maintenance population is obtained, and the infrastructure operation and maintenance population is added to the multi-scenario operation and maintenance population.

[0037] Preferably, in the evaluation optimization module 14, when evaluating the infrastructure operation space, the embedded infrastructure security evaluation model in the multi-scenario security analysis model will be activated first. This infrastructure security evaluation model is a pre-trained machine learning model, which can usually be built with a deep neural network. Then, each infrastructure operation scheme in the infrastructure operation space is input into the city digital twin model in turn. The city digital twin model will make corresponding adjustments according to each scheme and generate operation simulation data corresponding to each scheme through simulation. This simulation data covers dynamic information such as infrastructure operation state changes, fault probability evolution trends, maintenance duration, and resource consumption. After that, multiple facility operation simulation data are input into the infrastructure security evaluation model. This infrastructure security evaluation model analyzes the received simulation data according to the learned mapping relationship and calculates the infrastructure security evaluation coefficient of each infrastructure operation scheme. By sorting these infrastructure security evaluation coefficients according to the order of infrastructure operation scheme simulation, an infrastructure security evaluation sequence is obtained. The larger the infrastructure security evaluation coefficient in this sequence, the better the comprehensive performance of the corresponding scheme in ensuring the safe operation of urban infrastructure. Then, according to the pre-set infrastructure security threshold, the infrastructure security evaluation sequence is optimized and screened to identify infrastructure operation schemes with infrastructure security evaluation coefficients greater than or equal to the threshold and eliminate schemes that do not meet the requirements, ensuring that the screening results can meet the minimum security needs of city operation. Finally, all operation schemes that pass the threshold screening will form an infrastructure operation population, and this infrastructure operation population will be added to the multi-scenario operation population, laying a reliable data and scheme foundation for subsequent global evolution optimization.

[0038] Further, the evaluation optimization module 14 includes:

[0039] The multi-scenario security analysis model includes an infrastructure security evaluation model, a traffic security evaluation model, and an energy security evaluation model.

[0040] Preferably, the multi-scenario guarantee analysis model includes an infrastructure guarantee evaluation model, a traffic guarantee evaluation model, and an energy guarantee evaluation model. The infrastructure guarantee evaluation model takes infrastructure operation simulation data as input, focuses on the operation safety and stability of key facilities such as buildings, bridges, pipe networks, and power lines, and generates corresponding infrastructure guarantee evaluation coefficients to measure the effectiveness of each infrastructure operation scheme in risk control and operation guarantee in actual implementation. The traffic guarantee evaluation model takes traffic operation simulation data as input, understands the impact of different control strategies on complex traffic flow, and generates corresponding traffic guarantee evaluation coefficients to reflect its comprehensive ability in improving road operation efficiency and reducing risks. The energy guarantee evaluation model takes energy operation simulation data as input, comprehensively evaluates the pros and cons of energy status, and generates corresponding energy guarantee evaluation coefficients to measure the effectiveness of each energy operation scheme in guaranteeing energy security supply and optimizing resource utilization. These three sub-models are built by deep neural networks or multi-layer perception mechanisms, and have been iteratively trained using corresponding sample operation data and sample guarantee evaluation coefficients through steps such as forward propagation, loss calculation (such as mean square error), back propagation, and parameter optimization (such as Adam optimizer). They can comprehensively evaluate and screen different operation schemes in the multi-scenario operation space, providing a basis and unified standard for subsequent multi-scenario operation population construction and global optimization.

[0041] The global optimization module 15 is configured to perform global evolutionary optimization on the multi-scenario operation population according to a predetermined number of evolutions to determine a multi-scenario operation optimization result.

[0042] Specifically, in the global optimization module 15, the multi-scenario operation population is iteratively evolved according to a predetermined number of evolutions and evolution rules. In each evolution process, the guarantee evaluation of each operation scheme in the current population is identified to determine the best-performing group of schemes, which is used as a guide factor. Then, based on the guide factor, the other schemes in the population are evolved to generate new candidate schemes. Subsequently, the new candidate schemes are evaluated again using the guarantee evaluation models of each sub-scenario, and the best-performing operation evolution population is selected through guarantee threshold screening and identification. This evolution and screening process is repeated until the predetermined number of evolutions is completed, and the multi-scenario operation optimization result is obtained. This multi-scenario operation optimization result is used as the basis for decision-making of the city risk adaptive cloud management, thereby realizing the output of the globally optimal operation strategy of the city under complex scenarios.

[0043] Further, the global optimization module 15 includes:

[0044] The first infrastructure operation guiding factor is obtained by performing protection evaluation and optimal identification on the infrastructure operation population; the first operation evolution optimization group is obtained by performing evolution optimization on the infrastructure operation population based on the first infrastructure operation guiding factor according to the infrastructure protection evaluation model; the second infrastructure operation guiding factor is obtained by performing protection evaluation and optimal identification on the first operation evolution optimization group; the operation evolution optimization groups meeting the predetermined evolution times are obtained by continuously performing evolution optimization on the first operation evolution optimization group based on the second infrastructure operation guiding factor according to the infrastructure protection evaluation model; the infrastructure operation optimization result is obtained by performing protection evaluation global optimization on the infrastructure operation population and the operation evolution optimization groups, and the infrastructure operation optimization result is added to the multi-scenario operation optimization result.

[0045] Preferably, in the global optimization module 15, for the infrastructure operation population in the multi-scenario operation population, the protection evaluation optimal identification is performed according to the size of the infrastructure protection evaluation coefficient of each operation scheme in the infrastructure operation population, and the operation scheme with the largest protection evaluation coefficient is extracted as the first infrastructure operation guiding factor, which represents the optimal infrastructure operation scheme at the current stage. Subsequently, based on the first infrastructure operation guiding factor, the evolution optimization is performed on the infrastructure operation population, in which the difference comparison is performed, a plurality of evolution adjustment vectors are generated, and the other operation schemes in the population are adjusted and mutated according to the evolution adjustment vectors, thereby generating new candidate schemes to form the first operation evolution optimization group. The protection evaluation optimal identification is performed on the first operation evolution optimization group, and the operation scheme with the largest protection evaluation coefficient is selected as the second infrastructure operation guiding factor. Then, the difference evolution and screening are continuously performed on the first operation evolution optimization group based on the second infrastructure operation guiding factor, a new evolution group is formed again, and the evolution optimization and optimal identification process are repeated. Through multiple rounds of iterative evolution, the operation evolution optimization groups are generated until the predetermined evolution times are reached. Finally, the initial infrastructure operation population and the operation evolution optimization groups generated by each round of evolution are compared, the operation scheme with the largest protection evaluation coefficient is found as the global optimal solution, and the infrastructure operation optimization result is determined. For the traffic operation population and the energy operation population in the multi-scenario operation population, the similar steps are also performed to obtain the corresponding traffic operation optimization result and energy operation optimization result. By adding the infrastructure operation optimization result, the traffic operation optimization result and the energy operation optimization result to the multi-scenario operation optimization result, it can be ensured that the infrastructure operation scheme not only performs well in local evolution, but also reaches the optimal solution in the global range, thereby providing the optimal infrastructure protection decision for urban operation.

[0046] Further, the global optimization module 15 comprises:

[0047] According to the facility operation and maintenance first guide factor, difference detection is performed on the infrastructure operation and maintenance population, a plurality of operation and maintenance difference vectors are obtained, random variation is performed on each operation and maintenance difference vector, a plurality of evolution adjustment vector sets are obtained, the infrastructure operation and maintenance population is adjusted according to the plurality of evolution adjustment vector sets, a first operation and maintenance evolution group is obtained, and based on an infrastructure guarantee threshold, the first operation and maintenance evolution group is evaluated and optimized according to the infrastructure guarantee evaluation model, and the first operation and maintenance evolution optimization group is obtained.

[0048] Optionally, in the global optimization module 15, when evolution optimization is performed, first, the facility operation and maintenance first guide factor is taken as a benchmark scheme, difference detection is performed on the remaining operation and maintenance schemes in the infrastructure operation and maintenance population, and a plurality of operation and maintenance difference vectors are generated by comparing the differences of the schemes in the construction period, material type, cost budget, construction period and other key parameters. Each difference vector represents the performance difference characteristics between a candidate scheme and the guide factor. Subsequently, random variation is performed on each operation and maintenance difference vector, and the size or direction of part of the difference parameter values is randomly adjusted, thereby obtaining a plurality of evolution adjustment vector sets. Based on the evolution adjustment vector set, the original infrastructure operation and maintenance population is parameter corrected and scheme adjusted, a new candidate scheme group is generated, and the candidate scheme group is taken as the first operation and maintenance evolution group. Then, the city digital twin model is called, and each candidate scheme in the first operation and maintenance evolution group is simulated to obtain simulation data of each candidate scheme. Then, the embedded infrastructure guarantee evaluation model in the multi-scenario guarantee analysis model is activated, the infrastructure guarantee evaluation coefficients corresponding to the simulation data are calculated, and based on the preset infrastructure guarantee threshold, the candidate schemes with the guarantee evaluation coefficients greater than or equal to the threshold are extracted to generate a first operation and maintenance evolution optimization group. The first operation and maintenance evolution optimization group is composed of high-quality operation and maintenance schemes generated by evolution and verified by the guarantee threshold, and provides basic data support for subsequent iteration evolution and global optimization.

[0049] The cloud management module 16 is configured to perform risk adaptive cloud management of the city according to the multi-scenario operation and maintenance optimization result.

[0050] Specifically, in the cloud management module 16, after obtaining the multi-scene operation and maintenance optimization result, the multi-scene operation and maintenance optimization result is uploaded to the city-level cloud management platform. The multi-scene operation and maintenance optimization result includes the optimal operation and maintenance scheme filtered and optimized in multiple scenes such as infrastructure, transportation and energy. The cloud management platform uniformly analyzes and distributes these optimization results, and maps the optimal schemes in different scenes to each functional module of city operation and management. For the infrastructure field, the cloud management platform issues inspection plans, maintenance and reinforcement instructions, emergency disposal schemes, etc. to the corresponding facility management subsystem according to the optimization result, to realize dynamic protection of facilities such as buildings, pipelines and bridges; in the transportation field, the cloud management platform adjusts the signal lamp control system, road control system, public transportation scheduling system, etc. based on the traffic operation and maintenance optimization result, to ensure smooth roads and minimize accident risks; in the energy field, the cloud management platform optimizes the use of power load distribution, gas scheduling, renewable energy, etc. according to the energy operation and maintenance optimization result, to ensure energy supply and demand balance and supply safety. Through the above process, the city realizes risk adaptive management in the cloud environment, not only ensuring the optimality and real-time of operation and maintenance strategies, but also improving the overall safety, resilience and resource utilization efficiency of the city.

[0051] In summary, the city-level digital twin cloud management platform supporting multi-scene access provided by the embodiments of the present application has the following technical effects:

[0052] The multi-dimensional modeling module 11 is configured to perform multi-dimensional modeling according to multi-scene characteristic data of the city to construct a city digital twin model. The risk analysis module 12 is configured to perform multi-scene risk analysis according to the city digital twin model to construct a multi-scene risk atlas. The operation and maintenance decision module 13 is configured to perform operation and maintenance decision for the city according to the multi-scene risk atlas to obtain a multi-scene operation and maintenance space. The evaluation and optimization module 14 is configured to perform evaluation and optimization on the multi-scene operation and maintenance space through a multi-scene guarantee analysis model to obtain a multi-scene operation and maintenance population. The global optimization module 15 is configured to perform global evolutionary optimization on the multi-scene operation and maintenance population according to a predetermined number of evolutions to determine a multi-scene operation and maintenance optimization result. The cloud management module 16 is configured to perform risk adaptive cloud management of the city according to the multi-scene operation and maintenance optimization result. Through the above steps, the technical problem of difficulty in multi-scene cooperation in the existing digital twin city management, resulting in one-sided and lagging risk analysis and low management efficiency, is solved. By constructing a multi-scene risk atlas and combining an evolutionary optimization algorithm, the technical effects of improving city operation and maintenance efficiency and enhancing risk response capability are achieved.

[0053] Embodiment two, based on the same inventive concept as the city-level digital twin cloud management platform supporting multi-scene access in the foregoing embodiments, like Figure 2As shown, the embodiments of the present application provide a city-level digital twin cloud management method supporting multi-scenario access, which comprises:

[0054] According to the multi-scenario feature data of the city, a multi-dimensional modeling is performed to construct a city digital twin model; according to the city digital twin model, a multi-scenario risk analysis is performed to construct a multi-scenario risk atlas; according to the multi-scenario risk atlas, an operation and maintenance decision is made for the city to obtain a multi-scenario operation and maintenance space; through a multi-scenario guarantee analysis model, the multi-scenario operation and maintenance space is evaluated and optimized to obtain a multi-scenario operation and maintenance population; according to a predetermined evolution number, the multi-scenario operation and maintenance population is globally evolved and optimized to determine a multi-scenario operation and maintenance optimization result; and according to the multi-scenario operation and maintenance optimization result, a risk self-adaptive cloud management is performed for the city.

[0055] Further, the method comprises:

[0056] According to the city digital twin model, an infrastructure risk analysis is performed to obtain an infrastructure risk atlas; according to the city digital twin model, a traffic risk analysis is performed to obtain a traffic risk atlas; according to the city digital twin model, an energy risk analysis is performed to obtain an energy risk atlas; and the infrastructure risk atlas, the traffic risk atlas and the energy risk atlas are integrated to obtain the multi-scenario risk atlas.

[0057] Further, the method comprises:

[0058] According to the city digital twin model, a multi-point facility anomaly detection is performed to obtain each facility anomaly detection result; through the city digital twin model, an accident deduction is performed on the each facility anomaly detection result to obtain each facility accident deduction path; according to a facility accident risk record set of the city, a facility risk analysis network is trained; through the facility risk analysis network, a risk analysis is performed on the each facility accident deduction path to generate the infrastructure risk atlas.

[0059] Further, the method comprises:

[0060] According to the infrastructure risk atlas, an infrastructure operation and maintenance decision is made for the city to obtain an infrastructure operation and maintenance space; according to the traffic risk atlas, a traffic operation and maintenance decision is made for the city to obtain a traffic operation and maintenance space; according to the energy risk atlas, an energy operation and maintenance decision is made for the city to obtain an energy operation and maintenance space; and the infrastructure operation and maintenance space, the traffic operation and maintenance space and the energy operation and maintenance space are integrated to generate the multi-scenario operation and maintenance space.

[0061] Further, the method comprises:

[0062] activating an infrastructure guarantee evaluation model embedded in the multi-scenario guarantee analysis model; simulating operation and maintenance of each infrastructure operation and maintenance scheme in an infrastructure operation and maintenance space on the city digital twin model to obtain a plurality of facility operation and maintenance simulation data; inputting the plurality of facility operation and maintenance simulation data into the infrastructure guarantee evaluation model to obtain an infrastructure guarantee evaluation sequence; and based on the infrastructure guarantee evaluation sequence, screening the infrastructure operation and maintenance space according to an infrastructure guarantee threshold to obtain an infrastructure operation and maintenance population, and adding the infrastructure operation and maintenance population to the multi-scenario operation and maintenance population.

[0063] Further, the method comprises:

[0064] According to the infrastructure operation and maintenance population, guarantee evaluation and winner identification are performed to obtain a facility operation and maintenance first guide factor; based on the facility operation and maintenance first guide factor, evolution optimization of the infrastructure operation and maintenance population is performed according to the infrastructure guarantee evaluation model to obtain a first operation and maintenance evolution optimization group; according to the first operation and maintenance evolution optimization group, guarantee evaluation and winner identification are performed to obtain a facility operation and maintenance second guide factor; based on the facility operation and maintenance second guide factor, evolution optimization of the first operation and maintenance evolution optimization group is continued according to the infrastructure guarantee evaluation model until a plurality of operation and maintenance evolution optimization groups satisfying a predetermined evolution number of times are obtained; guarantee evaluation global optimization is performed according to the infrastructure operation and maintenance population and the plurality of operation and maintenance evolution optimization groups to obtain an infrastructure operation and maintenance optimization result, and the infrastructure operation and maintenance optimization result is added to the multi-scenario operation and maintenance optimization result.

[0065] Further, the method comprises:

[0066] According to the facility operation and maintenance first guide factor, difference detection is performed on the infrastructure operation and maintenance population to obtain a plurality of operation and maintenance difference vectors; random mutation is performed according to each operation and maintenance difference vector to obtain a plurality of evolution adjustment vector sets; adjustment is performed on the infrastructure operation and maintenance population according to the plurality of evolution adjustment vector sets to obtain a first operation and maintenance evolution group; based on an infrastructure guarantee threshold, guarantee evaluation optimization is performed on the first operation and maintenance evolution group according to the infrastructure guarantee evaluation model to obtain the first operation and maintenance evolution optimization group.

[0067] Further, the method comprises:

[0068] A multi-scenario monitoring data set of the city is obtained, the multi-scenario monitoring data set comprising infrastructure monitoring data, traffic monitoring data and energy monitoring data; data cleaning is performed on the multi-scenario monitoring data set to generate the multi-scenario feature data.

[0069] Further, the method comprises:

[0070] The multi-scenario guarantee analysis model comprises an infrastructure guarantee evaluation model, a traffic guarantee evaluation model and an energy guarantee evaluation model.

[0071] Any step of the method described above can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor to implement any method in the embodiments of the present application, and no redundant limitation is made herein.

[0072] Further, the first or the second described above may not only represent an order relationship, but also may represent a certain specific concept, and / or refer to the selection of multiple elements individually or collectively. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technologies, the present application is intended to include these modifications and variations.

Claims

1. A city-level digital twin cloud management platform supporting multi-scenario access, characterized in that, The platform comprises: a multi-dimensional modeling module for multi-dimensional modeling according to multi-scene characteristic data of a city to construct a city digital twin model; a risk analysis module for multi-scene risk analysis according to the city digital twin model to construct a multi-scene risk atlas; an operation and maintenance decision module for operation and maintenance decision of the city according to the multi-scene risk atlas to obtain a multi-scene operation and maintenance space; an evaluation and optimization module for evaluation and optimization of the multi-scene operation and maintenance space through a multi-scene guarantee analysis model to obtain a multi-scene operation and maintenance population; a global optimization module for global evolutionary optimization of the multi-scene operation and maintenance population according to a predetermined evolution number to determine a multi-scene operation and maintenance optimization result; a cloud management module for risk adaptive cloud management of the city according to the multi-scene operation and maintenance optimization result; the global optimization module comprises: guarantee evaluation and winner identification according to an infrastructure operation and maintenance population to obtain an infrastructure operation and maintenance first guide factor; evolutionary optimization of the infrastructure operation and maintenance population according to an infrastructure guarantee evaluation model based on the infrastructure operation and maintenance first guide factor to obtain a first operation and maintenance evolutionary optimization group; guarantee evaluation and winner identification according to the first operation and maintenance evolutionary optimization group to obtain an infrastructure operation and maintenance second guide factor; continuing evolutionary optimization of the first operation and maintenance evolutionary optimization group according to the infrastructure guarantee evaluation model based on the infrastructure operation and maintenance second guide factor until a plurality of operation and maintenance evolutionary optimization groups satisfying the predetermined evolution number are obtained; guarantee evaluation global optimization according to the infrastructure operation and maintenance population and the plurality of operation and maintenance evolutionary optimization groups to obtain an infrastructure operation and maintenance optimization result, and adding the infrastructure operation and maintenance optimization result to the multi-scene operation and maintenance optimization result; the global optimization module comprises: difference detection of the infrastructure operation and maintenance population according to the infrastructure operation and maintenance first guide factor to obtain a plurality of operation and maintenance difference vectors; a plurality of evolutionary adjustment vector sets are obtained by random mutation according to each operation and maintenance difference vector; adjustment of the infrastructure operation and maintenance population according to the plurality of evolutionary adjustment vector sets to obtain a first operation and maintenance evolutionary group; guarantee evaluation and optimization of the first operation and maintenance evolutionary group according to the infrastructure guarantee evaluation model based on an infrastructure guarantee threshold to obtain the first operation and maintenance evolutionary optimization group.

2. The city-level digital twin cloud management platform supporting multi-scene access according to claim 1, wherein, the risk analysis module comprises: infrastructure risk analysis according to the city digital twin model to obtain an infrastructure risk atlas; traffic risk analysis according to the city digital twin model to obtain a traffic risk atlas; energy risk analysis according to the city digital twin model to obtain an energy risk atlas; integrating the infrastructure risk atlas, the traffic risk atlas and the energy risk atlas to obtain the multi-scene risk atlas.

3. The city-level digital twin cloud management platform supporting multi-scene access according to claim 2, wherein, the risk analysis module comprises: multi-point facility anomaly detection according to the city digital twin model to obtain each facility anomaly detection result; accident deduction of each facility anomaly detection result through the city digital twin model to obtain each facility accident deduction path; training of a facility risk analysis network according to a facility accident risk record set of the city; The infrastructure risk analysis network analyzes the accident deduction paths of the infrastructures to generate an infrastructure risk map.

4. The city-level digital twin cloud management platform supporting multi-scene access of claim 1, wherein, The operation and maintenance decision module includes: According to the infrastructure risk map, an infrastructure operation and maintenance space is obtained. According to the traffic risk map, a traffic operation and maintenance space is obtained. According to the energy risk map, an energy operation and maintenance space is obtained. The infrastructure operation and maintenance space, the traffic operation and maintenance space, and the energy operation and maintenance space are integrated to generate the multi-scenario operation and maintenance space.

5. The city-level digital twin cloud management platform supporting multi-scene access according to claim 1, wherein, The evaluation optimization module includes: An infrastructure guarantee evaluation model embedded in the multi-scenario guarantee analysis model is activated. According to each infrastructure operation and maintenance scheme in the infrastructure operation and maintenance space, the city digital twin model is simulated and operated to obtain a plurality of infrastructure operation and maintenance simulation data. The plurality of infrastructure operation and maintenance simulation data are input into the infrastructure guarantee evaluation model to obtain an infrastructure guarantee evaluation sequence. Based on the infrastructure guarantee evaluation sequence, the infrastructure operation and maintenance space is screened according to an infrastructure guarantee threshold to obtain an infrastructure operation and maintenance population, and the infrastructure operation and maintenance population is added to the multi-scenario operation and maintenance population.

6. The city-level digital twin cloud management platform supporting multi-scene access of claim 1, wherein, The multi-dimensional modeling module includes: A multi-scenario monitoring data set of the city is obtained, including infrastructure monitoring data, traffic monitoring data, and energy monitoring data. The multi-scenario monitoring data set is data cleaned to generate multi-scenario feature data.

7. The city-level digital twin cloud management platform supporting multi-scene access of claim 1, wherein, The multi-scenario guarantee analysis model includes an infrastructure guarantee evaluation model, a traffic guarantee evaluation model, and an energy guarantee evaluation model.

8. The city-level digital twin cloud management method supporting multi-scenario access, characterized in that, The city-level digital twin cloud management platform supporting multi-scenario access based on any one of claims 1-7 is implemented, and the method includes: According to the multi-scenario feature data of the city, a multi-dimensional model is established to construct a city digital twin model; According to the city digital twin model, multi-scenario risk analysis is performed to construct a multi-scenario risk map; According to the multi-scenario risk map, operation and maintenance decisions are made for the city to obtain a multi-scenario operation and maintenance space; Through the multi-scenario guarantee analysis model, the multi-scenario operation and maintenance space is evaluated and optimized to obtain a multi-scenario operation and maintenance population; According to the predetermined number of evolutions, the multi-scenario operation and maintenance population is globally evolved and optimized to determine a multi-scenario operation and maintenance optimization result; and According to the multi-scenario operation and maintenance optimization result, risk adaptive cloud management of the city is performed.

Images