Urban water ecological complex simulator, simulator operation method, computer readable storage medium and terminal

Abstract

The application relates to the technical field of data processing, and discloses a simulator of a city water ecological complex, a simulator operation method, a computer readable storage medium and a terminal. In the simulator, a data layer carries out real-time monitoring on an original data source and then carries out preprocessing to obtain a space-time profile data set; a model layer utilizes the space-time profile data set to carry out multiple simulations to obtain key interactive variables, so as to train a deep neural agent model, gradually approach a simulation result in a second-level time to realize water information prediction, and analyze the prediction result through a decision layer to obtain a scheduling-evaluation report; an interactive layer dynamically renders the space-time profile data set and the prediction result, constructs a virtual experience environment according to the scheduling-evaluation report, and obtains a visual simulation result; and a platform layer provides model library and knowledge base management, a standardized API interface, and cloud computing and high-performance computing support. The application improves the fine management and scientific decision-making capability of a water ecological system.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a simulator for an urban water ecological complex, a simulator operation method, a computer-readable storage medium, and a terminal. Background Technology

[0002] Urban water ecological complexes are a typical microcosm of the human-nature-society coupled system. The core of its scientific connotation lies in applying the principles of ecology, hydrology, environmental engineering and systems science to regard the gray (water engineering system), green (natural ecosystem) and blue (water environment system) infrastructure in the city as an organic living system. By simulating and strengthening the interaction and integration between natural hydrological processes, ecological functions and engineering facilities, it collaboratively solves water resources, water environment, water ecology and water security issues. The aim is to build a comprehensive and sustainable urban infrastructure that can self-regulate, provide diverse ecosystem services and achieve harmonious coexistence of "human-water-city".

[0003] However, existing governance and modeling frameworks are mostly based on natural hydrological processes, separating water allocation, water quality protection and ecological restoration. They also fail to adequately depict the coupling of human activity elements such as population density, industrial layout and engineering operation, thus making it difficult to truly reflect the feedback mechanism and collaborative goals of the "human-water-city" system.

[0004] Furthermore, at the data level, although multi-source observations are constantly being enriched, they exhibit significant dispersion and structural fragmentation. Ground monitoring points are dense but have limited spatial coverage, while satellite remote sensing has wide coverage but is limited by spatiotemporal resolution. Socioeconomic and human activity data mostly exist in the form of indirect proxies, resulting in a lack of consistent high-frequency and high-precision observation foundations between water quantity, water quality, ecology, and human activities.

[0005] The aforementioned complexity of urban water ecosystems necessitates the reliance on digital simulators for their research and management. Urban water ecosystems are highly complex, dynamic, and multi-objective coupled mega-systems, which traditional empirical methods and static planning are inadequate for effectively addressing: they cannot quantify the dynamic interactions between gray, green, and blue infrastructures, nor can they provide timely and reliable decision-making support under extreme weather conditions or sudden pollution events. Digital simulators, by constructing a virtual "digital twin," can integrate multi-dimensional data and models from hydrology, water quality, ecology, and society. In virtual space, they can accurately simulate and evaluate the comprehensive response of the system under different planning schemes, management strategies, or extreme scenarios. This allows for optimized design before construction, intelligent scheduling and risk warning during operation, and finding the optimal balance between ecological benefits, safety performance, and social services with minimal cost and risk. It is an indispensable core decision-making tool for achieving scientific and refined management of urban water ecosystems. Therefore, existing technologies require further improvement and development. Summary of the Invention

[0006] The main objective of this invention is to provide a simulator and method for urban water ecological complexes, aiming to solve the problems of fragmented and discrete structures, insufficient dynamic interaction characterization of three types of infrastructure, and inability to provide multi-objective collaborative optimization decision support in extreme scenarios in the process of simulating the water environment in existing technologies.

[0007] To achieve the above objectives, the present invention provides a simulator for an urban water ecological complex, wherein the simulator for the urban water ecological complex includes: a data layer, a model layer, a decision layer, an interaction layer, and a platform layer; The data layer is used to obtain original data sources of multiple dimensions of the target area through multi-source means, and to perform real-time monitoring and preprocessing on each of the original data sources to obtain the spatiotemporal profile dataset corresponding to each dimension. The model layer is used to perform natural system simulation and human system simulation using all the spatiotemporal profile datasets to obtain key interaction variables. The deep neural agent model is trained using the key interaction variables and the deep neural agent model is used to predict water information in different dimensions. The simulation results are gradually approximated within seconds, and the corresponding prediction results are output. The decision-making layer is used to construct multiple functional service models, evaluate the simulator response under different planning schemes, climate scenarios or emergencies using all the functional service models, and input all the prediction results into the corresponding functional service models respectively. It uses a multi-objective optimization algorithm to balance multiple objectives and generate and output the corresponding scheduling-evaluation report. The interaction layer is used to perform three-dimensional dynamic rendering of all the spatiotemporal profile datasets and all the prediction results based on GIS and game engine, construct a digital twin virtual experience environment according to all the scheduling-evaluation reports, and perform different inferences and comparisons on the virtual experience environment to obtain multiple visual simulation results. The platform layer provides model and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. It transmits all the visualization simulation results through data exchange with multiple external systems, including the City Information Modeling (CIM) platform and the City Water Dispatch System.

[0008] The data layer includes: a data acquisition unit, a data preprocessing unit, a data unification unit, and an intelligent parsing unit; The data acquisition unit is used to acquire original data sources of the target area in the dimensions of meteorological and hydrological monitoring, water quality, ecology, topography, human activities, transportation and travel and social development through multi-source means, and to monitor each of the original data sources in real time to obtain change data of each of the original data sources. The data preprocessing unit is used to correct the changed data, missing values ​​and outliers of each of the original data sources, use data cleaning algorithms to perform quality control and noise removal on the corrected original data sources, and then perform standardization processing to obtain the corresponding corrected data sources. The data unification unit is used to use a data assimilation algorithm to unify the resolution of all the corrected data sources into a preset grid in the time and space dimensions, so as to obtain the spatiotemporal profile dataset corresponding to each dimension. The intelligent parsing unit is used to parse unstructured text data through a large language model, extract key knowledge, and assist in building a background knowledge base. The unstructured text data includes historical research reports, rules, and operation and maintenance logs.

[0009] The model layer includes: a matching unit, a preliminary prediction unit, an optimization unit, and a final prediction unit; The matching unit is used to construct natural system models and human system models. It inputs all the spatiotemporal profile datasets into the natural system model and the human system model respectively through a standardized API interface for simulation, and performs pair fitting on the two types of simulation data to obtain key interaction variables. The preliminary prediction unit is used to predict water information of different dimensions based on the key interaction variables and the deep neural agent model, and output multiple initial prediction results, including: initial water volume simulation results, initial multi-source water demand, initial three-dimensional concentration field, initial water velocity field, initial ecological flow, and initial urban water affairs data. An optimization unit is used to calculate the error between the monitoring data and all initial prediction results. Whenever the error exceeds a threshold within a preset time, the weight parameters of the deep neural agent model are adjusted to iteratively optimize the deep neural agent model. The final prediction unit is used to predict water information in different dimensions using the current deep neural agent model, and outputs the current water volume simulation results, current multi-source water demand, current three-dimensional concentration field, current water flow velocity field, current ecological flow and current urban water affairs data.

[0010] The matching unit includes: a model building subunit, a first simulation subunit, a data input subunit, a second simulation subunit, a third simulation subunit, a fourth simulation subunit, a data analysis subunit, and a data fusion subunit; The model building subunit is used to build a hydrodynamic model, a water quality model, an ecological process model, and a socio-economic model of the target area, and to build a natural system model of the target area using the hydrodynamic model, the water quality model, the ecological process model, and the socio-economic model. The first simulation subunit is used to input all the spatiotemporal profile datasets into the natural system model through the standardized API interface. The natural system model performs scene simulation based on all the spatiotemporal profile datasets and outputs natural scene data. The data input subunit is used to construct a human system model by using the system dynamics model as the overall framework. Through the standardized API interface, all the spatiotemporal profile datasets are input into the infrastructure and engineering regulation submodule, socioeconomic and spatial development submodule, and dynamic feedback and evolution submodule of the human system model, respectively. The second simulation subunit is used to control the infrastructure and engineering control submodule to simulate the operation and control process of the gray-green-blue infrastructure in the target area based on all the spatiotemporal profile datasets, and output the intensity of human disturbance. The third simulation subunit is used to control the socio-economic and spatial development submodule to simulate the driving impact of social development on the aquatic ecosystem based on all the spatiotemporal profile datasets, and output the spatiotemporal distribution characteristics. The fourth simulation subunit is used to control the dynamic feedback and evolution submodule to simulate the evolution and mutual feedback mechanism of all elements in the target area according to all the spatiotemporal profile datasets, and output dynamic feedback data; The data analysis subunit is used to construct socially driven scenario data based on the intensity of human disturbance, the spatiotemporal distribution characteristics, and the dynamic feedback data. The data fusion subunit is used to pair and fit the natural scene data and the socially driven scene data using a data fusion algorithm to obtain coupled key interaction variables.

[0011] The decision-making layer includes: a scheme generation unit, an alarm and suggestion unit, a comprehensive quantification unit, a performance evaluation unit, and an intelligent optimization unit; The scheme generation unit is used to construct a water resource scheduling sub-model, input the current water volume simulation results and the current multi-source water demand into the water resource scheduling sub-model, and output the water resource scheduling scheme for the target area. The alarm and suggestion unit is used to construct a water environment risk early warning sub-model, input the current three-dimensional concentration field and the current water flow velocity field into the water environment risk early warning sub-model, and output multiple risk level signals and emergency suggestions corresponding to each risk level signal; The comprehensive quantitative unit is used to construct a water ecological integrity assessment sub-model. The current ecological flow is input into the water ecological integrity assessment sub-model, and a comprehensive quantitative assessment report of the water ecology is output. The performance evaluation unit is used to construct a water service performance evaluation sub-model, input the current urban water affairs data into the water service performance evaluation sub-model, and output a performance evaluation report; The intelligent optimization unit is used to automatically explore multiple management strategies using reinforcement learning algorithms, and to automatically correlate and analyze environmental distribution, meteorological and hydrological conditions and equipment operating status using knowledge graphs and large language models to obtain the decision basis for each management strategy.

[0012] The interaction layer includes: an interactive data input unit, a deduction unit, a comparison unit, a visualization unit, and an intelligent interaction unit; An interactive data input unit is used to construct an interactive model, inputting all the spatiotemporal profile datasets and all the prediction results into the interactive model; The inference unit is used to add multiple sets of parameters input by the user to the interaction model. Under different sets of parameters, the interaction model performs different scenario inferences based on all the spatiotemporal profile datasets and all the prediction results to obtain the corresponding scenario results. The comparison unit is used to compare the differences between the results of each scenario, obtain potential impact information under different parameter sets, and generate corresponding optimization scheduling suggestions based on all the potential impact information. A visualization unit is used to construct a virtual experience environment based on all the optimization scheduling suggestions, all the scheduling-evaluation reports and all the scenario results, and to render the virtual experience environment to obtain visualization simulation results under different scenarios; The intelligent interaction unit integrates a large language model as an intelligent interaction hub and automatically generates analysis reports containing data interpretation, chart explanations, and management suggestions based on simulation results. The intelligent interaction hub allows users to directly interact with the simulator in natural language to query simulation results.

[0013] The platform layer includes: a model library and knowledge base management unit, a standardized interface unit, and a computing support unit; The model library and knowledge base management unit is used to uniformly manage the versions, parameter configurations, and expert rules of all models in the simulator; A standardized interface unit is used to provide a standardized API interface for data and function interaction with all the aforementioned external systems; The computing support unit is used to support the rapid operation of all the aforementioned models and the efficient simulation of massive scenarios through elastic cloud computing resources and parallel computing capabilities.

[0014] Furthermore, to achieve the above objectives, the present invention also provides a simulator operation method for a simulator of an urban water ecological complex, wherein the simulator operation method for the urban water ecological complex simulator includes: The data layer obtains original data sources of multiple dimensions of the target region through multi-source methods, performs real-time monitoring on each of the original data sources and then performs preprocessing to obtain the spatiotemporal profile dataset corresponding to each dimension. The model layer uses all the spatiotemporal profile datasets to perform natural system simulation and human system simulation, obtains key interaction variables, uses the key interaction variables to train the deep neural agent model, and uses the deep neural agent model to predict water information in different dimensions, gradually approximating the simulation results in seconds, and outputting the corresponding prediction results. The decision-making level constructs multiple functional service models, uses all the functional service models to evaluate the simulator response under different planning schemes, climate scenarios or emergencies, and inputs all the prediction results into the corresponding functional service models respectively. It uses a multi-objective optimization algorithm to balance multiple objectives and generates and outputs the corresponding scheduling-evaluation report. The interaction layer performs three-dimensional dynamic rendering of all the spatiotemporal profile datasets and all the prediction results based on GIS and game engine, constructs a digital twin virtual experience environment based on all the scheduling-evaluation reports, and performs different simulations and comparisons on the virtual experience environment to obtain multiple visual simulation results. The platform layer provides model and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. It transmits all the visualization simulation results through data exchange with multiple external systems, including the City Information Modeling (CIM) platform and the City Water Dispatch System.

[0015] In addition, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a simulator program for an urban water ecological complex stored in the memory and executable on the processor, wherein when the simulator program for an urban water ecological complex is executed by the processor, it implements the steps of the simulator operation method for an urban water ecological complex as described above.

[0016] In addition, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a simulator program for an urban water ecological complex, and when the simulator program for an urban water ecological complex is executed by a processor, it implements the steps of the simulator operation method for an urban water ecological complex as described above.

[0017] This invention aims to solve water resource, water environment, water ecology and water security issues by simulating and enhancing the interaction and integration between natural hydrological processes, ecological functions and engineering facilities. It is designed to build a comprehensive and sustainable urban infrastructure that can self-regulate, provide diverse ecosystem services (such as rainwater storage, water purification, habitat creation and recreation) and achieve harmonious coexistence between people, water and city. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a simulator for the urban water ecological complex of the present invention; Figure 2 This is a schematic diagram of the processing flow in a preferred embodiment of the simulation operation method for urban water ecological complex of the present invention; Figure 3 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0021] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0022] One embodiment of the simulator for the urban water ecological complex described in the preferred embodiment of the present invention includes, as follows: Figure 1As shown, the simulator for the urban water ecosystem complex includes: a data layer 51, a model layer 52, a decision layer 53, an interaction layer 54, and a platform layer 55. The urban water ecosystem complex simulator of this invention is a simulation system and its hardware and software devices that couple natural processes with socio-economic processes, using the urban water ecosystem complex as the object and the water cycle as the link. Essentially, it is a large-scale scientific device designed to achieve integrated "monitoring-simulation-evaluation-early warning-decision-making" functions for the urban water ecosystem complex through big data and artificial intelligence technologies. A simulator is a tool that uses a computer model to simulate a real system or process for analysis, prediction, testing, and decision support. In the urban water ecosystem complex simulator, this concept is applied to the complex system of the urban water ecosystem complex. The simulator includes a terminal (i.e., a hardware layer) and a software layer; the software layer includes the simulator's operating program for the urban water ecosystem complex, used to achieve integrated operation of monitoring, simulation, evaluation, early warning, and decision-making for the target area. When the simulator executes the simulator's operating program in the software layer, it implements the steps of the simulator operation method for the urban water ecosystem complex as described above.

[0023] The data layer is used to obtain original data sources of multiple dimensions of the target area through multi-source means, and to perform real-time monitoring and preprocessing on each of the original data sources to obtain the spatiotemporal profile dataset corresponding to each dimension. The model layer is used to perform natural system simulation and human system simulation using all the spatiotemporal profile datasets to obtain key interaction variables. The deep neural agent model is trained using the key interaction variables and the deep neural agent model is used to predict water information in different dimensions. The simulation results are gradually approximated within seconds, and the corresponding prediction results are output. The decision-making layer is used to construct multiple functional service models, evaluate the simulator response under different planning schemes, climate scenarios or emergencies using all the functional service models, and input all the prediction results into the corresponding functional service models respectively. It uses a multi-objective optimization algorithm to balance multiple objectives and generate and output the corresponding scheduling-evaluation report. The interaction layer is used to perform three-dimensional dynamic rendering of all the spatiotemporal profile datasets and all the prediction results based on GIS (Geographic Information System) and game engine, construct a digital twin virtual experience environment according to all the scheduling-evaluation reports, and perform different inferences and comparisons on the virtual experience environment to obtain multiple visual simulation results; The platform layer provides model and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. It transmits all the visualization simulation results through data exchange with multiple external systems, including the City Information Modeling (CIM) platform and the city water management system.

[0024] In the existing hydrological simulation process, although the amount of observational data considered is constantly increasing, it is still significantly discrete and fragmented. This results in a lack of consistent high-frequency and high-precision observational foundations for water quantity, water quality, ecology, and human activities. Therefore, this invention manages all data in a unified manner when considering data from multiple dimensions, providing a unified data foundation for subsequent simulation and evaluation.

[0025] Specifically, the data layer includes: a data acquisition unit, a data preprocessing unit, a data unification unit, and an intelligent parsing unit; The data acquisition unit is used to acquire original data sources of the target area in the dimensions of meteorological and hydrological monitoring, water quality, ecology, topography, human activities, transportation and travel and social development through multi-source means, and to monitor each of the original data sources in real time to obtain change data of each of the original data sources. The data preprocessing unit is used to correct the changed data, missing values ​​and outliers of each of the original data sources, use data cleaning algorithms to perform quality control and noise removal on the corrected original data sources, and then perform standardization processing to obtain the corresponding corrected data sources. The data unification unit is used to use a data assimilation algorithm to unify the resolution of all the corrected data sources into a preset grid in the time and space dimensions, so as to obtain the spatiotemporal profile dataset corresponding to each dimension. The intelligent parsing unit is used to parse unstructured text data through a large language model, extract key knowledge, and assist in building a background knowledge base. The unstructured text data includes historical research reports, rules, and operation and maintenance logs.

[0026] Among them, the data layer can collect hydrological, water quality, ecological, meteorological and human activity data in real time or near real time through multiple sources such as IoT sensors, remote sensing and social big data, and clean, integrate and standardize the data to form a unified spatiotemporal data foundation.

[0027] In the embodiments disclosed in this invention, multi-dimensional spatiotemporal observation data of natural environment big data and urban operation big data of the target area are obtained by relying on the multi-source data open interface. This data covers key elements such as meteorological and hydrological monitoring, topography, human activities, transportation and social development information. On this basis, data cleaning algorithms are used to perform quality control and noise removal on heterogeneous data. At the same time, spatiotemporal registration and scale consistency of multi-source data are achieved based on assimilation algorithms to form a high-precision spatiotemporal profile dataset that can characterize urban water ecological processes. Then, it can be sent to the cloud service platform to establish data usage permissions and supervision mechanisms, providing a unified data foundation for subsequent simulation and evaluation (wherein, the data foundation refers to the high-precision spatiotemporal profile dataset that is updated in real time).

[0028] In the embodiments disclosed in this invention, meteorological and hydrological monitoring includes data such as rainfall from meteorological stations, radar precipitation, and water levels at hydrological stations. Topography can be described using digital elevation models, geological exploration reports, and soil permeability coefficients. Human activities refer to information on human interference with natural water systems, such as water treatment plant operation logs and data from smart water meters in communities. Social development information can be obtained using population heat maps and GDP (Gross Domestic Product). The data is described by the spatial distribution and functional area changes of the product, and finally includes traffic data. The original data sources for each dimension are constructed through meteorological and hydrological monitoring, water quality, ecology, topography, human activities, transportation and travel, and social development. Then, missing values ​​are processed and outliers are identified and corrected for the original heterogeneous data. Data cleaning algorithms combining statistics and deep learning (such as isolated forest and temporal autoencoders) are used to automatically identify and correct abnormal data caused by factors such as sensor drift, cloud cover, and signal loss. Quality control and noise removal are performed on the original data. Finally, spatiotemporal registration and scale consistency of multi-source data are achieved based on data assimilation algorithms, unifying time and space benchmarks and unifying the resolution to a model-driven custom standard (such as 1km×1km grid, 6-hour time step). The final result is a spatiotemporal profile dataset with high integrity (e.g., including location data and time data, with corresponding raster and rainfall values ​​for the corresponding time period during rainfall).

[0029] Based on the integrated data foundation, it can be deployed to an internal database platform. It adopts an incremental update mechanism to update and synchronize the changed data in the data foundation in real time, and sets up multiple backups for the updated data to deal with data loss and other situations, ensuring high availability of data. It also provides data catalog services and standardized application interfaces to support calls from multiple modules (i.e., systems for each dimension), improving the efficiency of data analysis.

[0030] Furthermore, the model layer includes: a matching unit, a preliminary prediction unit, an optimization unit, and a final prediction unit; The matching unit is used to construct natural system models and human system models. It inputs all the spatiotemporal profile datasets into the natural system model and the human system model respectively through a standardized API interface for simulation, and performs pair fitting on the two types of simulation data to obtain key interaction variables. The preliminary prediction unit is used to predict water information of different dimensions based on the key interaction variables and the deep neural agent model, and output multiple initial prediction results, including: initial water volume simulation results, initial multi-source water demand, initial three-dimensional concentration field, initial water velocity field, initial ecological flow, and initial urban water affairs data. An optimization unit is used to calculate the error between the monitoring data and all initial prediction results. Whenever the error exceeds a threshold within a preset time, the weight parameters of the deep neural agent model are adjusted to iteratively optimize the deep neural agent model. The final prediction unit is used to predict water information in different dimensions using the current deep neural agent model, and outputs the current water volume simulation results, current multi-source water demand, current three-dimensional concentration field, current water flow velocity field, current ecological flow and current urban water affairs data.

[0031] The model layer integrates multiple mechanistic models such as hydrology and hydrodynamics, water quality, ecological processes, and socio-economics. It uses all the spatiotemporal profile datasets to perform natural system simulations and human system simulations, obtains key interaction variables, and uses these key interaction variables to train a deep neural proxy model. The deep neural proxy model then approximates complex simulation results within seconds, performs ultra-real-time simulations of water information in different dimensions, and outputs corresponding prediction results.

[0032] To enhance the correlation between different dimensions and improve global analysis capabilities, this invention uses a hydrodynamic-water quality-biogeochemical model as the physical core to construct a natural system module that characterizes processes such as material migration and transformation, and energy cycling in water. Simultaneously, using system dynamics as the basic framework and infrastructure topology network as the core, a human system module is constructed to quantify the impact of human disturbances. Finally, a physically constrained deep neural network is introduced as a core proxy model for cross-scale fusion and computational accelerators to achieve synergistic coupling between physical and human units.

[0033] Based on the current data foundation, a large number of extreme and normal scenarios are run offline in the constructed natural system model to generate a massive number of physical process input-output pairs. Then, the human system model is used to simulate different scenarios of population growth, land use and engineering regulation (such as reservoir and pump station scheduling) offline, thereby quantifying the intensity, spatiotemporal distribution and feedback effects of human disturbances.

[0034] Among them, disturbance intensity refers to the degree of impact of human activities (water intake, drainage, and water transfer) on the original natural water cycle, and there are many indicators that can be quantified; spatiotemporal distribution is the spatiotemporal profile dataset (the data itself has time and space labels), which can be used to make the virtual topological nodes of this pipeline network have real geographical coordinates; feedback effect is a skill inherent in the system dynamics method, which allows the establishment of positive and negative feedback loops, such as a feedback relationship between total population, water pollution, and per capita GDP.

[0035] Furthermore, the matching unit includes: a model building subunit, a first simulation subunit, a data input subunit, a second simulation subunit, a third simulation subunit, a fourth simulation subunit, a data analysis subunit, and a data fusion subunit; The model building subunit is used to build a hydrodynamic model, a water quality model, an ecological process model, and a socio-economic model of the target area, and to build a natural system model of the target area using the hydrodynamic model, the water quality model, the ecological process model, and the socio-economic model. The first simulation subunit is used to input all the spatiotemporal profile datasets into the natural system model through the standardized API interface. The natural system model performs scene simulation based on all the spatiotemporal profile datasets and outputs natural scene data. The data input subunit is used to construct a human system model by using the system dynamics model as the overall framework. Through the standardized API interface, all the spatiotemporal profile datasets are input into the infrastructure and engineering regulation submodule, socioeconomic and spatial development submodule, and dynamic feedback and evolution submodule of the human system model, respectively. The second simulation subunit is used to control the infrastructure and engineering control submodule to simulate the operation and control process of the gray-green-blue infrastructure in the target area based on all the spatiotemporal profile datasets, and output the intensity of human disturbance. The third simulation subunit is used to control the socio-economic and spatial development submodule to simulate the driving impact of social development on the aquatic ecosystem based on all the spatiotemporal profile datasets, and output the spatiotemporal distribution characteristics. The fourth simulation subunit is used to control the dynamic feedback and evolution submodule to simulate the evolution and mutual feedback mechanism of all elements in the target area according to all the spatiotemporal profile datasets, and output dynamic feedback data; The data analysis subunit is used to construct socially driven scenario data based on the intensity of human disturbance, the spatiotemporal distribution characteristics, and the dynamic feedback data. The data fusion subunit is used to pair and fit the natural scene data and the socially driven scene data using a data fusion algorithm to obtain coupled key interaction variables.

[0036] It is important to note that the Infrastructure and Engineering Regulation submodule simulates the operation and regulation processes of gray, green, and blue infrastructure in the target area. Gray infrastructure refers to traditional artificially constructed municipal engineering systems, such as drainage networks, pumping stations, and sewage treatment plants, which are mainly based on engineering design and emphasize solving urban operation problems through engineering technology. Green infrastructure refers to natural or semi-natural ecosystems, such as green spaces, parks, bioretention ponds, permeable pavements, and green roofs, which enhance multiple ecosystem service functions, such as hydrological regulation, biodiversity conservation, and urban heat island mitigation, by simulating natural ecological processes. Blue infrastructure refers to urban water systems, including rivers, lakes, reservoirs, and waterways, which are used for stormwater storage, water resource management, and water ecology maintenance. They play an important buffering role under extreme rainstorms and also provide social and cultural services such as recreational landscapes and public health.

[0037] Furthermore, a human system model is constructed with the infrastructure topology network as the core. The human system model includes an infrastructure and engineering regulation sub-module, a socio-economic and spatial development sub-module, and a dynamic feedback and evolution sub-module.

[0038] Specifically, the Infrastructure and Engineering Regulation submodule, centered on the infrastructure topology network, is primarily used to simulate the operation and regulation processes of urban water physics engineering facilities. This module simulates engineering regulation scenarios such as reservoir and pumping station scheduling based on virtual topology nodes of the pipeline network with real geographic coordinates. Simultaneously, by combining spatiotemporal profile datasets such as water treatment plant operation logs and community smart water meters, it quantifies the disturbance intensity of specific human engineering activities such as water intake, drainage, and water transfer on the natural water cycle.

[0039] The Socioeconomic and Spatial Development submodule is primarily used to simulate the driving impact of macro-social development on aquatic ecosystems. This module takes into account spatiotemporal profile datasets of social development and human activities, such as population heat maps, GDP spatial distribution, functional zone changes, and transportation, to simulate different population growth and land use scenarios, thereby quantifying the spatiotemporal distribution characteristics of human disturbance.

[0040] The Dynamics Feedback and Evolution submodule, based on system dynamics, is primarily used to simulate the long-term evolution and mutual feedback mechanisms of various elements within the system. This module establishes positive and negative feedback loops, quantifies the dynamic feedback relationships between various variables such as "total population - water pollution - GDP per capita," and finally integrates the quantification results from the three submodules to output socially driven scenario data.

[0041] Based on the key interaction variables shared between the natural system model and the human system model (such as urban water demand, water discharge, hydrological status, etc.), a data fusion algorithm is used to pair and integrate the massive input and output data (i.e. natural scene data and socially driven scene data) of the offline simulation to form a unified coupled dataset. Its input includes environmental elements and human disturbances, and its output includes physical state and social response.

[0042] Furthermore, based on the coupled dataset, a deep neural network is trained as a surrogate model, enabling it to learn and internalize the complex nonlinear interaction relationships and feedback mechanisms in natural system models and human system models. The collaborative module, as a computing accelerator and cross-scale fusion unit, is a key core component for exploring the coupling relationship between urban water quality and quantity.

[0043] During model operation, simulation accuracy can be evaluated using multiple metrics such as Nash efficiency coefficient and root mean square error, and the error between monitoring data and coupled model simulation results can be dynamically calculated. A trigger mechanism is set to update the internal weight parameters of the collaborative module DNN (Deep Neural Network) when the error is continuously high or extreme events occur, enabling online correction and continuous optimization of the coupled model, ensuring its stability, adaptability, and predictive ability for extreme hydrological events in long-term operation.

[0044] This invention generates massive amounts of simulation data through natural system models and human system models, and trains a deep neural network proxy model. This model integrates the dual advantages of physical mechanisms and data-driven approaches, and is mainly used to simulate nonlinear response relationships that traditional mechanism models cannot describe, thereby exploring the coupling relationship between urban water quality and quantity.

[0045] Furthermore, the decision-making layer includes: a scheme generation unit, an alarm and suggestion unit, a comprehensive quantification unit, a performance evaluation unit, and an intelligent optimization unit; The scheme generation unit is used to construct a water resource scheduling sub-model, input the current water volume simulation results and the current multi-source water demand into the water resource scheduling sub-model, and output the water resource scheduling scheme for the target area. The alarm and suggestion unit is used to construct a water environment risk early warning sub-model, input the current three-dimensional concentration field and the current water flow velocity field into the water environment risk early warning sub-model, and output multiple risk level signals and emergency suggestions corresponding to each risk level signal; The comprehensive quantitative unit is used to construct a water ecological integrity assessment sub-model. The current ecological flow is input into the water ecological integrity assessment sub-model, and a comprehensive quantitative assessment report of the water ecology is output. The performance evaluation unit is used to construct a water service performance evaluation sub-model, input the current urban water affairs data into the water service performance evaluation sub-model, and output a performance evaluation report; The intelligent optimization unit is used to automatically explore multiple management strategies using reinforcement learning algorithms, and to automatically correlate and analyze environmental distribution, meteorological and hydrological conditions and equipment operating status using knowledge graphs and large language models to obtain the decision basis for each management strategy.

[0046] The decision-making level constructs multiple functional service models based on the model simulation results, evaluates the system response under different planning schemes, climate scenarios or emergencies, inputs all the prediction results into the corresponding functional service models, balances multiple objectives such as flood control, water quality, ecology and landscape through multi-objective optimization algorithms, and predicts and warns of flood risk, water quality exceeding standards, ecological degradation, etc., and outputs corresponding performance evaluation reports.

[0047] Among them, based on the trained surrogate model (i.e. the trained deep neural network model mentioned above), the advanced nature of the water ecosystem in the target area is initially assessed, and a joint scheduling algorithm (such as linear programming, dynamic programming or intelligent optimization algorithm) is used to formulate a zoned water allocation scheme and water source switching rules to generate a water resource scheduling scheme.

[0048] The above sub-models are constructed based on different algorithms or models. The water resource scheduling sub-model is constructed based on the joint scheduling algorithm, the water environment risk early warning model is constructed based on the pollutant migration and transformation numerical model, the water ecological integrity assessment sub-model is constructed based on the habitat suitability model, and the water service efficiency assessment sub-model is constructed based on the efficiency assessment algorithm. In the dimension of water resource allocation, the gridded water volume simulation results and multi-source water demand prediction results output by the proxy model are used as inputs. Joint scheduling algorithms (such as linear programming, dynamic programming, or intelligent optimization algorithms) are employed to formulate zonal water allocation schemes and water source switching rules, generating water resource allocation schemes. In the dimension of water environment risk early warning, the three-dimensional concentration field (objects affecting water quality) and water velocity field output by the proxy model are loaded with a material migration and transformation numerical model. Real-time and historical monitoring data are integrated to simulate pollutant diffusion. The simulated concentration is compared with preset thresholds, matched with the disposal procedures in the emergency plan library, and risk level signals and corresponding emergency response suggestions are generated. In the dimension of water ecological integrity assessment, parameters such as ecological flow and cross-sectional water quality indicators output by the proxy model are used to calculate the habitat suitability index of key aquatic organisms using a habitat suitability model. Graph theory and hydrological connectivity models are used to calculate water system connectivity. Ecosystem service assessment models are used to quantify key ecosystem services, and a comprehensive quantitative assessment report is output.

[0049] In the water service efficiency assessment dimension disclosed in this invention, urban water affairs data and drainage system simulation data in the data sharing platform are used to analyze the synergistic and complementary performance among traditional gray engineering facilities, mainly pipe networks, pumping stations and sewage treatment plants, green ecological facilities, mainly green spaces, parks and bioretention ponds, and blue water spaces, mainly rivers, lakes and reservoirs, using efficiency assessment reports and providing an interactive decision support interface.

[0050] Furthermore, in order to optimize the output performance evaluation report, based on the decision-making level, the intelligent optimization unit can use reinforcement learning algorithms to automatically explore massive management strategies and find the optimal solution for comprehensive benefits among multiple objectives; combined with the reasoning ability of knowledge graphs and large language models, when the system triggers a risk warning, it automatically correlates and analyzes the possible distribution of pollution sources, meteorological and hydrological conditions and equipment operating status, infers the most likely cause chain, and provides interpretable decision-making basis for managers.

[0051] Furthermore, the interaction layer includes: an interactive data input unit, a deduction unit, a comparison unit, a visualization unit, and an intelligent interaction unit; An interactive data input unit is used to construct an interactive model, inputting all the spatiotemporal profile datasets and all the prediction results into the interactive model; The inference unit is used to add multiple sets of parameters input by the user to the interaction model. Under different sets of parameters, the interaction model performs different scenario inferences based on all the spatiotemporal profile datasets and all the prediction results to obtain the corresponding scenario results. The comparison unit is used to compare the differences between the results of each scenario, obtain potential impact information under different parameter sets, and generate corresponding optimization scheduling suggestions based on all the potential impact information. A visualization unit is used to construct a virtual experience environment based on all the optimization scheduling suggestions, all the scheduling-evaluation reports and all the scenario results, and to render the virtual experience environment to obtain visualization simulation results under different scenarios; The intelligent interaction unit integrates a large language model as an intelligent interaction hub and automatically generates analysis reports containing data interpretation, chart explanations, and management suggestions based on simulation results. The intelligent interaction hub allows users to directly interact with the simulator in natural language to query simulation results.

[0052] The interactive layer uses GIS and a game engine to perform 3D dynamic rendering of all the spatiotemporal profile datasets and all the prediction results. It constructs a digital twin virtual experience environment based on all the scheduling-evaluation reports and performs different simulations and comparisons of the virtual experience environment. It allows managers to modify parameters or plans interactively and view the visualized simulation results under different scenarios in real time. It also integrates a large language model to support natural language interactive queries and automatic generation of analysis reports.

[0053] Based on the above analysis, a comprehensive water ecosystem simulator can be further constructed to expand its application in various dimensions of urban governance. A comprehensive regulatory framework is built from five aspects: systems engineering, risk response, ecological protection, service improvement, and public participation. This establishes a collaborative regulatory system covering five dimensions: water resource security and regulation, water environment risk prevention and control, water ecological integrity assessment, water service efficiency improvement, and water culture dissemination. A digital twin platform for urban water systems is also developed. Through multi-scenario simulations of the long-term evolution of water systems, professional decision-making is supported. Simultaneously, popular science display functions are integrated, and interactive technologies such as VR (Virtual Reality) are combined to form a three-dimensional, visualized dissemination mechanism, enhancing public understanding and participation in urban water systems and achieving a smart integration of decision support and public education.

[0054] Furthermore, by integrating a large language model into the intelligent interaction unit as an intelligent interaction hub, users can directly communicate with the simulator in natural language to query simulation results, and automatically generate analysis reports containing data interpretation, chart explanations, and management suggestions based on the simulation results.

[0055] Specifically, based on built-in decision analysis and optimization scheduling algorithms, it automatically generates multi-dimensional decision-making schemes, including water resource scheduling schedules, emergency plans, risk warning reports, and ecological environment assessment documents. It also embeds sensitivity analysis and uncertainty explanations into the results, supports multi-format interface output, ensures the scientific nature and transparency of decision-making schemes, and achieves the dual application value of providing standardized data for management departments and visual materials for public science popularization platforms.

[0056] Furthermore, a platform layer can be added to the simulator to provide model and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. Through data exchange with external systems such as the City Information Modeling (CIM) platform and the urban water management system, the stable operation, open expansion, and efficient computing of the above layers can be ensured.

[0057] The platform layer includes: a model library and knowledge base management unit, a standardized interface unit, and a computing support unit; The model library and knowledge base management unit is used to uniformly manage the versions, parameter configurations, and expert rules of all models in the simulator; A standardized interface unit is used to provide a standardized API interface for data and function interaction with all the aforementioned external systems; The computing support unit is used to support the rapid operation of all the aforementioned models and the efficient simulation of massive scenarios through elastic cloud computing resources and parallel computing capabilities.

[0058] Specifically, the model library and knowledge base management unit is used to uniformly manage the versions, parameter configurations, and expert rules of all models in the simulator, supporting dynamic updates and version rollback of models to ensure the traceability and maintainability of the model system; the standardized interface unit is used to provide standardized API interfaces for data and function interaction with external systems (including the City Information Modeling (CIM) platform, the urban water dispatching system, and the emergency management platform, etc.) to achieve cross-system data sharing and business collaboration; the computing support unit is used to support the rapid operation of large-scale complex models and the efficient simulation of massive scenarios through elastic cloud computing resources and high-performance parallel computing capabilities, ensuring the computing performance requirements of the system in extreme event response and real-time dispatch decision-making scenarios.

[0059] Urban water ecosystem complexes are comprehensive urban infrastructure projects that treat urban gray, green, and blue infrastructures as an organic whole, collaboratively addressing water resources, water environment, water ecology, and water security issues. Their highly complex, dynamic, and multi-objective coupled system characteristics necessitate reliance on digital simulators as a core decision-making tool. These simulators construct a "digital twin" of the urban water ecosystem complex, integrating multi-dimensional data and models from hydrology, water quality, ecology, and society to achieve accurate simulation, evaluation, and prediction of the system's comprehensive response under different planning schemes, management strategies, or extreme scenarios. The data layer utilizes multiple sources, including IoT sensors, remote sensing, and social big data, to collect, clean, and preprocess raw data in real time, forming a unified spatiotemporal data foundation. The model layer integrates multiple mechanistic models related to hydrology, hydrodynamics, water quality, ecological processes, and socioeconomics, and introduces a deep neural proxy model to achieve ultra-real-time simulation and rapid extrapolation of numerous scenarios. The decision-making layer constructs a functional service model through multi-objective optimization algorithms and intelligent early warning mechanisms, outputting scheduling-evaluation reports. The interaction layer dynamically renders datasets and prediction results, constructing a digital twin virtual experience environment to obtain visualized simulation results. The platform layer provides model library management, standardized API interfaces, and high-performance computing support. This invention integrates multi-source data to achieve accurate simulation of urban water ecosystem complexes at extremely low cost and risk, optimizing design before construction and enabling intelligent scheduling and risk early warning during operation, significantly improving the refined management and scientific decision-making capabilities of water ecosystems.

[0060] Furthermore, based on Figure 1 The simulator of the urban water ecological complex shown in the present invention, and the simulation operation method of the urban water ecological complex according to a preferred embodiment of the present invention, are as follows: Figure 2 As shown, the simulation operation method of the urban water ecological complex includes the following steps: Step S10: The data layer obtains original data sources of multiple dimensions of the target area through multi-source means, performs real-time monitoring on each of the original data sources and then performs preprocessing to obtain the spatiotemporal profile dataset corresponding to each dimension.

[0061] Furthermore, all the spatiotemporal profile datasets are deployed to an internal data platform. For each spatiotemporal profile dataset, privacy protection processing is performed to remove sensitive information from the spatiotemporal profile dataset, thereby achieving high-precision desensitization of the spatiotemporal profile dataset. Construct a data map, store the spatiotemporal profile dataset updated each time in the data map, and add a data index for the spatiotemporal profile dataset updated each time. Based on all the data indexes, each spatiotemporal profile dataset is encapsulated, and a corresponding standardized API interface is constructed.

[0062] In the embodiments disclosed in this invention, meteorological and hydrological monitoring includes data such as rainfall from meteorological stations, radar precipitation, and water levels at hydrological stations. Topography can be described using digital elevation models, geological exploration reports, and soil permeability coefficients. Human activities, i.e., information on human society's interference with natural water systems, include data such as water treatment plant operation logs and community smart water meters. Social development information can be described using population heat maps, GDP spatial distribution, and functional zone changes. Finally, traffic data is included. These five types of data together construct a five-dimensional original data source. Then, missing values ​​are processed, outliers are identified and corrected, and statistical methods are employed to address the issues. Data cleaning algorithms combined with deep learning (such as isolated forests and temporal autoencoders) automatically identify and correct abnormal data caused by factors such as sensor drift, cloud cover, and signal loss, and perform quality control and noise removal on the original data. Finally, based on data assimilation algorithms, spatiotemporal registration and scale consistency of multi-source data are achieved, unifying time and spatial benchmarks, and unifying the resolution to a model-driven custom standard (such as a 1km×1km grid and a 6-hour time step), ultimately forming a highly complete spatiotemporal profile dataset (e.g., including location data and time data, with corresponding grids and rainfall values ​​for the corresponding time period during rainfall).

[0063] Step S20: The model layer uses all the spatiotemporal profile datasets to perform natural system simulation and human system simulation to obtain key interaction variables. The deep neural agent model is trained using the key interaction variables, and the deep neural agent model is used to predict water information in different dimensions. The simulation results are gradually approximated within seconds, and the corresponding prediction results are output.

[0064] The key interactive variables include the physical state and social response of urban water quality in the target area; specifically, they include environmental elements and physical states that characterize natural processes in the target area, as well as human disturbances and social responses that characterize human activities, including urban water demand, sewage discharge, and human activity data.

[0065] Step S30: The decision-making layer constructs multiple functional service models, uses all the functional service models to evaluate the simulator response under different planning schemes, climate scenarios or emergencies, and inputs all the prediction results into the corresponding functional service models respectively. It uses a multi-objective optimization algorithm to balance multiple objectives and generates and outputs the corresponding scheduling-evaluation report.

[0066] Step S40: The interaction layer performs three-dimensional dynamic rendering of all the spatiotemporal profile datasets and all the prediction results based on GIS and game engine, constructs a digital twin virtual experience environment according to all the scheduling-evaluation reports, and performs different simulations and comparisons on the virtual experience environment to obtain multiple visual simulation results.

[0067] Step S50: The platform layer provides model library and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. It transmits all the visualization simulation results through data exchange with multiple external systems, including the City Information Modeling (CIM) platform and the City Water Dispatch System.

[0068] Furthermore, such as Figure 3 As shown, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 3 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0069] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard drive or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard drive, smart media card (SMC), secure digital card (SD), flash card, etc. Furthermore, the memory 20 may include both internal and external storage units. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output.

[0070] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data.

[0071] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components of the terminal communicate with each other via a system bus.

[0072] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a simulator program for an urban water ecological complex, and when the simulator program for an urban water ecological complex is executed by a processor, it implements the steps of the simulator operation method for an urban water ecological complex as described above.

[0073] This invention integrates multi-source data to achieve accurate simulation of urban water ecosystems, thereby improving the accuracy of water quality information prediction.

[0074] This invention provides a simulator and method for an urban water ecosystem complex. The simulator comprises a data layer, a model layer, a decision layer, an interaction layer, and a platform layer. The data layer is used to monitor in real time the original data sources of a target area across multiple dimensions obtained through multi-source methods. Each original data source is monitored in real time and then preprocessed to obtain a spatiotemporal profile dataset corresponding to each dimension. The model layer is used to perform natural system simulation and human system simulation using all the spatiotemporal profile datasets to obtain key interaction variables. These key interaction variables are used to train a deep neural proxy model, which then predicts water information in different dimensions, gradually approximating the simulation results within seconds and outputting the corresponding prediction results. The decision layer is used to construct multiple functional service models and evaluate the results using all the functional service models. The system estimates simulator responses under different planning schemes, climate scenarios, or emergencies, and inputs all predicted results into the corresponding functional service models. A multi-objective optimization algorithm is used to balance multiple objectives, generating and outputting corresponding scheduling-evaluation reports. An interaction layer is used to perform 3D dynamic rendering of all spatiotemporal profile datasets and prediction results based on GIS and a game engine. A digital twin virtual experience environment is constructed based on all scheduling-evaluation reports, and the virtual experience environment is simulated and compared in different ways to obtain multiple visualized simulation results. A platform layer provides model and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. It transmits all visualized simulation results through data exchange with multiple external systems, including a City Information Modeling (CIM) platform and an urban water management scheduling system. This invention integrates multi-source data to achieve accurate simulation of urban water ecosystems, improving the accuracy of water quality information prediction.

[0075] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0076] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A simulator for an urban water ecological complex, characterized in that, The simulator for the urban water ecosystem complex includes: a data layer, a model layer, a decision-making layer, an interaction layer, and a platform layer; The data layer is used to obtain original data sources of multiple dimensions of the target area through multi-source means, and to perform real-time monitoring and preprocessing on each of the original data sources to obtain the spatiotemporal profile dataset corresponding to each dimension. The model layer is used to perform natural system simulation and human system simulation using all the spatiotemporal profile datasets to obtain key interaction variables. The deep neural agent model is trained using the key interaction variables and the deep neural agent model is used to predict water information in different dimensions. The simulation results are gradually approximated within seconds, and the corresponding prediction results are output. The model layer includes: a matching unit, a preliminary prediction unit, an optimization unit, and a final prediction unit; The matching unit includes: a model building subunit, a first simulation subunit, a data input subunit, a second simulation subunit, a third simulation subunit, a fourth simulation subunit, a data analysis subunit, and a data fusion subunit; The model building subunit is used to build a hydrodynamic model, a water quality model, an ecological process model, and a socio-economic model of the target area, and to build a natural system model of the target area using the hydrodynamic model, the water quality model, the ecological process model, and the socio-economic model. The first simulation subunit is used to input all the spatiotemporal profile datasets into the natural system model through a standardized API interface. The natural system model performs scene simulation based on all the spatiotemporal profile datasets and outputs natural scene data. The data input subunit is used to construct a human system model by using the system dynamics model as the overall framework. Through the standardized API interface, all the spatiotemporal profile datasets are input into the infrastructure and engineering regulation submodule, socioeconomic and spatial development submodule, and dynamic feedback and evolution submodule of the human system model, respectively. The second simulation subunit is used to control the infrastructure and engineering control submodule to simulate the operation and control process of the gray-green-blue infrastructure in the target area based on all the spatiotemporal profile datasets, and output the intensity of human disturbance. The third simulation subunit is used to control the socio-economic and spatial development submodule to simulate the driving impact of social development on the aquatic ecosystem based on all the spatiotemporal profile datasets, and output the spatiotemporal distribution characteristics. The fourth simulation subunit is used to control the dynamic feedback and evolution submodule to simulate the evolution and mutual feedback mechanism of all elements in the target area according to all the spatiotemporal profile datasets, and output dynamic feedback data; The data analysis subunit is used to construct socially driven scenario data based on the intensity of human disturbance, the spatiotemporal distribution characteristics, and the dynamic feedback data. The data fusion subunit is used to pair and fit the natural scene data and the socially driven scene data using a data fusion algorithm to obtain coupled key interaction variables. The decision-making layer is used to construct multiple functional service models, evaluate the simulator response under different planning schemes, climate scenarios or emergencies using all the functional service models, and input all the prediction results into the corresponding functional service models respectively. It uses a multi-objective optimization algorithm to balance multiple objectives and generate and output the corresponding scheduling-evaluation report. The interaction layer is used to perform three-dimensional dynamic rendering of all the spatiotemporal profile datasets and all the prediction results based on GIS and game engine, construct a digital twin virtual experience environment according to all the scheduling-evaluation reports, and perform different inferences and comparisons on the virtual experience environment to obtain multiple visual simulation results. The platform layer provides model and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. It transmits all the visualization simulation results through data exchange with multiple external systems, including the City Information Modeling (CIM) platform and the City Water Dispatch System.

2. The simulator for an urban water ecological complex according to claim 1, characterized in that, The data layer includes: a data acquisition unit, a data preprocessing unit, a data unification unit, and an intelligent parsing unit; The data acquisition unit is used to acquire original data sources of the target area in the dimensions of meteorological and hydrological monitoring, water quality, ecology, topography, human activities, transportation and travel and social development through multi-source means, and to monitor each of the original data sources in real time to obtain change data of each of the original data sources. The data preprocessing unit is used to correct the changed data, missing values ​​and outliers of each of the original data sources, use data cleaning algorithms to perform quality control and noise removal on the corrected original data sources, and then perform standardization processing to obtain the corresponding corrected data sources. The data unification unit is used to use a data assimilation algorithm to unify the resolution of all the corrected data sources into a preset grid in the time and space dimensions, so as to obtain the spatiotemporal profile dataset corresponding to each dimension. The intelligent parsing unit is used to parse unstructured text data through a large language model, extract key knowledge, and assist in building a background knowledge base. The unstructured text data includes historical research reports, rules, and operation and maintenance logs.

3. The simulator for an urban water ecological complex according to claim 1, characterized in that, The matching unit is used to construct natural system models and human system models. It inputs all the spatiotemporal profile datasets into the natural system model and the human system model respectively through a standardized API interface for simulation, and performs pair fitting on the two types of simulation data to obtain key interaction variables. The preliminary prediction unit is used to predict water information of different dimensions based on the key interaction variables and the deep neural agent model, and output multiple initial prediction results, including: initial water volume simulation results, initial multi-source water demand, initial three-dimensional concentration field, initial water velocity field, initial ecological flow, and initial urban water affairs data. An optimization unit is used to calculate the error between the monitoring data and all initial prediction results. Whenever the error exceeds a threshold within a preset time, the weight parameters of the deep neural agent model are adjusted to iteratively optimize the deep neural agent model. The final prediction unit is used to predict water information in different dimensions using the current deep neural agent model, and outputs the current water volume simulation results, current multi-source water demand, current three-dimensional concentration field, current water flow velocity field, current ecological flow and current urban water affairs data.

4. The simulator for an urban water ecological complex according to claim 3, characterized in that, The decision-making layer includes: a scheme generation unit, an alarm and suggestion unit, a comprehensive quantification unit, a performance evaluation unit, and an intelligent optimization unit; The scheme generation unit is used to construct a water resource scheduling sub-model, input the current water volume simulation results and the current multi-source water demand into the water resource scheduling sub-model, and output the water resource scheduling scheme for the target area. The alarm and suggestion unit is used to construct a water environment risk early warning sub-model, input the current three-dimensional concentration field and the current water flow velocity field into the water environment risk early warning sub-model, and output multiple risk level signals and emergency suggestions corresponding to each risk level signal; The comprehensive quantitative unit is used to construct a water ecological integrity assessment sub-model. The current ecological flow is input into the water ecological integrity assessment sub-model, and a comprehensive quantitative assessment report of the water ecology is output. The performance evaluation unit is used to construct a water service performance evaluation sub-model, input the current urban water affairs data into the water service performance evaluation sub-model, and output a performance evaluation report; The intelligent optimization unit is used to automatically explore multiple management strategies using reinforcement learning algorithms, and to automatically correlate and analyze environmental distribution, meteorological and hydrological conditions and equipment operating status using knowledge graphs and large language models to obtain the decision basis for each management strategy.

5. The simulator for an urban water ecological complex according to claim 1, characterized in that, The interaction layer includes: an interactive data input unit, a deduction unit, a comparison unit, a visualization unit, and an intelligent interaction unit; An interactive data input unit is used to construct an interactive model, inputting all the spatiotemporal profile datasets and all the prediction results into the interactive model; The inference unit is used to add multiple sets of parameters input by the user to the interaction model. Under different sets of parameters, the interaction model performs different scenario inferences based on all the spatiotemporal profile datasets and all the prediction results to obtain the corresponding scenario results. The comparison unit is used to compare the differences between the results of each scenario, obtain potential impact information under different parameter sets, and generate corresponding optimization scheduling suggestions based on all the potential impact information. A visualization unit is used to construct a virtual experience environment based on all the optimization scheduling suggestions, all the scheduling-evaluation reports and all the scenario results, and to render the virtual experience environment to obtain visualization simulation results under different scenarios; The intelligent interaction unit integrates a large language model as an intelligent interaction hub and automatically generates analysis reports containing data interpretation, chart explanations, and management suggestions based on simulation results. The intelligent interaction hub allows users to directly interact with the simulator in natural language to query simulation results.

6. The simulator for an urban water ecological complex according to claim 1, characterized in that, The platform layer includes: a model library and knowledge base management unit, a standardized interface unit, and a computing support unit; The model library and knowledge base management unit is used to uniformly manage the versions, parameter configurations, and expert rules of all models in the simulator; A standardized interface unit is used to provide a standardized API interface for data and function interaction with all the aforementioned external systems; The computing support unit is used to support the rapid operation of all the aforementioned models and the efficient simulation of massive scenarios through elastic cloud computing resources and parallel computing capabilities.

7. A simulator operation method for an urban water ecosystem complex as described in any one of claims 1-6, characterized in that, The simulator operation method for the urban water ecological complex includes: The data layer obtains original data sources of multiple dimensions of the target region through multi-source methods, performs real-time monitoring on each of the original data sources and then performs preprocessing to obtain the spatiotemporal profile dataset corresponding to each dimension. The model layer uses all the spatiotemporal profile datasets to perform natural system simulation and human system simulation, obtains key interaction variables, uses the key interaction variables to train the deep neural agent model, and uses the deep neural agent model to predict water information in different dimensions, gradually approximating the simulation results in seconds, and outputting the corresponding prediction results. The decision-making level constructs multiple functional service models, uses all the functional service models to evaluate the simulator response under different planning schemes, climate scenarios or emergencies, and inputs all the prediction results into the corresponding functional service models respectively. It uses a multi-objective optimization algorithm to balance multiple objectives and generates and outputs the corresponding scheduling-evaluation report. The interaction layer performs three-dimensional dynamic rendering of all the spatiotemporal profile datasets and all the prediction results based on GIS and game engine, constructs a digital twin virtual experience environment based on all the scheduling-evaluation reports, and performs different simulations and comparisons on the virtual experience environment to obtain multiple visual simulation results. The platform layer provides model and knowledge base management, standardized API interfaces, and cloud computing and high-performance computing support. It transmits all the visualization simulation results through data exchange with multiple external systems, including the City Information Modeling (CIM) platform and the City Water Dispatch System.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a simulator program for an urban water ecological complex, which, when executed by a processor, implements the steps of the simulator operation method for an urban water ecological complex as described in claim 7.

9. A terminal, characterized in that, The terminal includes a memory, a processor, and a display. When the processor executes the simulator running program for the urban water ecological complex, it implements the steps of the simulator running method for the urban water ecological complex as described in claim 7.

Images