A multi-dimensional dynamic evaluation and path optimization method for carbon neutrality in ecological parks

Abstract

The present invention relates to the field of carbon neutrality technology, and specifically to a multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method; the present invention collects heat load data in real time through a sensor network, constructs a heat energy flow model, analyzes pipeline efficiency fluctuations and locates energy efficiency loss points. In response to the loss points, the present invention uses thermodynamic simulation and reinforcement learning algorithms to optimize the pipeline layout and generate new configuration schemes. Through iterative optimization and real-time adjustment, the present invention can sustainably improve the system operation status, reduce carbon emissions and operating costs. At the same time, the present invention can also predict future heat load distribution trends and provide a basis for optimization of the next cycle. This method realizes the intelligent management of the park's thermal energy network, improves energy utilization efficiency, and provides an effective solution for achieving low-carbon operations.

Description

Technical Field

[0001] The present invention belongs to the field of carbon neutrality technology, and specifically provides a multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method. Background Art

[0002] Carbon neutrality is the core of responding to global climate change. As an important carrier of sustainable urban development, the low-carbon transformation of the energy system of the ecological park is particularly critical. Achieving carbon neutrality in the park requires comprehensive consideration of the optimization of the entire chain of energy supply, transmission and consumption, and dynamic evaluation and path optimization methods provide systematic technical support for achieving this goal. However, existing methods have significant limitations in dealing with the complexity of the park's energy system. They mostly stay at the static analysis level, making it difficult to adapt to the dynamic changes in heat load and the needs of multi-energy coordination, and lack accurate capture of real-time energy efficiency losses. These limitations make it difficult to optimize the park's heating system to balance low-carbon goals and economic efficiency.

[0003] The complexity of the eco-park's heating system stems from the spatiotemporal heterogeneity of heat load distribution, a characteristic that makes accurate modeling of heat flow difficult. This uneven distribution of heat loads further exacerbates fluctuations in the network's transmission efficiency, making it difficult to identify heat loss points in real time. Fluctuations in transmission efficiency directly impact the system's overall energy efficiency, leading to reduced precision in carbon emissions control. Existing thermal network models are mostly based on traditional thermodynamic analysis and lack deep integration with digital twin technology, making it impossible to dynamically monitor heat flow and optimize decision-making.

[0004] Therefore, how to construct a heating flow analysis method that can dynamically capture the heat load distribution characteristics, monitor the flow and loss of heat energy in real time, and optimize the pipeline layout has become a key issue in optimizing the carbon neutrality path of ecological parks. Summary of the Invention

[0005] The purpose of the present invention is to address the above-mentioned deficiencies in the prior art and provide a multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method.

[0006] The purpose of the present invention is achieved through the following technical solutions: A multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method, comprising the following steps: obtaining real-time data on heat loads in various areas of the park, collecting temperature, flow and pressure information through a sensor network, and determining the spatiotemporal heterogeneity characteristics of heat load distribution in combination with time series analysis; constructing a thermal energy flow model based on digital twins according to the spatiotemporal heterogeneity characteristics of heat load distribution, and using a graph neural network to simulate the dynamic transmission path of heat energy in the pipeline network to obtain the real-time state of heat energy flow; extracting the temperature gradient and flow rate changes of each node in the pipeline network from the real-time state of heat energy flow, calculating the transmission efficiency fluctuation index, judging whether there are nodes whose efficiency fluctuations exceed a preset threshold, and determining the energy efficiency loss point; for the energy efficiency loss point, obtaining the geometric parameters and material properties of the corresponding pipeline segment, analyzing the distribution law of heat energy loss through thermodynamic simulation, and obtaining high-precision spatial positioning of the loss point; according to High-precision spatial positioning of loss points, using reinforcement learning algorithms to optimize pipeline layout, adjust pipe diameters and valve openings, and generate optimized pipeline configuration plans; through the optimized pipeline configuration plan, update the pipeline parameters in the digital twin model, re-simulate the heat flow state, calculate carbon emissions and operating costs, and determine whether low-carbon goals and economic requirements are met; if carbon emissions or operating costs exceed the preset threshold, extract new heat load distribution data from the digital twin model, iteratively optimize the pipeline layout, and generate an updated configuration plan; according to the updated configuration plan, adjust the valve and pump station parameters in the pipeline control system in real time, obtain the adjusted heat flow data, and verify whether the system operation status achieves the expected optimization effect; through the verified heat flow data, update the heat load distribution prediction model, use the time series prediction algorithm to generate the heat load distribution trend for the next 24 hours, and determine the optimization starting point for the next cycle.

[0007] The present invention is further configured as follows: the spatiotemporal heterogeneity characteristics of the heat load distribution include the temporal periodicity and spatial aggregation of the heat load; the real-time status of the heat energy flow includes the temperature field distribution and dynamic changes of the flow rate at the pipeline nodes; the high-precision spatial positioning of the energy efficiency loss points is specifically the local heat loss distribution and loss hotspot area identification of the pipeline segment; the optimized pipeline configuration plan includes the pipeline diameter adjustment plan, the valve opening optimization strategy and the pump station power adjustment plan; the evaluation results of the carbon emissions and operating costs include the carbon emission reduction potential and the operating cost saving rate; the updated configuration plan includes the pipeline topology adjustment and control parameter optimization; the heat load distribution prediction model includes the short-term heat load trend prediction and the long-term heat demand pattern analysis.

[0008] The present invention is further configured to obtain real-time data on heat loads in various areas within the park, collect temperature, flow, and pressure information through a sensor network, and determine the spatiotemporal heterogeneity characteristics of the heat load distribution in combination with time series analysis. The specific steps are as follows: based on the sensor network arranged in various areas within the park, real-time data on temperature, flow, and pressure are collected, and time synchronization and spatial calibration are performed on the multi-sensor data to generate a unified heat load data set; based on the heat load data set, a time series decomposition algorithm is used to separate trend items, period items, and random items of the heat load data to reveal the temporal periodic characteristics of the heat load; based on the temporal periodic characteristics, a cluster analysis method is used to divide the spatial distribution of the heat load into multiple clustered areas to generate the spatial clustering characteristics of the heat load; and by integrating the temporal periodic characteristics and the spatial clustering characteristics, a spatiotemporal heterogeneity model of the heat load distribution is generated to describe the spatiotemporal heterogeneity characteristics of the heat load distribution.

[0009] The present invention is further configured to construct a thermal energy flow model based on digital twins according to the spatiotemporal heterogeneity characteristics of the heat load distribution, and use a graph neural network to simulate the dynamic transmission path of thermal energy in the pipeline network to obtain the real-time state of thermal energy flow. The specific steps are: based on the spatiotemporal heterogeneity characteristics of the heat load distribution, construct a digital twin model of thermal energy flow, and associate the physical pipeline network with the digital model through virtual mapping technology; based on the digital twin model, use a graph neural network algorithm to abstract the pipeline network nodes and connection relationships into a graph structure to simulate the dynamic transmission path of thermal energy in the pipeline network; based on the dynamic transmission path of thermal energy, extract the temperature field distribution and flow dynamic change data of the pipeline network nodes to generate the real-time state of thermal energy flow; combined with the real-time state of thermal energy flow, verify the accuracy of the model through the thermodynamic equilibrium equation to generate a real-time state description of thermal energy flow.

[0010] The present invention is further configured to extract the temperature gradient and flow rate change of each node in the pipeline network from the real-time state of thermal energy flow, calculate the transmission efficiency fluctuation index, and judge whether there is a node whose efficiency fluctuation exceeds a preset threshold. The specific steps of determining the energy efficiency loss point are: based on the real-time state of thermal energy flow, extract the temperature gradient and flow rate change data of each node in the pipeline network, and generate a thermal characteristic description of the pipeline network node; based on the thermal characteristic description, calculate the transmission efficiency fluctuation index of each node in the pipeline network, and evaluate the stability of thermal energy transmission; based on the transmission efficiency fluctuation index, set a fluctuation threshold, judge whether there is a node whose efficiency fluctuation exceeds the preset threshold, and generate a candidate set of energy efficiency loss points; based on the candidate set of energy efficiency loss points, verify the accuracy of the loss point through thermodynamic simulation, and generate high-precision spatial positioning of the energy efficiency loss point.

[0011] The present invention is further configured to obtain the geometric parameters and material properties of the corresponding pipeline segment for the energy efficiency loss point, analyze the distribution law of heat energy loss through thermodynamic simulation, and obtain the high-precision spatial positioning of the loss point. The specific steps are: based on the high-precision spatial positioning of the energy efficiency loss point, obtain the geometric parameters and material properties of the corresponding pipeline segment, and generate a basic physical property description of the pipeline segment; based on the basic physical property description, use thermodynamic simulation technology to analyze the distribution law of heat energy loss in the pipeline segment, and generate a spatial distribution map of heat energy loss; based on the spatial distribution map of heat energy loss, identify the local heat loss distribution through heat flux density analysis, and generate a loss hotspot area; comprehensively integrate the local heat loss distribution and the loss hotspot area to generate a high-precision spatial positioning description of the loss point.

[0012] The present invention is further configured to optimize the pipeline layout based on the high-precision spatial positioning of the loss point, adjust the pipe diameter and valve opening, and generate an optimized pipeline configuration plan. The specific steps are: based on the high-precision spatial positioning of the loss point, construct a reinforcement learning model for pipeline optimization, and explore the optimal solution through the interaction between the intelligent agent and the environment; based on the reinforcement learning model, adjust the pipe diameter and valve opening of the pipeline section, optimize the heat energy transmission path, and generate a preliminary optimization plan; based on the preliminary optimization plan, verify the optimization effect through thermodynamic simulation, and generate an optimized pipeline configuration plan; combined with the optimized pipeline configuration plan, update the pipeline parameters in the digital twin model, and generate an optimized heat energy flow state description.

[0013] The present invention is further configured to update the pipeline parameters in the digital twin model through the optimized pipeline configuration plan, re-simulate the heat flow state, calculate the carbon emissions and operating costs, and judge whether the low-carbon goals and economic requirements are met. The specific steps are: based on the optimized pipeline configuration plan, update the pipeline parameters in the digital twin model, and re-simulate the heat flow state; based on the heat flow state, calculate the optimized carbon emissions and operating costs, and generate low-carbon goals and economic evaluation results; based on the evaluation results, judge whether the low-carbon goals and economic requirements are met, and if not, generate new optimization requirements; based on the new optimization requirements, extract heat load distribution data, and generate a new round of optimization starting point.

[0014] The present invention is further configured to adjust the valve and pump station parameters in the pipeline control system in real time according to the updated configuration scheme, obtain the adjusted thermal energy flow data, and verify whether the system operation status has achieved the expected optimization effect. The specific steps are: based on the updated configuration scheme, adjust the valve opening and pump station power parameters in the pipeline control system in real time; based on the adjusted parameters, collect thermal energy flow data, and generate an adjusted thermal energy flow state description; based on the adjusted thermal energy flow state description, verify whether the system operation status has achieved the expected optimization effect by comparing with the optimization target; based on the verification results, generate a system optimization effect report.

[0015] The present invention is further configured to update the prediction model of heat load distribution through verified heat energy flow data, and use a time series prediction algorithm to generate a heat load distribution trend for the next 24 hours. The steps of determining the optimization starting point of the next cycle are specifically as follows: based on the verified heat energy flow data, update the prediction model of heat load distribution; based on the prediction model, use a time series prediction algorithm to generate a heat load distribution trend for the next 24 hours; based on the heat load distribution trend, combined with the current optimization effect, determine the optimization starting point of the next cycle; based on the optimization starting point, generate an optimization task description for the next cycle.

[0016] Beneficial effects of the present invention: The present invention collects heat load data in real time through a sensor network, constructs a heat flow model, analyzes fluctuations in pipeline network efficiency, and locates energy efficiency loss points. In response to loss points, the present invention uses thermodynamic simulation and reinforcement learning algorithms to optimize the pipeline network layout and generate new configuration schemes. Through iterative optimization and real-time adjustment, the present invention can sustainably improve the system operating status, reduce carbon emissions and operating costs. At the same time, the present invention can also predict future heat load distribution trends and provide a basis for optimization of the next cycle. This method realizes the intelligent management of the park's thermal energy network, improves energy utilization efficiency, and provides an effective solution for achieving low-carbon operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The invention is further described with reference to the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the invention. A person skilled in the art can obtain other drawings based on the following drawings without making any creative effort.

[0018] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0019] The present invention is further described with reference to the following examples.

[0020] Depend on Figure 1It can be seen that the embodiment of the present invention provides a multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method, including the following steps: obtaining real-time data of heat load in each area of ​​the park, collecting temperature, flow and pressure information through a sensor network, and combining time series analysis to determine the spatiotemporal heterogeneity characteristics of heat load distribution; constructing a thermal energy flow model based on digital twins according to the spatiotemporal heterogeneity characteristics of heat load distribution, and using graph neural networks to simulate the dynamic transmission path of heat energy in the pipeline network to obtain the real-time state of heat energy flow; extracting the temperature gradient and flow change of each node in the pipeline network from the real-time state of heat energy flow, calculating the transmission efficiency fluctuation index, judging whether there are nodes whose efficiency fluctuation exceeds a preset threshold, and determining the energy efficiency loss point; for the energy efficiency loss point, obtaining the geometric parameters and material properties of the corresponding pipeline segment, analyzing the distribution law of heat energy loss through thermodynamic simulation, and obtaining high-precision spatial positioning of the loss point; according to the loss High-precision spatial positioning of the point, using reinforcement learning algorithm to optimize the pipeline layout, adjust the pipe diameter and valve opening, and generate an optimized pipeline configuration plan; through the optimized pipeline configuration plan, update the pipeline parameters in the digital twin model, re-simulate the heat flow state, calculate the carbon emissions and operating costs, and judge whether it meets the low-carbon goals and economic requirements; if the carbon emissions or operating costs exceed the preset threshold, extract new heat load distribution data from the digital twin model, iteratively optimize the pipeline layout, and generate an updated configuration plan; according to the updated configuration plan, adjust the valve and pump station parameters in the pipeline control system in real time, obtain the adjusted heat flow data, and verify whether the system operation status achieves the expected optimization effect; through the verified heat flow data, update the heat load distribution prediction model, use the time series prediction algorithm to generate the heat load distribution trend for the next 24 hours, and determine the optimization starting point for the next cycle. This embodiment achieves accurate capture of the spatiotemporal heterogeneity of heat load distribution and improves the modeling accuracy of heat flow state through the application of sensor network and time series analysis. This method uses digital twin technology and graph neural networks to effectively simulate the dynamic transmission path of thermal energy, which helps to identify energy efficiency loss points and their distribution patterns. In addition, the application of reinforcement learning algorithms makes the optimization of pipeline layout more scientific, providing a solid foundation for balancing low-carbon goals and economic requirements. By combining thermodynamic simulation and time series prediction, this method can deeply analyze the interaction between heat energy loss and heat load distribution, and accurately predict future heat load trends. By using life cycle assessment and cost-benefit analysis, it provides comprehensive technical support for the optimization of the carbon neutrality path of the ecological park, thereby promoting the low-carbon transformation and sustainable development of the park's heating system.

[0021] During the implementation process, real-time heat load data for each area within the park must be obtained. This process is accomplished through a sensor network deployed throughout the park. The sensor network collects real-time data on temperature, flow, and pressure. Time synchronization and spatial calibration are performed on the multi-sensor data to generate a unified heat load dataset. Time synchronization utilizes GPS-based timestamp alignment technology to ensure a consistent time base for all sensor data. Spatial calibration uses a geographic information system to accurately calibrate sensor locations, eliminating data bias caused by installation errors. Subsequently, a time series decomposition algorithm is used to analyze the heat load dataset, separating the trend term, periodic term, and random term. This can be expressed as: X(t) = T(t) + C(t) + R(t), where X(t) is the raw heat load data, T(t) is the trend term, C(t) is the periodic term, and R(t) is the random term. The trend term is extracted using a moving average method, the periodic term uses a fast Fourier transform (FFT) to identify the primary frequency components, and the random term is the residual component. Based on the temporal periodicity, the K-means clustering analysis method is used to divide the spatial distribution of the heat load into multiple clustered areas, generating the spatial clustering characteristics of the heat load. Finally, by combining the temporal periodicity and spatial clustering characteristics, the temporal and spatial heterogeneity characteristics describing the heat load distribution are generated through spatiotemporal heterogeneity modeling.

[0022] Next, based on the spatiotemporal heterogeneity of heat load distribution, a digital twin-based heat flow model was constructed. Virtual mapping technology was used to link the physical pipeline network with the digital model, creating a digital twin model of heat flow. The core of this model is to map the geometric parameters, material properties, and operational status of the physical pipeline network into a digital environment, creating a highly accurate virtual representation. A graph neural network algorithm was used to abstract the network nodes and connections into a graph structure, simulating the dynamic transmission paths of heat energy within the pipeline network. The inputs to the graph neural network include the node feature matrix F and the adjacency matrix A, and the output is the state vector H for each node. The update formula is: H^(l+1)=σ(A·H^(l)·W^(l)), where H^(l) is the hidden state of the lth layer, W^(l) is the weight matrix, and σ is the activation function. Based on the dynamic heat transmission paths, the temperature field distribution and flow rate dynamics of the pipeline network nodes are extracted to generate the real-time state of heat flow. The model's accuracy was verified using thermodynamic equilibrium equations to ensure that the digital twin model accurately reflects the operational status of the physical pipeline network.

[0023] After obtaining the real-time status of thermal energy flow, further analysis is needed to determine the transmission efficiency fluctuations at each node in the pipeline network and identify energy efficiency loss points. The temperature gradient and flow rate variation data for each node are extracted from the real-time status of thermal energy flow to generate a thermal characteristic description of the network node. The temperature gradient is calculated using the formula: ∇T = (T_out - T_in) / L, where T_out is the outlet temperature, T_in is the inlet temperature, and L is the pipe length. The flow rate variation is calculated using the formula: ΔQ = Q_t - Q_(t-1), where Q_t is the current flow rate and Q_(t-1) is the previous flow rate. Based on the thermal characteristic description, the transmission efficiency fluctuation index is calculated for each node in the pipeline network. The transmission efficiency fluctuation index is defined as: η = ΔQ / ΔT, where ΔQ is the flow rate change and ΔT is the temperature gradient change. A fluctuation threshold is set to determine whether there are nodes with efficiency fluctuations exceeding the preset threshold, generating a set of candidate energy efficiency loss points. The accuracy of the loss points is verified through thermodynamic simulation, generating high-precision spatial locations of energy efficiency loss points.

[0024] For energy efficiency loss points, further analysis of their heat loss distribution is required. Based on the high-precision spatial positioning of the energy efficiency loss points, the geometric parameters and material properties of the corresponding pipe network segment are obtained to generate a basic physical property description of the pipe network segment. Geometric parameters include pipe diameter D, length L, and wall thickness δ, and material properties include thermal conductivity λ and specific heat capacity c. Thermodynamic simulation technology is used to analyze the distribution of heat loss within the pipe network segment and generate a spatial distribution map of heat loss. The heat loss calculation formula is: Q_loss=k·A·(T_1-T_2), where k is the heat transfer coefficient, A is the heat transfer area, and T_1 and T_2 are the inner and outer surface temperatures, respectively. Local heat loss distribution is identified through heat flux analysis, and loss hotspots are generated. Combining the local heat loss distribution and loss hotspots, a high-precision spatial positioning description of the loss points is generated.

[0025] Based on the high-precision spatial positioning of loss points, a reinforcement learning algorithm is used to optimize the pipeline network layout. A reinforcement learning model for pipeline network optimization is constructed, and the intelligent agent explores the optimal solution through interaction with the environment. The state space of the reinforcement learning model includes the temperature, flow rate, and pressure of each pipeline node, and the action space includes pipe diameter adjustment, valve opening adjustment, and pump station power change. The reward function is defined as: R = -α·C-β·E, where C is the operating cost, E is the carbon emission, and α and β are weight coefficients. Based on the reinforcement learning model, the pipe diameter and valve opening of each pipeline segment are adjusted to optimize the heat transfer path and generate a preliminary optimization solution. The optimization results are verified through thermodynamic simulation, and an optimized pipeline network configuration solution is generated. Based on the optimized pipeline network configuration solution, the pipeline network parameters in the digital twin model are updated to generate an optimized description of the heat flow state.

[0026] The optimized pipeline network configuration plan needs to re-simulate the heat energy flow state, calculate carbon emissions and operating costs, update the pipeline network parameters in the digital twin model, and re-simulate the heat energy flow state. Based on the heat energy flow state, calculate the optimized carbon emissions and operating costs, and generate low-carbon targets and economic evaluation results. The carbon emission calculation formula is: E=∑(Q_i·EF_i), where Q_i is the heat energy consumption of the i-th pipeline network section, and EF_i is the corresponding carbon emission factor. The operating cost calculation formula is: C=∑(P_i·Q_i), where P_i is the unit operating cost of the i-th pipeline network section. Determine whether the low-carbon target and economic requirements are met. If not, generate new optimization requirements. Extract heat load distribution data to generate a new round of optimization starting point.

[0027] Based on the updated configuration plan, valve and pump station parameters in the pipeline network control system are adjusted in real time. Based on the adjusted parameters, thermal energy flow data is collected and a description of the adjusted thermal energy flow state is generated. By comparing the system to the optimization target, it is verified that the system operation status has achieved the expected optimization results. Based on the verification results, a system optimization effect report is generated.

[0028] Finally, the heat load distribution prediction model is updated based on the verified heat flow data. A time series prediction algorithm is used to generate the heat load distribution trend for the next 24 hours. The prediction algorithm uses a long short-term memory (LSTM) network with the following formula: h_t = f(W_hh·h_(t-1)+W_xh·x_t+b_h), where h_t is the current hidden state, W_hh and W_xh are weight matrices, and b_h is the bias term. Based on the heat load distribution trend and the current optimization results, the optimization starting point for the next cycle is determined. Based on this optimization starting point, the optimization task description for the next cycle is generated.

[0029] In summary, the present invention realizes the accurate modeling and optimization of the thermal energy flow state of the park through the comprehensive application of sensor networks, digital twin technology, graph neural networks, thermodynamic simulation and reinforcement learning algorithms, providing comprehensive technical support for the optimization of the carbon neutrality path of the ecological park.

[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method, characterized by: The following steps are involved: Obtain real-time data on heat loads in various areas of the park. Collect temperature, flow, and pressure information through a sensor network, and combine time series analysis to determine the spatiotemporal heterogeneity of heat load distribution. Based on the spatiotemporal heterogeneity of heat load distribution, a digital twin-based thermal energy flow model was constructed. Graph neural networks were used to simulate the dynamic transmission path of heat energy in the pipe network to obtain the real-time status of heat energy flow. The system extracts the temperature gradient and flow rate changes of each node in the pipeline network from the real-time status of heat energy flow, calculates the transmission efficiency fluctuation index, determines whether there are nodes where the efficiency fluctuation exceeds the preset threshold, and determines the energy efficiency loss point; For energy efficiency loss points, the geometric parameters and material properties of the corresponding pipe network segments are obtained. The distribution pattern of heat energy loss is analyzed through thermodynamic simulation to obtain high-precision spatial positioning of the loss points. Based on the high-precision spatial positioning of loss points, a reinforcement learning algorithm is used to optimize the pipe network layout, adjust pipe diameters and valve openings, and generate an optimized pipe network configuration plan; Through the optimized pipeline network configuration plan, the pipeline network parameters in the digital twin model are updated, the heat flow state is re-simulated, the carbon emissions and operating costs are calculated, and whether the low-carbon goals and economic requirements are met; If carbon emissions or operating costs exceed a preset threshold, new heat load distribution data is extracted from the digital twin model, and the pipe network layout is iteratively optimized to generate an updated configuration plan; According to the updated configuration plan, the valve and pump station parameters in the pipe network control system are adjusted in real time, and the adjusted heat flow data is obtained to verify whether the system operation status has achieved the expected optimization effect; The heat load distribution prediction model is updated based on the verified heat flow data. The time series prediction algorithm is used to generate the heat load distribution trend for the next 24 hours and determine the optimization starting point for the next cycle. Among them, the spatiotemporal heterogeneity characteristics of the heat load distribution are determined based on the temporal periodicity and spatial aggregation characteristics of the heat load. The real-time status of thermal energy flow includes the temperature field distribution and dynamic changes of flow at the pipeline network nodes. The high-precision spatial positioning process of the loss point includes the identification of the local heat loss distribution and loss hotspot areas in the pipeline network section.

2. A multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1, characterized in that: The optimized pipeline network configuration plan includes a pipeline diameter adjustment plan, a valve opening optimization strategy, and a pump station power regulation plan. The carbon emissions and operating cost assessment results include the carbon emission reduction potential and operating cost savings rate. The updated configuration plan includes pipeline network topology adjustment and control parameter optimization. The heat load distribution prediction model includes short-term heat load trend prediction and long-term heat demand pattern analysis.

3. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: The steps to obtain real-time data on heat loads in various areas of the park, collect temperature, flow, and pressure information through a sensor network, and combine time series analysis to determine the spatiotemporal heterogeneity of heat load distribution are as follows: Based on the sensor network deployed in various areas of the park, real-time data on temperature, flow and pressure are collected, and multi-sensor data is synchronized in time and space to generate a unified heat load data set; Based on the heat load data set, a time series decomposition algorithm is used to separate the trend term, periodic term and random term of the heat load data, revealing the temporal periodic characteristics of the heat load. Based on the temporal periodicity characteristics, the cluster analysis method is used to divide the spatial distribution of heat load into multiple clustering areas, generating the spatial clustering characteristics of heat load; By integrating the temporal periodicity characteristics and spatial aggregation characteristics, the spatiotemporal heterogeneity characteristics of heat load distribution are modeled to generate a description of the spatiotemporal heterogeneity characteristics of heat load distribution.

4. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: Based on the spatiotemporal heterogeneity of heat load distribution, a digital twin-based thermal energy flow model is constructed. Graph neural networks are used to simulate the dynamic transmission path of heat energy in the pipe network. The specific steps to obtain the real-time status of thermal energy flow are as follows: Based on the spatiotemporal heterogeneity of heat load distribution, a digital twin model of heat flow is constructed, and the physical pipe network is linked to the digital model through virtual mapping technology; Based on the digital twin model, a graph neural network algorithm is used to abstract the nodes and connection relationships of the pipeline network into a graph structure to simulate the dynamic transmission path of heat energy in the pipeline network; Based on the dynamic transmission path of thermal energy, the temperature field distribution and flow dynamic change data of the pipeline network nodes are extracted to generate the real-time status of thermal energy flow; Combined with the real-time state of thermal energy flow, the accuracy of the model is verified by the thermodynamic equilibrium equation, and a real-time state description of thermal energy flow is generated.

5. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: The steps for extracting the temperature gradient and flow rate changes at each node in the pipeline network from the real-time state of heat flow, calculating the transmission efficiency fluctuation index, and determining whether there are nodes where the efficiency fluctuation exceeds the preset threshold are as follows: Based on the real-time status of thermal energy flow, the temperature gradient and flow change data of each node in the pipeline network are extracted to generate a description of the thermal characteristics of the pipeline network node; Based on the description of thermal characteristics, the transmission efficiency fluctuation index of each node in the pipeline network is calculated to evaluate the stability of heat energy transmission; Based on the transmission efficiency fluctuation index, a fluctuation threshold is set to determine whether there are nodes whose efficiency fluctuation exceeds the preset threshold, and a candidate set of energy efficiency loss points is generated; Based on the candidate set of energy efficiency loss points, the accuracy of the loss points is verified through thermodynamic simulation, and high-precision spatial positioning of the energy efficiency loss points is generated.

6. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: For energy efficiency loss points, the geometric parameters and material properties of the corresponding pipe network segments are obtained. The distribution pattern of heat loss is analyzed through thermodynamic simulation to obtain high-precision spatial positioning of the loss points. The specific steps are as follows: Based on the energy efficiency loss points, the geometric parameters and material properties of the corresponding pipe network segment are obtained to generate a basic physical property description of the pipe network segment; Based on the description of basic physical properties, thermodynamic simulation technology is used to analyze the distribution of heat loss in the pipe network section and generate a spatial distribution map of heat loss; Based on the spatial distribution map of heat energy loss, the local heat loss distribution is identified through heat flux analysis to generate loss hotspot areas; The local heat loss distribution and loss hotspot areas are integrated to generate a high-precision spatial positioning description of the loss points.

7. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: Based on the high-precision spatial positioning of loss points, a reinforcement learning algorithm is used to optimize the pipe network layout, adjust the pipe diameter and valve opening, and generate an optimized pipe network configuration plan. The specific steps are as follows: Based on the high-precision spatial positioning of loss points, a reinforcement learning model for pipe network optimization is constructed, exploring the optimal solution through the interaction between the intelligent agent and the environment; Based on the reinforcement learning model, the pipe diameter and valve opening of the pipe network section are adjusted to optimize the heat energy transmission path and generate a preliminary optimization plan; Based on the preliminary optimization plan, the optimization effect is verified through thermodynamic simulation to generate an optimized pipe network configuration plan; Combined with the optimized pipe network configuration plan, the pipe network parameters in the digital twin model are updated to generate an optimized description of the thermal energy flow state.

8. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: The steps to determine whether low-carbon goals and economic requirements are met by optimizing the pipeline network configuration plan, updating the pipeline network parameters in the digital twin model, resimulating the heat flow state, calculating carbon emissions and operating costs, and the following are specific steps: Based on the optimized pipe network configuration plan, the pipe network parameters in the digital twin model are updated and the heat energy flow state is re-simulated; Based on the thermal energy flow state, the optimized carbon emissions and operating costs are calculated to generate low-carbon targets and economic evaluation results; Based on the evaluation results, determine whether the low-carbon goals and economic requirements are met. If not, generate new optimization requirements; Based on new optimization requirements, heat load distribution data is extracted to generate a new round of optimization starting point.

9. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: According to the updated configuration plan, the valve and pump station parameters in the pipe network control system are adjusted in real time, the adjusted heat flow data is obtained, and the steps to verify whether the system operation status has achieved the expected optimization effect are as follows: Based on the updated configuration plan, the valve opening and pump station power parameters in the pipeline network control system are adjusted in real time; Based on the adjusted parameters, heat energy flow data is collected to generate an adjusted heat energy flow state description; Based on the adjusted description of the thermal energy flow state, by comparing with the optimization target, verify whether the system operation state has achieved the expected optimization effect; Based on the verification results, a system optimization effect report is generated.

10. The multi-dimensional ecological park carbon neutrality dynamic evaluation and path optimization method according to claim 1 is characterized by: The heat load distribution prediction model is updated based on the verified heat flow data. The time series prediction algorithm is used to generate the heat load distribution trend for the next 24 hours. The steps to determine the optimization starting point for the next cycle are as follows: Update the prediction model of heat load distribution based on the verified heat flow data; Based on the prediction model, a time series prediction algorithm is used to generate the heat load distribution trend for the next 24 hours; Based on the heat load distribution trend and the current optimization effect, determine the optimization starting point for the next cycle; Based on the optimization starting point, generate the optimization task description for the next cycle.

Images