A heat supply pipe network energy storage prediction method based on power grid peak regulation

Abstract

This invention relates to the field of energy storage prediction and discloses a method for predicting energy storage in heating networks based on power grid peak shaving, providing a solution for energy storage prediction in heating networks. The method includes acquiring relevant data through a multi-source data acquisition system, generating a network state feature set, constructing a nonlinear heat conduction model based on thermodynamic principles, incorporating characteristics such as thermal inertia, outputting predicted heat distribution results, establishing an energy flow optimization calculation framework, generating an optimal energy storage strategy evaluation report, constructing a multi-timescale energy storage potential calculation model, and generating energy storage performance comparison data. This invention, combined with future peak shaving plans, outputs a decision support report, providing comprehensive support for energy storage prediction and optimized operation of heating networks under power grid peak shaving scenarios.

Description

Technical Field

[0001] This invention relates to the field of energy storage prediction, and in particular to a method for predicting energy storage in heating networks based on grid peak shaving. Background Technology

[0002] With the transformation of my country's energy structure and the deepening of the electricity market, the participation of combined heat and power (CHP) units in grid peak shaving has become an inevitable trend. In northern heating areas, the coupling relationship between heating and power supply is becoming increasingly close, highlighting the contradiction between heat and power. Especially during the severe cold season, there is a significant time mismatch between grid peak shaving demand and residential heating demand: the output of heat sources decreases sharply during the evening peak shaving period, while this coincides with the peak of residential heating demand, resulting in insufficient heating capacity and seriously affecting the quality of heating.

[0003] The existing technology has the following shortcomings: Current predictions of energy storage in heating networks rely heavily on empirical formulas or statistical models based on historical data. These methods fail to accurately describe the nonlocal characteristics and memory effects of heat energy transmission in the network, resulting in limited prediction accuracy. This is especially true during peak-shaving periods when operating conditions change drastically, where prediction errors are more pronounced. Existing methods often simplify heating networks into centralized parameter systems, ignoring the spatial distribution characteristics of the networks. The transmission of heat energy in long-distance pipelines has significant spatial correlation and time delay effects, leading to inaccurate judgment of energy storage timing. Existing forecasting systems generally lack effective adaptive calibration mechanisms. Once the model parameters are determined, it is difficult to dynamically adjust them according to changes in operating conditions. When operating conditions change, such as extreme weather or pipeline renovation, the forecasting accuracy drops significantly, affecting the practicality and reliability of the system.

[0004] Therefore, we propose a prediction method for energy storage in heating networks based on grid peak shaving to address the above problems. Summary of the Invention

[0005] This invention provides a method for predicting energy storage in heating networks based on power grid peak shaving, which provides a solution for predicting energy storage in heating networks.

[0006] The first aspect of this invention provides a method for predicting energy storage in heating networks based on power grid peak shaving. This method acquires operating parameters of the heating network and power grid peak shaving demand information to generate a network state feature set. Based on the network state feature set, by introducing the historical dependence and spatial correlation characteristics of thermal inertia, dynamic response parameters for network heat energy transmission are obtained, and heat energy distribution prediction results are output. An energy flow optimization calculation framework is established, transforming power grid peak shaving demand into temperature constraints. Based on the heat energy distribution prediction results, an optimal energy storage strategy evaluation report is generated. Based on the optimal energy storage strategy evaluation report, energy storage performance comparison data is generated through thermodynamic simulation analysis. An adaptive calibration mechanism for prediction accuracy is designed, and by analyzing the deviation between real-time monitoring data and the energy storage performance comparison data, the model parameter correction amount is calculated, and updated heat energy distribution prediction results are output.

[0007] Optionally, in a first implementation of the first aspect of the present invention, the method includes: obtaining historical dependence parameters of thermal inertia based on the temperature gradient distribution and flow change patterns in the pipeline network state feature set, combined with the temperature change trends in historical operating data, and generating a thermal inertia feature vector; analyzing the thermal conduction correlation degree of each node in the pipeline network based on the pressure field characteristics and pipeline network topology data in the pipeline network state feature set, obtaining spatial correlation characteristic parameters, and generating a spatial correlation model; and obtaining dynamic response parameters of pipeline network heat energy transmission based on the thermal inertia feature vector and the spatial correlation model, and outputting heat energy distribution prediction results.

[0008] Optionally, in a second implementation of the first aspect of the present invention, temperature change sequences for multiple consecutive operating cycles are extracted from the historical operating data, and the periodicity of temperature changes is identified through time series analysis to generate a temperature change pattern library; based on the temperature change pattern library, the decay characteristics of temperature changes at different time scales are calculated through thermal inertia effect analysis to generate a thermal inertia feature parameter set; the thermal inertia feature parameter set is fused with real-time acquired temperature gradient distribution data to obtain a comprehensive thermal inertia historical dependency parameter.

[0009] Optionally, in a third implementation of the first aspect of the present invention, the method includes: based on the peak-shaving demand time-series signal from the power grid dispatch center, converting the peak-shaving power demand into a temperature variation range constraint of the heating network, and generating a set of temperature boundary conditions; based on the predicted heat energy distribution and the set of temperature boundary conditions, constructing a multi-objective optimization calculation model by establishing an energy functional extremum problem, and generating an optimization solution framework; based on the optimization solution framework, solving the optimal trajectory of energy distribution using a variational method to obtain the optimal temperature control strategy during different peak-shaving periods, and generating an optimal energy storage strategy evaluation report.

[0010] Optionally, in the fourth implementation of the first aspect of the present invention, based on the set of temperature boundary conditions and the predicted results of thermal energy distribution, a set of thermodynamic constraints is generated by identifying the allowable range and rate of change of temperature through a constraint analysis module; based on the peak-shaving demand time-series signal from the power grid dispatch center, a multi-objective optimization function including energy utilization efficiency, peak-shaving response speed and operation stability indicators is established through an objective function construction module, generating a set of objective function components; based on the set of thermodynamic constraints and the set of objective function components, a comprehensive optimization model including thermodynamic characteristics and power grid peak-shaving demand is constructed through a multi-objective optimization algorithm, generating an optimization solution framework.

[0011] Optionally, in the fifth implementation of the first aspect of the present invention, the method includes: establishing a multi-timescale analysis framework containing short-term, medium-term, and long-term analysis windows based on the pipeline topology and medium characteristic parameters, and generating a timescale division scheme; obtaining the thermal energy storage capacity and release characteristics at different timescales based on the optimal energy storage strategy evaluation report and the timescale division scheme, and generating a thermodynamic state evolution sequence; and based on the thermodynamic state evolution sequence, constructing a multi-scenario simulation environment including normal operating conditions, extreme operating conditions, and peak-shaving operating conditions by establishing a dynamic scenario simulation mechanism, and generating energy storage performance comparison data.

[0012] Optionally, in the sixth implementation of the first aspect of the present invention, the estimated equivalent energy storage capacity of the pipeline network is... ,but: Where ρ and c p It refers to the density and specific heat capacity of the medium, V. i The volume of the pipe segment associated with node i, ΔT i (t) represents the temperature change of node i at time t caused by the strategy. It is a time constant that reflects heat loss.

[0013] Optionally, in the seventh implementation of the first aspect of the present invention, based on the pipeline operation cycle data in the historical operation database and the peak-shaving demand characteristics collected in real time, the heat energy change pattern under different time scales is analyzed by the time feature extraction module to generate a multi-scale division basis; based on the multi-scale division basis and the pipeline transmission characteristic parameters from the pipeline topology database, a time window of three levels—short-term, medium-term, and long-term—is established by the dynamic time window division algorithm to generate time scale division parameters; based on the time scale division parameters, a multi-time scale analysis framework with hierarchical characteristics is constructed by establishing a time scale correlation model, and an analysis framework description file is generated.

[0014] Optionally, in the eighth implementation of the first aspect of the present invention, the method includes: acquiring real-time monitored pipeline operation data, performing deviation analysis on the data and comparing it with the energy storage performance data to generate a deviation feature vector; based on the deviation feature vector, determining the degree of influence of each parameter in the nonlinear heat conduction calculation model on the prediction accuracy through parameter sensitivity analysis, and generating a parameter correction sequence; calculating the specific correction amount of each parameter according to the parameter correction sequence by establishing a recursive correction algorithm, and generating a parameter correction set; updating the parameters of the nonlinear heat conduction calculation model based on the parameter correction set, generating an updated heat energy distribution prediction model through model verification testing, and outputting the updated heat energy distribution prediction result.

[0015] Optionally, in the ninth implementation of the first aspect of the present invention, the method further includes: based on the optimal energy storage strategy evaluation report and energy storage performance comparison data, a comprehensive verification mechanism is established to cross-verify the prediction results with actual operating data to generate a verification analysis report; based on the verification analysis report, an economic benefit evaluation model is established to calculate the peak-shaving revenue and operating costs under different energy storage strategies to generate a comprehensive benefit evaluation report; and based on the comprehensive benefit evaluation report and the future peak-shaving plan from the power grid dispatch center, an operation optimization suggestion generation mechanism is established to output a decision support report.

[0016] The mechanism of this invention is as follows: by establishing a complete digital processing flow from data acquisition and model calculation to optimization decision-making, accurate prediction and strategy optimization of the energy storage process of heating pipe networks are realized; Beneficial effects: It fully considers the complex characteristics of heat energy transfer process, can accurately calculate the dynamic response of heat energy transfer in pipeline network, and outputs heat energy distribution prediction results including time delay effect, which effectively improves the accuracy of heat energy distribution prediction and provides a more reliable basis for power grid peak shaving. It achieves comprehensive optimization of the grid peak-shaving demand and the thermodynamic characteristics of the heating network, and can generate the optimal temperature control strategy during different peak-shaving periods. This improves the scientificity and effectiveness of the energy storage strategy, gives full play to the energy storage advantages of the heating network, and helps to alleviate the grid peak-shaving pressure. It can comprehensively consider the thermal energy storage and release characteristics of the pipeline network at different time scales, providing a more complete basis for energy storage decisions under different operating conditions, which helps to formulate more reasonable and flexible energy storage strategies and improve energy utilization efficiency. By adjusting model parameters in a timely manner according to changes in pipeline operating conditions, the accuracy and reliability of the prediction model are maintained. This solves the problems of traditional prediction methods, such as the lack of adaptive calibration and the decrease in prediction accuracy as operating conditions change, and ensures the effectiveness and stability of energy storage strategies based on prediction results. Attached Figure Description

[0017] Figure 1This is a schematic diagram of an embodiment of the heating network energy storage prediction method based on power grid peak shaving in this invention. Figure 2 This is a schematic diagram of another embodiment of the heating network energy storage prediction method based on power grid peak shaving in this invention. Figure 3 This is a schematic diagram of an embodiment of the heating network energy storage prediction device based on power grid peak shaving in this invention. Detailed Implementation

[0018] This invention provides a method for predicting energy storage in heating networks based on grid peak shaving, offering a solution for predicting energy storage in heating networks. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the heating network energy storage prediction method based on power grid peak shaving in this invention includes: 101. Obtain heating network operation parameters and power grid peak-shaving demand information through a multi-source data acquisition system, and generate a network status feature set including temperature gradient distribution, pressure field characteristics and flow change law; It is understood that the executing entity of this invention can be a heating network energy storage prediction device based on grid peak shaving, or it can be a terminal or a server; the specific implementation is not limited here. This embodiment of the invention will be described using a server as an example.

[0020] It should be noted that, taking a typical urban district heating system as an example, the system covers a 5-kilometer pipeline network and includes a heat source station, 3 pumping stations, and 10 user-end nodes. The data collection period is the peak hours from 8:00 to 9:00 on a certain winter day.

[0021] Three types of data sources are integrated: Heating network operating parameters: acquired through a sensor network deployed at key nodes of the network. These include: Temperature sensors: 15 measuring points are set up at the heat source outlet, pump station inlet, and user end, recording temperature values ​​(unit: °C) every 5 minutes. At 8:00 AM, the heat source outlet temperature was 95 °C, the pump station 1 inlet temperature was 85 °C, and the lowest temperature at the user end was 65 °C. Pressure sensors: measuring pressure (unit: kPa) at the same nodes. The heat source outlet pressure was 600 kPa, the pump station 1 inlet pressure was 550 kPa, and the user end pressure fluctuated between 480-520 kPa. Flow meters: installed on the main pipeline and branch points to monitor flow rate (unit: m³ / s). 3 / h), the main pipeline flow rate is 200 m³ / h. 3 / h, 8:30 AM, due to increased user demand, the speed increased to 230m. 3 / h.

[0022] Grid peak-shaving demand information: obtained in real time from the grid dispatch center, including peak-shaving period (8:00-10:00), expected load reduction (in this example, the heating system is required to reduce heat output by 50kWh within 1 hour), and electricity price signal (peak electricity price of 0.8 yuan / kWh).

[0023] After preprocessing (outlier removal and time alignment), the collected raw data is used to generate a feature set through spatial interpolation and time series analysis: Temperature gradient distribution: Based on temperature data from 15 measuring points, the temperature distribution along the entire pipeline network is calculated using inverse distance weighted interpolation. From the heat source to the farthest user end, the temperature gradient shows a decrease of 6°C per kilometer, forming a linear distribution. The feature set is recorded as a list of temperature values ​​corresponding to spatial coordinates, and a mapping table between location coordinates (x, y) and temperature. Pressure field characteristics: Combining pressure sensor data, the pressure field is derived through a fluid dynamics model. The pressure near pump station 1 is relatively high (550 kPa), while the pressure in the user end area is relatively low (average 500 kPa). The feature set describes the pressure decay law in the form of pressure contour maps. Flow rate variation law: The time series data of the flow meter is analyzed to identify periodic patterns. In this example, the flow rate shows a linear increase from 8:00 to 8:30, with a growth rate of 10 m³ / s. 3 / hper30 minutes, the feature set is summarized into a traffic-time curve, and key turning points such as peaks and valleys are marked.

[0024] The feature set is output in the form of a structured table, which includes fields such as timestamp, spatial location, temperature gradient value, pressure field intensity, and flow rate change.

[0025] 102. Based on the first principles of thermodynamics, a nonlinear heat conduction calculation model is constructed. The network state feature set is used as input. By introducing the historical dependence and spatial correlation characteristics of thermal inertia, the dynamic response parameters of network heat energy transmission are calculated, and the predicted heat energy distribution results including time delay effects are output. It should be noted that, following the same heating system scenario mentioned above, the network status feature set collected from 8:00 to 9:00 is used as input.

[0026] The model is built upon the law of conservation of energy, discretizing a 5-kilometer-long heating pipeline network into 100 computational units. The thermal dynamics of each unit are determined by its own heat capacity, conduction with adjacent units, and water convection. The input feature set data includes: temperature gradient distribution (linear distribution from 95°C at the heat source to 65°C at the end), pressure field characteristics (pressure decaying from 600 kPa to 480 kPa), and flow rate variation (flow rate from 200 m³ / h to 400 m³ / h). 3 / h linearly increases to 230m 3 / h).

[0027] Thermal inertia is introduced in the following ways: Historical dependence: The calculation not only relies on the current data from 8:00 to 9:00, but also incorporates historical operating data from the previous hour (7:00 to 8:00) as the initial field. The system underwent a small temperature increase at 7:30, and this historical heat input is included in the calculation as the system's "thermal memory," affecting the current thermal response speed. Spatial correlation: The model considers the thermal interaction between adjacent pipe segments in the pipeline network topology. The pipe segment where pump station 1 is located has a higher temperature, and its heat will be conducted to the lower-temperature user-end pipe segments on both sides. This spatial correlation is quantified by setting the thermal conductivity of the pipe wall material (low-carbon steel) and the characteristics of the insulation layer. The model obtains key dynamic response parameters through iterative calculation: Heat propagation velocity: Calculated at the current flow rate of 230m³ / h... 3 At a rate of / h, it takes approximately 30 minutes for a temperature change from the heat source to reach the furthest user. Temperature decay coefficient: Combining flow rate and pipe diameter, it is calculated that the temperature drop per kilometer during heat transport in the pipeline network is approximately 5.8℃.

[0028] The model outputs a predicted heat distribution for the next hour (9:00-10:00), incorporating time-delay effects. The results are presented in a table format, predicting temperatures at different locations at various future time points. The prediction shows that if the water supply temperature is lowered to 90°C at the heat source at 9:00, the temperature at the user's measuring point at the end of the pipe network will remain above 63°C until 9:30, but will begin to drop significantly to 60°C after 9:40. This prediction clearly demonstrates the delay effect of heat transfer within the pipe network.

[0029] 103. Establish an energy flow optimization calculation framework based on variational principles, transform the grid peak-shaving demand into temperature constraints, take the thermal energy distribution prediction results as input, and generate an optimal energy storage strategy evaluation report by solving the energy functional extremum problem. It should be noted that the predicted heat distribution for the next hour (9:00-10:00) output in step 102 is used as the core input.

[0030] An optimization framework is established to respond as much as possible to the power grid's peak-shaving instructions while meeting users' basic heating needs. The input consists of two parts: the power grid's peak-shaving demand is transformed into a temperature constraint: the power grid requires a reduction of 50 kWh of heat output within this hour (9:00-10:00). The server translates this demand into a temperature constraint on the pipeline network operation: while ensuring that the indoor temperature of the furthest user is not lower than the legally mandated 18℃ (corresponding to a water supply temperature not lower than 60℃), a suitable reduction in the heat source's factory outlet temperature is allowed. The heat distribution prediction result is used as input: the key information obtained from step 102 is that there is a thermal delay of approximately 30 minutes in the pipeline network; the prediction indicates that changes in the heat source temperature at 9:00 will not significantly affect end users until after 9:30.

[0031] The optimization framework defines this problem as: finding an optimal heat source temperature regulation curve that minimizes the total energy consumption of the entire pipeline network during peak periods (i.e., responding to grid demand) while ensuring that the temperature at each user point does not fall below safety constraints throughout the process. This is constructed as a problem of solving for the extremum of an energy functional. The server performs the calculation, balancing the contradiction between "reducing the heat source temperature to save energy" and "maintaining the terminal temperature to meet the standard." The calculation process comprehensively considers the thermal inertia (delay effect) and spatial structure (temperature gradient) of the pipeline network to find a systematic optimal solution, rather than simply reducing the temperature uniformly.

[0032] After solving the problem, the server generates an evaluation report containing specific operational plans. The core content of the report is as follows: Optimal Strategy: It is recommended to gradually reduce the heat source outlet temperature from the current 95℃ to 88℃ at 9:00 AM, maintain this temperature until 9:45 AM, and then gradually increase it back to 92℃. This strategy fully utilizes the 30-minute thermal delay characteristic of the pipeline network. Expected Effect Assessment: Calculations show that after implementing this strategy, the water temperature of the end user in the pipeline network will begin to slowly decrease from 63℃ at 9:35 AM, but will still be maintained at 60.5℃ at 9:55 AM, exceeding the constraint conditions. Simultaneously, the total heat energy reduction is expected to reach 52 kWh, meeting the grid's 50 kWh peak-shaving demand.

[0033] Risk Description: The report points out that if the actual flow rate is higher than the predicted value, the terminal temperature may approach the critical point more quickly. It is recommended to strengthen the monitoring of terminal temperature after 9:30.

[0034] 104. Based on the pipeline network topology and medium characteristic parameters, construct a multi-timescale energy storage potential calculation model, take the optimal energy storage strategy evaluation report as input, and generate comparative data on energy storage performance under different operating conditions through thermodynamic simulation analysis. It should be noted that the optimal energy storage strategy evaluation report generated in step 103 (core recommendation: reduce the heat source temperature from 95℃ to 88℃ at 9:00) is used as input for in-depth simulation analysis of multiple scenarios.

[0035] The model is constructed based on the specific physical characteristics of the heating system: Pipeline topology: Considering a 5 km long branched pipeline network, including 1 heat source, 3 pumping stations, and 10 user branches. The model accurately reflects the pipeline length, diameter, burial depth, and insulation layer thickness. Medium characteristic parameters: The heating medium is water, and its specific heat capacity, density, and viscosity at different temperatures are set as the basic properties of the model. The input to the model is the optimal strategy report from step 103, i.e., the specific heat source temperature adjustment curve (dropping to 88°C at 9:00, then rising again after 9:45).

[0036] The model was simulated on three typical time scales to assess the energy storage potential: Short-term scale (next 2 hours): The simulation focused on the direct impact of implementing the strategy on the subsequent grid peak-shaving window (9:00-10:00) and the following hour (10:00-11:00). The simulation showed that after cooling at 9:00, the total system heat output remained stable below the base load between 9:15 and 9:50, with a cumulative heat reduction of 54.8 kWh over 2 hours, slightly higher than the strategy's expected 52 kWh. Medium-term scale (next 6 hours): The simulation analyzed the subsequent impact of the strategy. The simulation found that after the temperature recovery began at 9:45, the system needed a small energy compensation phase between 10:00 and 11:00 to restore the network temperature field to normal levels, with a compensation amount of approximately 15 kWh. This indicates that the energy storage strategy has a certain "rebound effect." Long-term scale (next 24 hours): The simulation considered the potential for implementing a similar strategy multiple times within a day. The assessment shows that, while ensuring user comfort, the system can perform similar operations approximately three times a day, with a maximum theoretical energy storage potential of 160 kWh throughout the day.

[0037] By setting different boundary conditions, the adaptability of the optimal strategy was tested, and comparative data was generated: Scenario 1 (Normal Operation): Outdoor temperature is -5 degrees Celsius, user load is stable. Simulation results show: Successful peak shaving of 54.8 kWh, minimum terminal temperature of 60.5 degrees Celsius, strategy successful. Scenario 2 (Sudden Load Increase): Simulation shows a sudden increase in traffic of 10% due to a new user connection at 9:30. Simulation results show that the terminal temperature will drop to 59.8 degrees Celsius at 9:50, slightly below the safety constraint of 60 degrees Celsius, posing a risk. Scenario 3 (Extreme Cold): Outdoor temperature drops to -15 degrees Celsius. Simulation shows that if this strategy is still implemented, the terminal temperature will prematurely fall below the constraint line; this strategy is not applicable, and a gentler peak shaving solution (cooling down to 90 degrees Celsius) is needed.

[0038] Output a data table comparing energy storage performance, clearly showing the effectiveness, boundaries and risks of the recommended strategy in different scenarios, providing dispatchers with comprehensive decision support, and clarifying under what conditions the energy storage strategy can be executed safely and efficiently.

[0039] 105. Design an adaptive calibration mechanism for prediction accuracy. By analyzing the deviation between real-time monitoring data and energy storage performance comparison data, calculate the correction amount of model parameters and output the updated thermal energy distribution prediction results for subsequent prediction cycles. It should be noted that the system has executed the optimal energy storage strategy generated in step 103 (reducing the heat source temperature to 88°C at 9:00) and has obtained the actual operating data for the period from 9:00 to 10:00.

[0040] The calibration mechanism was initiated at 10:00 AM, and its workflow is as follows: Real-time monitoring and data acquisition: The system obtains actual operating data from 9:00 AM to 10:00 AM from the data acquisition system. Key data includes: Actual heat source outlet temperature: It stabilized at 88.2℃ after 9:00 AM, which is basically consistent with the strategy setting of 88℃. Actual temperature of the furthest user (User 10): It started to decrease from 63.0℃ at 9:00 AM, but the measured temperature was 61.2℃ at 9:50 AM.

[0041] Actual total system traffic: An unexpected peak occurred at 9:30, reaching 245m. 3 / h, higher than the predicted 230m 3 / h. Deviation analysis is performed on the comparison data with energy storage performance: The server performs a comparative analysis on the above measured data with the predicted data generated by the "normal operating condition" simulation in step 104.

[0042] Key deviation found: For user 10's temperature, the model predicted 60.5℃ at 9:50, but the actual value was 61.2℃, a positive deviation of +0.7℃. This means that the actual heat loss in the pipeline is less than the model's preset value, or the brief increase in flow rate brought additional heat transfer, offsetting part of the temperature drop.

[0043] Another discrepancy is the thermal delay time. It was predicted that it would take 30 minutes for heat to transfer from the heat source to the end, but the actual data showed that the temperature of User 10 had already started to drop significantly at 9:25, with a delay time of about 25 minutes.

[0044] Calculate the correction amount of the model parameters: Based on the deviation analysis results, the calibration mechanism quantitatively calculates the correction amount of the parameters of the nonlinear heat conduction model in step 102.

[0045] To address the deviation in heat loss, the model adjusts the overall thermal conductivity of the pipeline network downward by 5 percent to reflect the situation where the pipeline insulation performance is better than expected.

[0046] To address the deviation in thermal delay time, the model slightly adjusts the effective flow velocity parameter of the medium in the pipe upwards by 3%, making it more consistent with the faster heat transfer rate actually observed.

[0047] After the calculations were completed, the heat transfer prediction model was immediately updated with the new parameter corrections mentioned above. This updated model will be used for the calculations of the next prediction cycle (predicting the 10:00-11:00 period starting at 10:00). This means that when the system receives new peak-shaving demands from the grid, the prediction model it is based on has been calibrated and can more accurately predict temperature changes and delay effects, thereby generating a more feasible and lower-risk energy storage strategy.

[0048] In this embodiment of the invention, a multi-source data acquisition system integrates heating network operating parameters and power grid peak-shaving demand information, covering multi-dimensional data such as temperature, pressure, and flow rate. This system provides a detailed and accurate depiction of the network's state characteristics, offering a solid data foundation for subsequent prediction and strategy formulation. Based on the first principles of thermodynamics, and considering the historical dependence and spatial correlation characteristics of thermal inertia, the network is discretized to make the model more closely reflect actual network operation. By incorporating historical operating data and considering the thermal interaction between adjacent pipe sections, the dynamic response parameters of network heat energy transfer can be accurately calculated, outputting heat energy distribution prediction results that include time delay effects. This clearly reflects the heat transfer delay characteristics, providing a key basis for peak-shaving strategies. An optimization framework based on variational principles is established, transforming power grid peak-shaving demand into temperature constraints. Combined with heat energy distribution prediction results, the energy functional extremum problem is solved, generating an optimal energy storage strategy evaluation report. This framework comprehensively considers the thermal inertia and spatial structure of the pipeline network, balancing the contradiction between energy conservation and maintaining the required terminal temperature, and provides a systematic optimal solution rather than a simple operational scheme. It constructs a multi-timescale energy storage potential calculation model, and conducts multi-scenario simulation analysis of the optimal energy storage strategy based on the pipeline network topology and medium characteristic parameters. Energy storage potential is assessed at short-term, medium-term, and long-term timescales, while different boundary conditions are set to test the adaptability of the strategy, generating comparative data on energy storage performance. This provides dispatchers with comprehensive decision support and clarifies the conditions for strategy execution. An adaptive calibration mechanism for prediction accuracy is designed. By analyzing the deviation between real-time monitoring data and comparative data on energy storage performance, the model parameter correction is calculated, and the thermal energy distribution prediction model is updated in a timely manner. This dynamic calibration method allows the model to continuously optimize according to actual operating conditions, improving the accuracy of subsequent prediction cycles, thereby generating more operable and lower-risk energy storage strategies that effectively cope with various changes in actual operation.

[0049] Please see Figure 2 Another embodiment of the heating network energy storage prediction method based on power grid peak shaving in this invention includes: 201. Obtain heating network operation parameters and power grid peak-shaving demand information through a multi-source data acquisition system, and generate a network status feature set including temperature gradient distribution, pressure field characteristics and flow change law; Specifically, by deploying temperature sensor arrays, pressure transmitters, and flow meters at key nodes of the heating network, real-time temperature monitoring data, pressure monitoring data, and flow monitoring data are collected during the operation of the heating network, generating a real-time dataset containing timestamps. Based on the real-time dataset, the collected monitoring data is processed by a data preprocessing module to perform time series alignment and spatial interpolation, generating a multi-dimensional feature matrix with a unified time reference and complete spatial coverage. Based on the multi-dimensional feature matrix, combined with the peak-shaving demand time-series signal received from the power grid dispatch center, a network status feature set containing temperature gradient distribution, pressure field characteristics, and flow change patterns is constructed through spatiotemporal correlation analysis.

[0050] It should be noted that a simplified heating network system is set up, containing three key nodes: Node A (heat source outlet), Node B (mid-section of the network), and Node C (user-end inlet). Deployed sensors include temperature sensor arrays (multiple sensors at each node measuring different depths), pressure transmitters, and flow meters.

[0051] Data Acquisition: From 10:00 to 10:30 on December 1, 2023, the sensors collected real-time data: Temperature data: At 10:00, 10:05, and 10:10, the temperatures at node A were 85℃, 84℃, and 83℃, respectively; at node B, they were 80℃, 79℃, and 78℃; and at node C, they were 75℃, 74℃, and 73℃. Pressure data: The pressures at node A were 0.5MPa, 0.51MPa, and 0.49MPa; at node B, they were 0.45MPa, 0.46MPa, and 0.44MPa; and at node C, they were 0.4MPa, 0.39MPa, and 0.38MPa. Flow rate data: The flow rate at node A was 100 m³ / s. 3 / h、102m 3 / h、98m 3 / h; Node B is 95m 3 / h、93m 3 / h、97m 3 / h; Node C is 90m 3 / h、88m 3 / h、92m 3 / h. All data includes a timestamp (2023-12-01 10:00:00).

[0052] The collected monitoring data were standardized to 5-minute time intervals (i.e., 10:00, 10:05, 10:10, etc.). Since pressure data for node B at 10:05 was missing, spatial interpolation was used (based on the pressure values ​​of nodes A and C, the pressure of node B was estimated to be 0.455 MPa through linear interpolation). After processing, a multidimensional feature matrix was generated: rows represent time points (3 time points), columns represent node parameters (each node has three columns: temperature, pressure, and flow rate, for a total of 9 columns), and matrix values ​​are [10:00, A temperature 85, A pressure 0.5, A flow rate 100, B temperature 80, B pressure 0.45, B flow rate 95, C temperature 75, C pressure 0.4, C flow rate 90].

[0053] Based on the grid peak-shaving demand: The peak-shaving demand time sequence signal is received from the grid dispatch center. During the period from 10:00 to 10:30, the peak-shaving power demand is to reduce the load by 50MW. This signal is converted into a time sequence (50MW demand at 10:00, 48MW demand at 10:05, and 52MW demand at 10:10).

[0054] Based on the multidimensional feature matrix and peak-shaving demand, correlation analysis is performed. The temperature gradient distribution is obtained by calculating the temperature difference between nodes; the gradient from node A to node B is 5℃ / km at 10:00 (node ​​spacing is set at 1km), and its change over time is analyzed. Pressure field characteristics are identified through the spatial distribution of pressure values; node A has high pressure, and node C has low pressure, forming a pressure field vector. Flow rate variation is analyzed through time series analysis of flow fluctuations; the flow rate increases at 10:05 in response to peak-shaving demand. Finally, a pipeline network state feature set is constructed, including temperature gradient distribution (gradient sequence), pressure field characteristics (pressure distribution map), and flow rate variation (flow trend line), for subsequent steps.

[0055] 202. Based on the first principles of thermodynamics, a nonlinear heat conduction calculation model is constructed. The network state feature set is used as input. By introducing the historical dependence and spatial correlation characteristics of thermal inertia, the dynamic response parameters of network heat energy transmission are calculated, and the predicted heat energy distribution results including time delay effects are output. Specifically, based on the temperature gradient distribution and flow rate variation patterns in the pipeline network state feature set, an energy conservation equation is established using the first law of thermodynamics. Combined with the temperature variation trend in historical operating data, the historical dependence parameters of thermal inertia are calculated, generating a thermal inertia feature vector that includes the time accumulation effect. Based on the pressure field characteristics and pipeline topology data in the pipeline network state feature set, the thermal conduction correlation degree of each node in the pipeline network is analyzed by establishing a spatial correlation function, and spatial correlation characteristic parameters are calculated, generating a spatial correlation model that includes nonlocal conduction effects. Using the thermal inertia feature vector and the spatial correlation model as input, the dynamic response parameters of pipeline network heat energy transfer are calculated by solving a set of nonlinear heat conduction equations, and the predicted results of heat energy distribution including the time delay effect are output. The historical dependence parameters of thermal inertia are obtained by: extracting temperature change sequences from multiple consecutive operating cycles from historical operating data, identifying the periodicity of temperature changes through time series analysis, and generating a temperature change pattern library containing periodic characteristics; based on the temperature change pattern library, calculating the decay characteristics of temperature changes at different time scales through thermal inertia effect analysis, and generating a set of thermal inertia feature parameters containing time decay factors; and fusing the set of thermal inertia feature parameters with real-time acquired temperature gradient distribution data, and calculating the comprehensive historical dependence parameters of thermal inertia through a weighted average algorithm.

[0056] It should be noted that the current time is 10:30 AM on December 1, 2023. We will use the "pipeline network status feature set" obtained in step 201 as input. This feature set includes the temperature gradient distribution of the pipeline network from 10:00 to 10:30 (the gradient from A to B is 5℃ / km), pressure field characteristics (the pressure at node A is the highest at 0.5MPa, and the pressure at node C is the lowest at 0.4MPa), and flow rate variation pattern (the flow rate shows a small peak of 102m³ at 10:05). 3 / h).

[0057] A temperature change pattern library was constructed: The system extracted temperature data from the historical operation database for the same time period (10:00-11:00) over the past 7 days. Analysis revealed that due to the daily peak heating patterns, the temperature at node A typically showed a slow decreasing trend between 10:00 and 10:45 (an average decrease of 0.03℃ per minute), and then stabilized after 10:45 due to peak-shaving preparations. This daily recurring pattern was identified and stored in the temperature change pattern library.

[0058] Generating a set of thermal inertia characteristic parameters: Based on a model library, thermal inertia effect analysis was performed. It was found that temperature changes exhibit strong inertia in the short term (next 5 minutes) and decay slowly (attenuation factor set to 0.9, indicating a significant weighting of historical trends); however, in the medium term (next 30 minutes), inertia weakens and decays significantly (attenuation factor becomes 0.6). This yields a set of characteristic parameters containing attenuation factors at different time scales.

[0059] Calculating the comprehensive historical dependency parameters: The above-mentioned set of feature parameters is fused with real-time data. Current real-time data shows that the temperature at node A dropped from 85℃ at 10:00 to 83℃ at 10:30, with a rate of decrease (0.067℃ / minute) slightly faster than the historical pattern. Using a weighted average algorithm, a higher weight (0.7) is assigned to the real-time data, while the historical pattern has a weight of 0.3. The calculated comprehensive historical dependency parameters indicate that the current "thermal inertia" of the pipeline network is that the temperature decrease trend will maintain strong inertia in the next 5 minutes, but the rate of decrease will slow down slightly.

[0060] Based on the pipeline topology (the distance from node A to B is known to be 1km, the distance from B to C is known to be 1km, the pipeline material is steel, and the insulation layer characteristics are known) and the current pressure field characteristics (the pressure at node A is 0.5MPa, the pressure at node C is 0.4MPa, forming a stable pressure difference driving the flow).

[0061] Analysis was conducted using a spatial correlation function. This function considers pipe segment length, material thermal conductivity, medium velocity (estimated from flow rate and pipe diameter), and pressure difference. The calculations show that the thermal conduction correlation between node A and node B is very high (correlation weight 0.8), the correlation between node A and node C weakens due to the greater distance (correlation weight 0.5), and the correlation between node B and node C is also very high (weight 0.7). This generates a quantitative spatial correlation model, indicating that the transfer of heat from A to C depends not only on the direct path but also significantly on the state of intermediate node B (non-local conduction effect).

[0062] The thermal inertia eigenvectors (including time-cumulative effects) and spatial correlation models (including nonlocal conduction effects) obtained above are used as key inputs and substituted into a set of nonlinear heat conduction equations based on the first law of thermodynamics. This set of equations describes the complex transfer process of heat energy in the pipe over time and space.

[0063] The system of equations was solved numerically to obtain the dynamic response parameters of heat transfer in the pipeline network. The calculation showed that if a control action (fine-tuning the temperature) was applied at node A at 10:30, it would take about 3 minutes (time delay 1) for heat to be transferred to node B and about 7 minutes (time delay 2) for heat to be transferred to node C.

[0064] The output includes a predicted heat distribution result that incorporates the time delay effect. The prediction is as follows: Within the next hour, the temperature at node A can be maintained at 82±1℃ under control; the temperature change at node B will lag by approximately 3 minutes, expected to reach 79.5℃ at 10:33; and the temperature change at node C will lag by approximately 7 minutes, expected to reach 74.5℃ at 10:37. This prediction clearly includes information about the time delay in the heat transfer process.

[0065] 203. Establish an energy flow optimization calculation framework based on variational principles, transform the grid peak-shaving demand into temperature constraints, take the thermal energy distribution prediction results as input, and generate an optimal energy storage strategy evaluation report by solving the energy functional extremum problem. Specifically, based on the peak-shaving demand time-series signal from the power grid dispatch center, the peak-shaving power demand is transformed into a temperature variation range constraint of the heating network through a constraint condition transformation algorithm, generating a set of temperature boundary conditions containing time-dimensional constraints. Using the predicted heat energy distribution and the set of temperature boundary conditions as input, a multi-objective optimization calculation model containing the thermodynamic characteristics of the network and the peak-shaving demand of the power grid is constructed by establishing an energy functional extremum problem, generating an optimization solution framework containing weight coefficients. Based on the optimization solution framework, the optimal trajectory of energy distribution is solved using the variational method, and the optimal temperature control strategy in different peak-shaving periods is calculated, generating an optimal energy storage strategy evaluation report containing time series. A multi-objective optimization computational model incorporating the thermodynamic characteristics of the pipeline network and the peak-shaving demand of the power grid is constructed, including: based on the set of temperature boundary conditions and the predicted results of heat energy distribution, the allowable range and rate of temperature change are identified through the constraint analysis module, generating a set of thermodynamic constraints containing multiple constraints; based on the peak-shaving demand time-series signal from the power grid dispatch center, a multi-objective optimization function including energy utilization efficiency, peak-shaving response speed, and operational stability indicators is established through the objective function construction module, generating a set of objective function components including weight coefficients; using the set of thermodynamic constraints and the set of objective function components as input, a comprehensive optimization model incorporating thermodynamic characteristics and the peak-shaving demand of the power grid is constructed through a multi-objective optimization algorithm, generating an optimization solution framework including a dynamic weight adjustment mechanism.

[0066] It should be noted that the "heat distribution prediction results" for the next hour (10:30 to 11:30 on December 1, 2023) have been obtained from step 202: the temperature controllable range of node A is 82±1℃, and the temperature changes of nodes B and C have delays of approximately 3 minutes and 7 minutes, respectively. Meanwhile, the peak-shaving demand timing signal received from the power grid dispatch center clearly requires that, during the 15-minute period from 11:00 to 11:15, the heating network needs to rapidly increase its heat output to handle the peak load equivalent to 50MW of generating capacity.

[0067] The 50MW peak-shaving power demand is transformed into temperature constraints on key nodes of the heating network using a constraint transformation algorithm. This transformation is based on the physical relationship between heat power, medium flow rate, specific heat capacity, and temperature changes. Calculations show that to meet this demand, the network needs to increase heat output during peak-shaving periods. Therefore, the following set of temperature boundary conditions is generated: Time range constraint: applies from 11:00 to 11:15.

[0068] Temperature variation range constraints: The outlet temperature of node A must be increased from the current baseline of 82℃ to no less than 85℃, but must not exceed the safety limit of 88℃. The temperature of user-end node C must not be lower than 75℃ to ensure heating quality. Rate of change limit: The rate of temperature increase must not exceed 0.5℃ / minute to prevent thermal shock.

[0069] A set of thermodynamic constraints was established: the above temperature boundary conditions were combined with the predicted heat energy distribution. The constraint analysis module identified multiple constraints, including: the upper temperature limit (88℃) and lower temperature limit (75℃) of node A, the lower temperature limit (75℃) of node C, and the maximum allowable temperature change rate of the entire pipeline network (0.5℃ / minute).

[0070] Establish a set of objective function components: The objective function building module defines three optimization objectives: Maximize energy utilization efficiency: Strive to achieve maximum effective heating with minimum pumping power consumption. Maximize peak-shaving response speed: Strive for rapid temperature rise to reach the target peak-shaving power as quickly as possible. Maximize operational stability: Strive for stable temperature changes and avoid large fluctuations.

[0071] To coordinate different objectives, weighting coefficients are assigned: response speed has the highest weight (0.5), followed by stability (0.3), and efficiency has a slightly lower weight (0.2), forming the objective function. The weighted summation can be expressed as: .

[0072] Generate an optimization solution framework: Input the set of thermodynamic constraints and the set of objective function components into a multi-objective optimization algorithm to construct a comprehensive optimization model. This framework includes a dynamic weight adjustment mechanism. If the algorithm finds that rapid heating leads to a sharp decrease in stability, it may fine-tune the weights during the solution process to seek a new balance between response speed and stability.

[0073] Based on this optimization framework, the optimal trajectory of energy distribution is solved using a variational method. This method does not seek a single optimal setpoint, but rather finds the smoothest, most energy-efficient, and fastest path (trajectory) of temperature change from the current state to the state that meets peak-shaving requirements.

[0074] The time series of the optimal temperature control strategy was calculated. The evaluation report indicates: 10:50: Begin slowly increasing the temperature of node A to prepare for a subsequent rapid increase. 11:00 (Peak shaving begins): Increase the temperature of node A to 86℃ at a rate of 0.4℃ / minute. 11:07 (Considering delay): The temperature of node B is expected to reach the set value. 11:15 (Peak shaving ends): Smoothly reduce the temperature of node A back to 83℃ at a rate not exceeding 0.3℃ / minute.

[0075] 204. Based on the pipeline network topology and medium characteristic parameters, construct a multi-timescale energy storage potential calculation model, take the optimal energy storage strategy evaluation report as input, and generate comparative data on energy storage performance under different operating conditions through thermodynamic simulation analysis. Specifically, based on the pipeline topology and medium characteristic parameters, a multi-timescale analysis framework including short-term, medium-term, and long-term analysis windows is established through a timescale partitioning algorithm, generating a timescale partitioning scheme with hierarchical characteristics. Taking the optimal energy storage strategy evaluation report and the timescale partitioning scheme as input, the thermal energy storage capacity and release characteristics at different timescales are calculated by establishing thermodynamic state transition equations, generating a thermodynamic state evolution sequence containing multi-time-dimensional characteristics. Based on the thermodynamic state evolution sequence, a dynamic scenario simulation mechanism is established to construct a multi-scenario simulation environment including normal operating conditions, extreme operating conditions, and peak-shaving operating conditions, generating comparative data on energy storage performance under different operating conditions. A multi-timescale analysis framework incorporating short-term, medium-term, and long-term analysis windows is established through a timescale partitioning algorithm. This includes: analyzing the thermal energy variation patterns at different timescales based on pipeline operation cycle data from historical operational databases and real-time peak-shaving demand characteristics using a time feature extraction module, generating a multi-scale partitioning basis containing time feature parameters; using the multi-scale partitioning basis and pipeline transmission characteristic parameters from a pipeline topology database as input, establishing short-term, medium-term, and long-term time windows through a dynamic time window partitioning algorithm, generating timescale partitioning parameters including time window boundaries and overlapping areas; and constructing a hierarchical multi-timescale analysis framework based on the timescale partitioning parameters by establishing a timescale correlation model, generating an analysis framework description file containing timescale transformation rules.

[0076] It should be noted that we have obtained an optimal energy storage strategy evaluation report from step 203. Its core recommendation is to raise the temperature of node A from 82℃ to 86℃ during the peak-shaving period from 11:00 to 11:15 on December 1st, thereby providing an additional 50MW of peak-shaving capacity to the grid. Now, we need to evaluate the energy storage potential of the pipeline network itself under this strategy.

[0077] Multi-scale classification criteria were generated based on historical data and system analysis. The analysis revealed that short-term (next 30 minutes) thermal energy changes are primarily driven by peak-shaving directives; medium-term (next 4 hours) changes are influenced by daily heat consumption patterns; and long-term (next 24 hours) changes are related to weather trends. These classification criteria were then combined with pipeline transmission characteristics (it takes approximately 7 minutes for thermal energy to travel from point A to point C).

[0078] Time window segmentation: A dynamic time window segmentation algorithm is used to establish three levels of time scales: Short-term: 11:00-11:30 (covering peak-shaving periods and their direct impact). Medium-term: 11:00-15:00 (covering midday peak heating consumption). Long-term: 11:00 on December 1st to 11:00 on December 2nd (analyzing the impact of the whole-day cycle).

[0079] Analytical framework construction: A hierarchical analytical framework was finally generated, which clarified the focus and transformation rules for different time scales. Short-term analysis focuses on instantaneous power response, while medium- and long-term analysis focuses on total energy balance and equipment fatigue.

[0080] Using the above timescale division scheme and the optimal energy storage strategy (node ​​A heating) as input, a thermodynamic state transition equation is established based on the pipeline topology (ABC structure) and medium parameters (specific heat capacity of water, pipeline volume) to calculate the thermal energy storage and release characteristics at different timescales.

[0081] The estimated equivalent energy storage capacity (Et) of the pipeline network is expressed as: Where ρ and c p It refers to the density and specific heat capacity of the medium, V. i The volume of the pipe segment associated with node i, ΔT i (t) represents the temperature change of node i at time t caused by the strategy. It is a time constant that reflects heat loss and embodies the node location (topology) and insulation characteristics.

[0082] Evolutionary Sequence Generation: Calculations yielded the following predicted sequence: Short-term: At 11:15, when peak shaving ended, the pipeline network stored an additional 500 kWh of thermal energy due to temperature rise; by 11:30, due to heat loss, this energy storage decreased to 480 kWh. Medium-term: Before 15:00, this portion of stored energy can be gradually released to support midday peak heating, and is expected to be fully released by 15:00. Long-term: Within a 24-hour cycle, this energy storage and release cycle has a relatively small impact on total energy consumption, but increases the number of thermal stress cycles in the pipeline network.

[0083] Based on the above evolutionary sequence, three simulation scenarios were constructed: Normal operating condition: Outdoor temperature 5℃, stable user load. After executing the strategy, energy storage release is smooth, and system efficiency is high. Extreme operating condition: Outdoor temperature drops sharply to -5℃, user load surges. Simulation shows that the 500kWh of thermal energy stored by the strategy can effectively supplement insufficient heating, reducing user-end temperature fluctuations by 1.5℃ and improving reliability. Peak-shaving operating condition: Frequent fluctuations in grid demand. Simulation shows that the strategy can effectively respond to a maximum of 3 similar peak-shaving commands within 2 hours, but the energy storage effect decreases by 15% on the third command.

[0084] Finally, comparative data on energy storage performance, including key indicators, is generated, as shown in the following example: 205. Design an adaptive calibration mechanism for prediction accuracy. By analyzing the deviation between real-time monitoring data and energy storage performance comparison data, calculate the correction amount of model parameters and output the updated thermal energy distribution prediction results for subsequent prediction cycles. Specifically, real-time monitoring of pipeline operation data is acquired through a data acquisition system, and deviation analysis is performed between this data and energy storage performance comparison data to generate a deviation feature vector containing the degree and distribution characteristics of the deviation. Based on the deviation feature vector, parameter sensitivity analysis is used to determine the influence of each parameter in the nonlinear heat conduction calculation model on the prediction accuracy, generating a parameter correction sequence containing parameter priority ranking. Using the parameter correction sequence as input, a recursive correction algorithm is established to calculate the specific correction amount of each parameter, generating a parameter correction set containing updated parameter values. Based on the parameter correction set, the nonlinear heat conduction calculation model is updated, and an updated heat energy distribution prediction model is generated through model validation testing, outputting the updated heat energy distribution prediction results.

[0085] It should be noted that the current time is 12:00 on December 1, 2023. Between 11:00 and 11:30, the system executed the optimal energy storage strategy generated in step 203 (i.e., increasing the temperature of node A to respond to grid peak shaving), and predicted the heat distribution of the pipeline network under this strategy based on the model in step 204 (energy storage performance comparison data). Now, we need to use actual operating data to verify the accuracy of the prediction and calibrate the model.

[0086] The system retrieves actual operating data from the data acquisition system for the period 11:00-11:30. Key data includes: the actual temperature of node A reached 85.8℃ at 11:15 (predicted value 86℃), and the temperature of node C reached 75.5℃ at 11:22 (predicted value 75℃, with an accurate prediction after a 7-minute delay). Deviation calculation: The actual data is compared with the predicted data from step 204. The main deviations are calculated as follows: the steady-state temperature deviation of node A is -0.2℃ (slightly lower than the prediction), and the temperature rise of node B is about 1 minute faster than predicted. A deviation feature vector is generated: this vector quantitatively describes the degree and distribution of the deviation: [average absolute temperature deviation: 0.15℃, maximum time deviation: 1 minute, deviation mainly distributed at node A].

[0087] Parameter sensitivity analysis: The nonlinear heat conduction calculation model contains several key parameters, including the equivalent heat transfer coefficient of the pipe wall and the flow rate correction factor of the medium. Analysis determined that the parameter with the greatest impact on the deviation of "node A not reaching the predicted temperature" is the equivalent heat transfer coefficient of the pipe wall (K), which directly affects the calculation of heat loss. The second most important parameter is the flow rate correction factor (Cv). Therefore, a parameter correction priority sequence is generated: 1. Heat transfer coefficient (K), 2. Flow rate correction factor (Cv).

[0088] A recursive correction algorithm is employed. This algorithm assumes that the correction amount should be proportional to the current deviation and inversely proportional to the historical stability of the parameter, in order to avoid overcorrection. The parameter correction formula is expressed as: Where: ΔP is the suggested correction amount for parameter P. Ecurrent is the current deviation value (for the physical quantity affected by parameter P, the temperature of node A). Shistorical is the stability index (variance) of this parameter over a period of time; a larger value indicates that the parameter is more unstable and the correction should be more cautious. λ is the learning rate, a constant between 0 and 1, which controls the correction step size.

[0089] Calculation process: For the heat transfer coefficient K, the current temperature deviation Ecurrent = -0.2℃, and historical data shows that K is very stable (Shistorical = 0.1). The learning rate λ is set to 0.5. Substituting into the formula, the calculation is as follows: This means that the K value should be adjusted downwards by about 0.09 to reflect the situation where the actual heat loss is slightly greater than the original model prediction.

[0090] Generate parameter correction set: get {heat transfer coefficient K: updated to the original value -0.09, flow rate correction factor Cv: kept unchanged}.

[0091] Model Update and Output: The nonlinear heat conduction calculation model was refreshed with the updated K value. The system was quickly validated using new data from 11:30 to 12:00, and it was found that the updated model reduced the prediction error of the temperature at node A by 60%. Finally, the system outputs this updated heat distribution prediction model for the next prediction cycle (targeting the peak-shaving demand on the afternoon of December 1st), thus achieving an adaptive improvement in prediction accuracy.

[0092] 206. Based on the optimal energy storage strategy evaluation report and energy storage performance comparison data, a comprehensive verification mechanism is established to cross-verify the prediction results with actual operating data, generating a verification analysis report that includes prediction accuracy indicators and reliability assessment. Based on the verification analysis report, an economic benefit evaluation model is established to calculate the peak-shaving revenue and operating costs under different energy storage strategies, generating a comprehensive benefit evaluation report that includes economic indicators. Using the comprehensive benefit evaluation report and future peak-shaving plans from the power grid dispatch center as input, an operation optimization suggestion generation mechanism is established to output a decision support report that includes suggestions for adjusting operating parameters and optimizing dispatch strategies.

[0093] It should be noted that the peak-shaving task from 11:00 to 11:30 on December 1st has been completed. The system now needs to evaluate the actual effectiveness of the implemented energy storage strategy.

[0094] Cross-validation: The system compares the predicted thermal energy distribution results from step 202, the energy storage performance comparison data from step 204, with the actual operating data. Prediction accuracy metrics: The target temperature for node A was set at 86℃, but it actually stabilized at 85.8℃, with an absolute error of 0.2℃ and an accuracy of 99.7%. The predicted time delay for temperature change at node C was 7 minutes, which was also 7 minutes, demonstrating complete accuracy. This indicates that the model is highly reliable in predicting spatiotemporal delays.

[0095] Reliability assessment: During the 15-minute peak-shaving period, the system's heating power remained stably above 98% of the target value (equivalent to 50MW) without drastic fluctuations, demonstrating the robustness of the strategy. The verification analysis report concludes that the predicted and optimized control in this study has high accuracy, and the system operates stably and reliably.

[0096] Establish an economic benefit assessment model: The model calculates the benefits and costs of this peak-shaving operation.

[0097] Peak shaving revenue: According to the ancillary service agreement signed with the power grid, revenue can be generated for successfully providing 50MW of peak shaving capacity for 15 minutes. The calculation method is: 50MW × 0.25 hours × service price of 100 yuan per megawatt-hour = 1250 yuan.

[0098] Operating costs: mainly due to the additional fuel consumption and increased pumping power consumption caused by the increased temperature, which is estimated to be about 300 yuan.

[0099] Economic indicators: Therefore, the net benefit of this energy storage peak shaving is 1250 yuan - 300 yuan = 950 yuan. The comprehensive benefit assessment report clearly points out that this strategy is economically positive.

[0100] Information integration: The comprehensive benefit assessment report (net benefit of 950 yuan, high reliability) and the future peak shaving plan released by the power grid dispatch center (similar peak shaving demand on December 2, 14:00-15:00) were used as inputs.

[0101] Operation optimization suggestion generation mechanism: System analysis indicates that the current strategy was successful and can serve as a benchmark template for similar tasks in the future. Simultaneously, operational parameter adjustment suggestions are proposed: Given that the actual temperature (85.8℃) is slightly lower than the target (86℃), it is recommended that the target temperature setting for node A be fine-tuned to 86.2℃ during the next execution to more accurately meet the power requirements.

[0102] Furthermore, the following optimization of the scheduling strategy is proposed: Since the system response was good as shown in this verification, it is recommended that the response time be shortened by another minute in future peak shaving tasks in order to obtain better grid service compensation.

[0103] Output Decision Support Report: The report ultimately provides the core recommendation: to approve and recommend the use of the optimized energy storage strategy (set temperature 86.2℃) in the peak shaving on December 2, which is expected to achieve economic benefits no less than those of this operation and further improve response performance.

[0104] In this embodiment of the invention, a comprehensive set of heating network operating parameters and power grid peak-shaving demand information is acquired through a multi-source data acquisition system. Combined with data preprocessing and spatiotemporal correlation analysis, a network state feature set including temperature gradient distribution, pressure field characteristics, and flow rate variation patterns is constructed, providing a rich and accurate data foundation for subsequent predictions. Simultaneously, an adaptive calibration mechanism for prediction accuracy is designed to automatically correct model parameters based on the deviation between real-time monitoring data and predicted data, further improving the accuracy of heat energy distribution prediction. A variational principle-based energy flow optimization calculation framework is established, transforming power grid peak-shaving demand into temperature constraints. Combined with heat energy distribution prediction results, an optimal energy storage strategy evaluation report is generated by solving the energy functional extremum problem. This strategy considers the multi-objective optimization of network thermodynamic characteristics and power grid peak-shaving demand, improving energy utilization efficiency, peak-shaving response speed, and operational stability while meeting peak-shaving requirements. Based on the pipeline network topology and medium characteristic parameters, a multi-timescale energy storage potential calculation model is constructed. Using the optimal energy storage strategy evaluation report as input, thermodynamic simulation analysis generates comparative data on energy storage performance under different operating conditions. Considering various operating conditions such as normal, extreme, and peak-shaving conditions, the model comprehensively evaluates the thermal energy storage capacity and release characteristics of the pipeline network at different time scales, providing a more comprehensive basis for decision-making. A comprehensive verification mechanism is established to cross-validate the prediction results with actual operating data, generating a verification analysis report including prediction accuracy indicators and reliability assessments. Based on this report, an economic benefit evaluation model is established to calculate the peak-shaving benefits and operating costs under different energy storage strategies, generating a comprehensive benefit evaluation report including economic indicators. Finally, using the comprehensive benefit evaluation report and future peak-shaving plans as input, a decision support report containing suggestions for adjusting operating parameters and optimizing scheduling strategies is output, providing a scientific and reliable decision-making basis for the operation of the power grid and heating pipeline network. The multi-scenario simulation environment and multi-timescale analysis framework constructed by this method enable heating networks to adapt to different operating conditions and timescale requirements. Whether it is short-term peak-shaving instructions, medium-term daily heat consumption patterns, or long-term weather trend impacts, accurate evaluation and optimized control can be achieved through corresponding analysis windows and simulation scenarios, improving the system's adaptability to complex and ever-changing power grid peak-shaving and heating demand. Based on the deviation analysis of real-time monitoring data and forecast results, as well as comprehensive benefit assessments and decision support reports, operators can flexibly adjust the operating parameters and scheduling strategies of the heating network. For example, based on the deviation between the actual temperature and the target temperature, the target temperature setpoint of point A can be fine-tuned; based on the system response, the response time can be shortened to strive for better power grid service compensation, etc., enabling the system to make optimal decisions in a timely manner according to the actual situation, improving operational flexibility and efficiency.

[0105] Figure 3 This is a schematic diagram of a heating network energy storage prediction device based on power grid peak shaving, provided by an embodiment of the present invention. The device 300 may include: a processor 301, a receiver 302, a transmitter 303, and a memory 303. The receiver 302, transmitter 303, and memory 303 are respectively connected to the processor 301 via a bus. It should be noted that in some possible implementations, the processor 301 and the memory 303 may be integrated together.

[0106] The processor 301 includes one or more processing cores. The processor 301 executes the methods performed by the base station in the random access method provided in this application embodiment by running software programs and modules. The memory 304 can be used to store software programs and modules. Specifically, the memory 304 can store an operating system 3041 and at least one application module 3042 required for a function. The receiver 302 is used to receive communication data sent by other devices, and the transmitter 303 is used to send communication data to other devices.

[0107] The present invention also provides a heating network energy storage prediction device based on power grid peak shaving. The heating network energy storage prediction device based on power grid peak shaving includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor performs the steps of the heating network energy storage prediction method based on power grid peak shaving in the above embodiments.

[0108] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the heating network energy storage prediction method based on grid peak shaving.

[0109] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0110] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0111] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A grid peak shaving based heat supply network energy storage prediction method, characterized in that, include: By acquiring heating network operation parameters and power grid peak-shaving demand information through a multi-source data acquisition system, a network status feature set containing temperature gradient distribution, pressure field characteristics, and flow change patterns is generated. A nonlinear heat conduction calculation model is constructed based on the first principles of thermodynamics. The network state feature set is used as input. By introducing the historical dependence and spatial correlation characteristics of thermal inertia, the dynamic response parameters of network heat energy transmission are calculated, and the predicted heat energy distribution results including time delay effects are output. A variational principle-based energy flow optimization calculation framework is established, which transforms the power grid peak-shaving demand information into temperature constraints, takes the thermal energy distribution prediction results as input, and generates an optimal energy storage strategy evaluation report by solving the energy functional extremum problem. Based on the pipeline network topology and medium characteristic parameters, a multi-timescale energy storage potential calculation model is constructed. The optimal energy storage strategy evaluation report is used as input, and thermodynamic simulation analysis is used to generate comparative data on energy storage performance under different operating conditions. An adaptive calibration mechanism for prediction accuracy is designed. By analyzing the deviation between real-time monitoring data and the energy storage performance comparison data, the parameter correction amount of the nonlinear heat conduction calculation model is calculated, and the updated heat energy distribution prediction result is output for subsequent prediction cycles.

2. The method for predicting energy storage in heating networks based on power grid peak shaving according to claim 1, characterized in that, include: Based on the temperature gradient distribution and flow rate variation patterns in the pipeline network status feature set, and combined with the temperature variation trends in historical operation data, the historical dependence parameters of thermal inertia are obtained, and a thermal inertia feature vector is generated. Based on the pressure field characteristics and pipeline topology data in the pipeline network state feature set, the heat conduction correlation degree of each node in the pipeline network is analyzed to obtain spatial correlation characteristic parameters and generate a spatial correlation model. Based on the thermal inertia feature vector and spatial correlation model, the dynamic response parameters of heat energy transmission in the pipeline network are obtained, and the predicted results of heat energy distribution are output.

3. The method for predicting energy storage in heating networks based on grid peak shaving according to claim 2, characterized in that, Temperature change sequences for multiple consecutive operating cycles are extracted from the historical operating data. Time series analysis is used to identify the periodic patterns of temperature changes and generate a temperature change pattern library. Based on the temperature change pattern library, the decay characteristics of temperature change at different time scales are calculated through thermal inertia effect analysis, and a set of thermal inertia characteristic parameters is generated. The thermal inertia characteristic parameter set is fused with the real-time acquired temperature gradient distribution data to obtain a comprehensive thermal inertia history dependency parameter.

4. The method for predicting energy storage in heating networks based on power grid peak shaving according to claim 2, characterized in that, include: Based on the peak-shaving demand time-series signal from the power grid dispatch center, the peak-shaving power demand is transformed into a temperature change range constraint of the heating network, generating a set of temperature boundary conditions. Based on the predicted thermal energy distribution and the set of temperature boundary conditions, a multi-objective optimization calculation model is constructed by establishing an energy functional extremum problem, and an optimization solution framework is generated. Based on the aforementioned optimization framework, the optimal trajectory of energy distribution is solved using the variational method, resulting in the optimal temperature control strategy during different peak-shaving periods, and generating an optimal energy storage strategy evaluation report.

5. The method for predicting energy storage in heating networks based on grid peak shaving according to claim 4, characterized in that, Based on the set of temperature boundary conditions and the predicted thermal energy distribution, the allowable range and rate of temperature change are identified by the constraint condition analysis module, and a set of thermodynamic constraint conditions is generated. Based on the peak-shaving demand time-series signal from the power grid dispatch center, a multi-objective optimization function including energy utilization efficiency, peak-shaving response speed and operation stability indicators is established through the objective function construction module, and a set of objective function components is generated. Based on the set of thermodynamic constraints and the set of objective function components, a comprehensive optimization model incorporating thermodynamic characteristics and power grid peak-shaving requirements is constructed using a multi-objective optimization algorithm, generating an optimization solution framework.

6. The method for predicting energy storage in heating networks based on power grid peak shaving according to claim 4, characterized in that, include: Based on the pipeline network topology and media characteristic parameters, a multi-timescale analysis framework including short-term, medium-term and long-term analysis windows is established, and a timescale division scheme is generated. Based on the optimal energy storage strategy evaluation report and the time scale division scheme, the thermal energy storage capacity and release characteristics at different time scales are obtained, and a thermodynamic state evolution sequence is generated. Based on the aforementioned thermodynamic state evolution sequence, a dynamic scenario simulation mechanism is established to construct a multi-scenario simulation environment that includes normal operating conditions, extreme operating conditions, and peak-shaving operating conditions, thereby generating energy storage performance comparison data.

7. The method for predicting energy storage in heating networks based on power grid peak shaving according to claim 6, characterized in that, The estimated equivalent energy storage capacity of the pipeline network is: ,but: Where ρ and c p It refers to the density and specific heat capacity of the medium, V. i The volume of the pipe segment associated with node i, ΔT i (t) represents the temperature change of node i at time t caused by the strategy. It is a time constant that reflects heat loss.

8. The method for predicting energy storage in heating networks based on grid peak shaving according to claim 6, characterized in that, Based on the pipeline operation cycle data in the historical operation database and the peak-shaving demand characteristics collected in real time, the heat energy change pattern under different time scales is analyzed through the time feature extraction module to generate a multi-scale division basis. Based on the multi-scale division criteria and the pipeline transmission characteristic parameters from the pipeline topology database, a time window with three levels—short-term, medium-term, and long-term—is established through a dynamic time window division algorithm to generate time scale division parameters. Based on the time scale division parameters, a multi-time scale analysis framework with hierarchical characteristics is constructed by establishing a time scale correlation model, and an analysis framework description file is generated.

9. The method for predicting energy storage in heating networks based on grid peak shaving according to claim 6, characterized in that, include: Obtain real-time monitoring data of pipeline operation, compare it with the energy storage performance data, perform deviation analysis, and generate a deviation feature vector; Based on the aforementioned deviation feature vector, the degree of influence of each parameter in the nonlinear heat conduction calculation model on the prediction accuracy is determined through parameter sensitivity analysis, and a parameter correction sequence is generated. Based on the parameter correction sequence, a recursive correction algorithm is established to calculate the specific correction amount of each parameter, thereby generating a parameter correction set; Based on the parameter correction set, the parameters of the nonlinear heat conduction calculation model are updated, and the updated heat energy distribution prediction model is generated through model verification test, and the updated heat energy distribution prediction result is output.

10. The method for predicting energy storage in heating networks based on power grid peak shaving according to claim 1, characterized in that, Also includes: Based on the optimal energy storage strategy evaluation report and energy storage performance comparison data, a comprehensive verification mechanism is established to cross-verify the prediction results with the actual operation data and generate a verification analysis report. Based on the aforementioned verification analysis report, an economic benefit assessment model is established to calculate the peak-shaving revenue and operating costs under different energy storage strategies, thereby generating a comprehensive benefit assessment report. Based on the comprehensive benefit assessment report and the future peak-shaving plan from the power grid dispatch center, a decision support report is output by establishing an operation optimization suggestion generation mechanism.

Images