An urban energy network intelligent deployment method and system

Abstract

The application discloses a kind of urban energy network intelligent deployment method and system, by obtaining building energy consumption data and energy supply network topology information and establishing supply-demand matching benchmark, identify load peak dense period and peak clipping potential area, export transferable load data to form regional supply-demand distribution map;Combining user response delay length and equipment start-stop cycle generates peak-shaving regulation space;For regional supply-demand distribution map and peak-shaving regulation space, carry out gap matching analysis, create hierarchical control instruction set and distinguish to generate immediate response instruction and delay response instruction;Through peak-shaving benefit coefficient and immediate response instruction collaborative configuration form control collaborative parameter;Finally determine control execution range and implement cost benefit evaluation, form urban energy scheduling execution scheme, realize the accurate matching and efficient scheduling of urban energy supply and demand, can provide comprehensive, efficient scheduling decision support for smart grid, regional heating, integrated energy services and other application scenarios.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy management, in particular to a city energy network intelligent deployment method and system. BACKGROUND

[0002] The city energy network undertakes the tasks of power, heat, gas and other energy transmission and distribution, and is the key infrastructure supporting the normal operation of the city. With the continuous expansion of the city size and the improvement of the residents' living standards, the total energy consumption is rising, and the peak-valley difference of energy consumption is widening. In some areas, there is a shortage of supply during peak periods, while the supply capacity is idle during off-peak periods. The traditional energy deployment mainly relies on dispatchers to arrange supply plans according to historical experience and simple load curves. This method has a relatively slow response to load changes, cannot deeply tap the flexible adjustment capacity of the user side, and cannot accurately grasp the operation characteristics of various energy-using equipment, resulting in low overall operation efficiency of the energy network.

[0003] Some existing technologies attempt to improve the deployment effect through load prediction, but these methods generally separate energy consumption data analysis from network deployment, lack systematic modeling of the correlation between the two, resulting in a large deviation between the prediction results and the actual network operation state. In addition, although some studies focus on the value of user-side adjustable load, there is still no clear identification and utilization method for issues such as which areas have peak shaving potential, which time periods are suitable for implementing adjustment, and the response speed differences of different users. More importantly, current control strategies often treat all adjustable resources equally, without classifying and managing resources according to their response speed and economic characteristics, which cannot meet the demand for fast response in emergency situations and cannot fully utilize the economic advantages of delayed adjustment. SUMMARY

[0004] The present application discloses a city energy network intelligent deployment method and system, which aims to accurately identify load fluctuation characteristics and peak shaving potential areas by deeply mining the correlation between building energy consumption data and energy supply network topology information, determine the peak shaving adjustment space in combination with user response delay time and equipment start-stop cycle, generate a set of hierarchical control instructions based on gap matching analysis, realize the collaborative configuration of immediate response instructions and delayed response instructions, and finally form a city-level energy scheduling execution scheme covering multiple nodes, multiple time periods and multiple energy types, providing comprehensive technical support for the intelligent and accurate control of city energy systems.

[0005] The present application discloses a city energy network intelligent deployment method and system, which aims to accurately identify load fluctuation characteristics and peak shaving potential areas by deeply mining the correlation between building energy consumption data and energy supply network topology information, determine the peak shaving adjustment space in combination with user response delay time and equipment start-stop cycle, generate a set of hierarchical control instructions based on gap matching analysis, realize the collaborative configuration of immediate response instructions and delayed response instructions, and finally form a city-level energy scheduling execution scheme covering multiple nodes, multiple time periods and multiple energy types, providing comprehensive technical support for the intelligent and accurate control of city energy systems.

[0006] Obtain building energy consumption data and energy supply network topology information, identify load fluctuation characteristics based on the energy consumption data, and establish a supply and demand matching benchmark between the load fluctuation characteristics and the energy supply network topology information;

[0007] Based on the supply and demand matching benchmark, the peak load period is detected, the peak shaving and valley filling potential area is extracted from the peak load period, the transferable load data is exported from the peak shaving and valley filling potential area, and a regional supply and demand distribution map is formed based on the transferable load data;

[0008] The system acquires user response delay duration and device start-stop cycle, identifies time elasticity intervals through the response delay duration, captures adjustment switching times based on the device start-stop cycle, and performs time-series comparison between the time elasticity interval and the adjustment switching times to generate peak-shaving adjustment space.

[0009] Gap matching analysis is performed on the regional supply and demand distribution map and the peak-shifting adjustment space to generate gap compensation parameters. The control target node is established through the gap compensation parameters, and a hierarchical control instruction set is created based on the control target node.

[0010] The hierarchical control instruction set is divided into immediate response instructions and delayed response instructions according to the response time difference. The delayed response instructions are used to identify the fluctuation adjustment benefit and obtain the peak-shaving benefit coefficient. The peak-shaving benefit coefficient and the immediate response instructions are coordinated to form control coordination parameters.

[0011] Based on the aforementioned regulation coordination parameters and the regional supply and demand distribution map, the scope of regulation implementation is determined, and a cost-benefit assessment is conducted on the scope of regulation implementation to form a city-level energy dispatch implementation plan.

[0012] A second aspect of this invention provides an intelligent dispatching system for urban energy networks, comprising:

[0013] The energy consumption acquisition module is used to acquire building energy consumption data and energy supply network topology information, identify load fluctuation characteristics based on the energy consumption data, and establish a supply and demand matching benchmark between the load fluctuation characteristics and the energy supply network topology information.

[0014] The load analysis module is used to detect periods of concentrated load peaks based on the supply and demand matching benchmark, extract potential areas for peak shaving and valley filling from the periods of concentrated load peaks, export transferable load data from the potential areas for peak shaving and valley filling, and form a regional supply and demand distribution map based on the transferable load data.

[0015] The timing matching module is used to obtain the user response delay duration and the device start-stop cycle, identify the time elasticity interval through the response delay duration, capture the adjustment switching time according to the device start-stop cycle, and perform a timing comparison between the time elasticity interval and the adjustment switching time to generate a peak-shaving adjustment space.

[0016] A regulation strategy module is configured to perform gap matching analysis on the regional supply-demand distribution map and the peak-shaving regulation space to generate gap compensation parameters, establish a regulation target node based on the gap compensation parameters, and create a hierarchical regulation instruction set based on the regulation target node;

[0017] An instruction coordination module is configured to distinguish between immediate response instructions and delayed response instructions according to response time differences based on the hierarchical regulation instruction set, identify and obtain a peak-shaving benefit coefficient based on fluctuation regulation benefits of the delayed response instructions, and form regulation coordination parameters based on coordination of the peak-shaving benefit coefficient and the immediate response instructions.

[0018] A scheme generation module is configured to determine a regulation execution range based on the regulation coordination parameters and the regional supply-demand distribution map, perform cost-benefit evaluation on the regulation execution range, and form a city-level energy dispatch execution scheme.

[0019] The beneficial effects of the present application are embodied in the following points: 1. By fusing the building energy behavior characteristics with the network topology structure, the supply and demand balance nodes and unbalanced nodes in the network are identified, the capacity of the balance nodes and the gap of the unbalanced nodes are analyzed, the supply and demand deviation quantification index is generated, and the supply and demand correlation benchmark is established. On this basis, by detecting the peak load intensive period, the peak value is continuously identified and the peak starting position is determined, the pre-regulation ability before the peak is obtained, the peak load shaving and valley filling potential area is divided, and the transferable load data is derived, and the regional supply and demand distribution situation is formed. This technical path which combines supply and demand correlation modeling and potential mining makes the adjustment decision based on accurate supply and demand state and clear resource distribution, improves the pertinence of adjustment, and solves the problem that the traditional method separates the energy consumption data analysis and adjustment decision, resulting in the disconnection between the adjustment instruction and the actual supply and demand state. 2. By obtaining the user response delay time length, the delay tolerance threshold is generated according to the delay time length, the response delay is upper limited to generate an elastic reserve constraint delay section, and the stability of the section is tested to determine a stable delay period, and the maximum continuous time length is extracted as a time elasticity interval. At the same time, the adjustment switching time is captured from the device start-stop cycle data, the time elasticity interval is compared with the adjustment switching time in time sequence, and the peak shifting regulation space which meets the user response ability and conforms to the device operation rule is generated, which provides a time dimension constraint condition for the accurate issuance of the adjustment instruction, and ensures the executability of the control measures. 3. The gap interaction transmission is tracked to identify the gap trigger chain, the gap trigger chain is compensated and damaged with the peak shifting regulation space, the gap compensation parameters are determined according to the compensation dissipation boundary, and the control target node is established, and the hierarchical control instruction set is created. The hierarchical control instruction set is divided into immediate response instruction and time-delay response instruction according to the response time difference, the time-delay response instruction is analyzed by the peak-valley difference of the load transfer vector to determine the peak shifting benefit coefficient, and the immediate response instruction is configured according to the peak shifting benefit coefficient. The control execution range is determined by the spatial superposition of the control coordination parameters and the regional supply and demand distribution graph, the total control cost and benefit are counted, and the benefit-cost ratio is calculated, the technical feasibility and execution reliability are comprehensively evaluated, the urban energy scheduling execution scheme is formed, the rapid response ability is ensured, the economic advantage of time-delay response is fully utilized, and the economy and feasibility of the scheme are ensured. BRIEF DESCRIPTION OF DRAWINGS

[0020] The drawings herein show specific examples of the technical solutions described in the present application, and constitute part of the specification together with the specific embodiments, for explaining the technical solutions, principles and effects of the present application.

[0021] Unless specifically stated or defined otherwise, the same reference signs in different drawings represent the same or similar technical features, and different reference signs may also be used to represent the same or similar technical features.

[0022] Figure 1 is a flow diagram of an urban energy network intelligent deployment method of the present application.

[0023] Figure 2 is a structural block diagram of an urban energy network intelligent deployment system of the present application. DETAILED DESCRIPTION

[0024] In the following description, for purposes of explanation and not limitation, specific details are set forth, such as a particular sequence of steps, in order to provide a thorough understanding of the present application. However, it will be apparent to one skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, and circuits are omitted so as not to obscure the description of the present application with unnecessary detail.

[0025] It is to be understood that the terminology "including", "comprising", "consisting" and "consisting essentially of" used in the specification and the appended claims, indicates the presence of the stated features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0026] Reference throughout this specification to "one embodiment" or "an embodiment" or "some embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. Thus, the appearances of the phrases "in one embodiment" or "in some embodiments" or "in other embodiments" or "in additional embodiments" in various places throughout this specification are not necessarily referring to the same embodiment, unless otherwise specified. The terms "including", "comprising", "consisting of" and "consisting essentially of" mean "including but not limited to", unless otherwise specified.

[0027] The technical solutions of the embodiments of the present application are introduced as follows.

[0028] As shown in Figure 1 The present application provides an urban energy network intelligent deployment method, which comprises the following steps S110-S160:

[0029] In step S110, building energy consumption data and energy supply network topology information are acquired, load fluctuation characteristics are identified according to the energy consumption data, and the load fluctuation characteristics and the energy supply network topology information are used to establish a supply-demand matching benchmark.

[0030] Specifically, building energy consumption data and energy supply network topology information are acquired. Building energy consumption data of various types of buildings are collected from the city energy management platform, including power consumption, heating demand, gas consumption, and cooling load of various energy forms. Building energy consumption data shows that the power load of a commercial complex during the peak period of weekdays is about twice that of the same period on weekends, showing obvious time regularity. Energy consumption characteristics of different building types are extracted from the energy consumption data, including residential communities, commercial buildings, industrial plants, and public facilities. Energy consumption data shows that the heating demand of a residential complex has a double-peak feature in the morning and evening during winter, while the electricity load of an office building is mainly concentrated in the working hours. The energy supply network topology information of the energy supply network is acquired, including the location of energy supply sites, the connection relationship of the pipe network, the capacity of the pipe network, and the transmission distance. Topology information shows that the heating pipe network in a city forms a central heat source station connected to several heat exchange stations through the main pipe network, and the network structure of each heat exchange station covers the surrounding area. Key nodes are identified from the energy supply network topology information, including energy stations, hub heat exchange stations, branch regulating valves, and end-user access points. Topology information shows that the regional power grid forms a double-ring power supply structure consisting of several high-voltage substations and several medium-voltage substations.

[0031] Load fluctuation characteristics are identified based on energy consumption data. The time series variation of energy consumption data is analyzed, and daily load curve, weekly load curve, and seasonal load curve are extracted from energy consumption data. Energy consumption data shows that the daily load curve of a commercial area presents three peak characteristics in the morning, noon, and evening. The amplitude and frequency parameters of load fluctuation characteristics are calculated, and the fluctuation degree is quantified by the ratio of maximum load, minimum load, and average load in energy consumption data. Energy consumption data shows that the peak-valley difference of power load in an industrial park is large, and the fluctuation amplitude of load fluctuation characteristics is significantly higher than that of residential areas. Fourier transform analysis is performed on energy consumption data to extract the dominant frequency component and amplitude distribution of load fluctuation characteristics. Frequency spectrum analysis of energy consumption data shows that the heating load of a residential area has a main frequency period of 24 hours, corresponding to the diurnal temperature variation, and a secondary frequency period of one week, corresponding to the difference between weekdays and weekends. The occurrence frequency and intensity of load mutation events in energy consumption data are counted, including load jump caused by sudden weather changes and load anomaly caused by equipment failure. Energy consumption data records show that the heating load of a region increases significantly in a short time during a cold wave, forming a load mutation event. Future load changes are predicted based on historical fluctuation characteristics of energy consumption data. Load fluctuation characteristics include time series regularity, fluctuation parameters, frequency components, and prediction data.

[0032] In some embodiments, the establishing of the supply-demand matching reference based on the load fluctuation feature and the energy supply network topology information comprises: identifying supply-demand balanced nodes and supply-demand unbalanced nodes according to the matching relationship between the load fluctuation feature and the energy supply network topology information; generating a supply-demand deviation value by capacity matching the surplus of the supply-demand balanced nodes and the gap of the supply-demand unbalanced nodes; determining a capacity adjustment threshold according to the supply-demand deviation value; and applying the capacity adjustment threshold to the energy supply network topology information to establish the supply-demand matching reference.

[0033] The supply-demand balanced nodes and the supply-demand unbalanced nodes are identified according to the matching relationship between the load fluctuation feature and the energy supply network topology information. A supply-demand balance index is calculated for each node in the energy supply network topology information, and the balance index calculation formula is B = 1 - |C_s - C_d| / C_d, where B is the balance index, C_s is the node supply capacity, and C_d is the node demand load. A supply-demand balance determination threshold is set, and the nodes with a balance index higher than the threshold are determined as supply-demand balanced nodes, and the nodes with a balance index lower than the threshold are determined as supply-demand unbalanced nodes. A reasonable balance threshold is set for a certain city, and the nodes in the energy supply network topology information are divided into two categories: supply-demand balanced nodes and supply-demand unbalanced nodes. In combination with the prediction data of the load fluctuation feature, the balance of each node in the energy supply network topology information in the future period is pre-judged. A certain node is currently a supply-demand balanced node, but the prediction of the load fluctuation feature shows that it will become a supply-demand unbalanced node in the evening peak period, and is marked as a potential unbalanced node in advance. The supply-demand balanced nodes and the supply-demand unbalanced nodes are identified by different colors on the network diagram of the energy supply network topology information, forming a supply-demand balance situation diagram. The situation diagram of a certain region shows that most of the nodes in the energy supply network topology information are in a balanced state, and a small number of nodes are in a supply-demand imbalance. The distribution changes of the supply-demand balanced nodes and the supply-demand unbalanced nodes in different periods are counted, and based on the time sequence law of the load fluctuation feature, the number of supply-demand unbalanced nodes in a certain city reaches a peak in the morning peak period of weekdays, and the number of supply-demand unbalanced nodes decreases to the lowest in the night period.

[0034] The capacity of the surplus of the supply-demand balance node and the gap of the supply-demand imbalance node are matched to generate a supply-demand deviation value. The adjustable capacity of the supply-demand balance node is calculated, which refers to the remaining supply capacity of the node after ensuring its own supply-demand balance. A certain supply-demand balance node still has a certain surplus that can be used to support other nodes after retaining the necessary safety margin. The supply gap of the supply-demand imbalance node is calculated, which refers to the part of the node's demand load that exceeds its supply capacity. The demand load of a certain supply-demand imbalance node significantly exceeds its supply capacity, resulting in a large supply gap. The spatial distance and connection path between the supply-demand balance node and the supply-demand imbalance node are evaluated, and the spatial distance and path constraints affect the feasibility and loss of energy allocation. A certain supply-demand balance node is connected to a supply-demand imbalance node through a transmission and distribution pipeline network, and the pipeline network has a certain transmission efficiency. The capacity matching analysis of the surplus and the gap is carried out, and the transmission distance, pipeline capacity and transmission loss are considered in the matching. A certain region has multiple supply-demand balance nodes that can provide surplus, which needs to be reasonably allocated to the supply-demand imbalance nodes according to the transmission efficiency. Based on the capacity matching analysis, the supply-demand deviation value is calculated. The supply-demand deviation value quantifies the overall supply-demand imbalance of the region, and the calculation formula is D = Σ(gap_i) - Σ(surplus_j x transmission efficiency_j), where D is the supply-demand deviation value, gap_i is the gap of the i-th supply-demand imbalance node, surplus_j is the surplus of the j-th supply-demand balance node, and transmission efficiency_j is the corresponding transmission efficiency. The supply-demand deviation value is positive, indicating that the overall supply is insufficient, and negative, indicating that there is surplus in the supply.

[0035] The capacity allocation threshold is determined according to the supply-demand deviation value. First, the distribution characteristics of the supply-demand deviation value are analyzed to identify the mean, variance and extreme value of the deviation value. The supply-demand deviation value of a certain urban area fluctuates within a week, with a larger value during peak hours. The operational risk under different supply-demand deviation value levels is evaluated, and the operation is stable when the supply-demand deviation value is small, and energy supply interruption or equipment overload may occur when the supply-demand deviation value is large. When the supply-demand deviation value of a certain region exceeds a certain limit, the pipeline pressure decreases significantly, and the heating quality is affected. The trigger condition for capacity allocation is determined based on the ratio of the supply-demand deviation value to the total capacity. The allocation is triggered when the supply-demand deviation value of a certain heating network exceeds a certain percentage of the total capacity. The dynamic adjustment of the capacity allocation threshold is set considering the load characteristics of different seasons and periods. A certain city uses a higher capacity allocation threshold during the winter heating period, appropriately reduces the threshold in the transition season, and adjusts according to the actual demand in the summer cooling period.

[0036] The capacity allocation threshold is applied to the energy supply network topology information to establish a supply-demand matching benchmark. The capacity allocation threshold is marked at each node of the energy supply network topology information to form a node-level allocation trigger mechanism. Each heat exchange station of a certain heat network sets an individualized capacity allocation threshold according to its own supply-demand state. Stations with high loads use higher thresholds, and stations with low loads use lower thresholds. A supply-demand matching evaluation system based on the capacity allocation threshold is established, which includes three dimensions: node matching degree, regional balance degree, and network stability degree. The supply-demand matching benchmark quantifies the degree of supply-demand matching through the evaluation system. A higher node matching degree in a certain region indicates that most nodes are in a balanced supply-demand state. A good regional balance degree indicates that the overall supply-demand coordination of the region is good. A network stability degree at a reasonable level indicates that the network is stable. Dynamic adjustment rules for the supply-demand matching benchmark are defined. The rules update the matching benchmark in real time according to the prediction of load fluctuation characteristics and the change of supply capacity. During the seasonal transition period of a certain city, the supply-demand matching benchmark is adjusted in advance based on temperature forecasts and predictions of load fluctuation characteristics. The capacity allocation threshold is gradually increased before the start of the heating period.

[0037] In step S120, the load peak intensive period is detected based on the supply-demand matching benchmark. The peak shaving and valley filling potential region is extracted from the load peak intensive period. The transferable load data is derived through the peak shaving and valley filling potential region. The regional supply-demand distribution map is formed based on the transferable load data.

[0038] Specifically, the load peak intensive period is detected based on the supply-demand matching benchmark. From the ideal supply-demand state of each node defined in the supply-demand matching benchmark, the period in which the actual load exceeds the matching benchmark is identified. The actual load of a certain commercial district exceeds the supply-demand matching benchmark by more than 20% during two periods: 9:00-11:00 in the morning and 18:00-20:00 in the evening. It is determined to be a peak period. The frequency and distribution of peak periods are counted. The load peak intensive period is identified based on the deviation of the supply-demand matching benchmark. The heating system of a certain city in winter has a continuous load peak during the morning from 6:00 to 8:00 and the evening from 17:00 to 22:00, forming a load peak intensive period. The peak intensity of the load peak intensive period is calculated. The peak intensity calculation formula is I=N_p / (T×ΔT), where I is the peak intensity, N_p is the number of peaks, T is the observation period, and ΔT is the total duration of the peak period. A certain region has a high peak intensity in the load peak intensive period, with a total of 15 peaks in a week of observation. The characteristics of the load peak intensive period of different regions and different energy types are analyzed. The heating peak of residential areas is mainly concentrated in the morning and evening, the electricity peak of commercial areas is mainly concentrated in the working hours, and the load peak of industrial areas is relatively dispersed.

[0039] In some embodiments, the extracting the peak shaving potential region from the load peak dense period comprises: performing peak duration identification on the load peak dense period to generate peak duration data; determining a peak starting position according to the peak duration data; obtaining a pre-regulation adjustable load capacity before the peak by the peak starting position; and dividing the peak shaving potential region according to the pre-regulation adjustable load capacity before the peak.

[0040] The peak duration identification on the load peak dense period generates peak duration data. The shape characteristics of the load curve in the load peak dense period are analyzed to identify the duration and trend of the peak. The evening peak load of a certain commercial district starts to rise at 18:00, reaches the peak at 19:00, and continues to run at a high level until 20:30, with a peak duration of about 2.5 hours. The peak duration of different dates and different regions is counted, and the statistical results are recorded in the peak duration data. The average duration of the morning peak in a certain residential area on weekdays is 1.5 hours, and the duration of the weekend morning peak is extended to 2 hours. These data are recorded in the peak duration data. The influencing factors of peak duration are identified, including weather conditions, user behavior habits, and holiday effects. In a certain city, the peak duration data shows that the duration of the heating load peak in extremely cold weather is more than 50% longer than in normal weather. The stability index of the peak duration is calculated, which reflects the predictability of the peak duration. The peak duration data of a certain industrial park shows that the duration is relatively stable and the variance is small, which is convenient for formulating control schemes in advance. Key parameters of the peak duration are extracted, including the peak start time, the peak end time, the duration, and the peak intensity.

[0041] For example, the peak starting position is determined according to the peak duration data, which comprises: analyzing the duration change key point in the peak duration data; determining the available peak prediction time at the duration change key point; evaluating the prediction accuracy of the peak prediction time to generate an accuracy evaluation result; and generating a peak starting position according to the peak prediction time and the accuracy evaluation result.

[0042] The time length change key point in the peak duration data is analyzed. The first derivative and second derivative of the load curve are extracted from the peak duration data to identify the sudden change point of the load change rate. The first derivative of the load curve in the peak duration data of a certain area shows that at 8:00, the first derivative suddenly increases from close to zero to 0.2 MW / hour, indicating that the load enters a rapid climbing stage, and this time is the time length change key point. The distribution of time length change key points on different days in the peak duration data is analyzed to identify the time regularity of the key points. The peak duration data of a certain commercial area shows that the time length change key points of weekdays are concentrated between 8:00-9:00, and those of weekends are delayed to 9:00-10:00. The time difference between the time length change key point and the peak arrival time is calculated, and the time difference reflects the duration from the start of load climbing to the peak. The peak duration data of a certain residential area shows that it takes an average of 1.5 hours from the time length change key point to the peak arrival, and this time window can be used to implement pre-regulation measures. External factors affecting the time length change key point are identified, including temperature changes, solar intensity, and user activity patterns. The peak duration data of a certain city shows that when the temperature drops suddenly, the time length change key point will appear earlier, and the load climbing speed will increase.

[0043] The available peak prediction time is determined at the time length change key point. The time length change key point is used as the trigger signal for peak prediction, and load forecasting and pre-regulation preparation are started at this time. A certain heating system starts load forecasting immediately after detecting the time length change key point and predicts the load change in the next 2 hours. Combined with historical data and real-time monitoring information, the peak arrival time is calculated at the time length change key point. The prediction of a certain area shows that from the current time length change key point, the load will reach the peak in 80-100 minutes. The advance amount of the peak prediction time is set, which needs to be long enough to implement pre-regulation measures, but not too long to increase the prediction error. A certain city sets the peak prediction time to 60 minutes before the predicted peak arrival time to ensure enough time to start the energy storage device and adjust the controllable load. Considering the response time characteristics of different types of loads, different peak prediction times are set for different adjustment methods. A certain energy storage responds quickly and can be started 30 minutes before the peak; a building pre-cooling needs a longer time and needs to be started 90 minutes before the peak.

[0044] The prediction accuracy of the peak prediction time is evaluated to generate an accuracy evaluation result. Historical prediction data and actual peak arrival data are collected, and prediction errors are calculated. The prediction error calculation formula is E = |T_prediction - T_actual|, where E is the prediction error, T_prediction is the predicted peak arrival time, and T_actual is the actual peak arrival time. The average prediction error of a certain prediction in the past 30 days is 12 minutes. The distribution characteristics of the prediction error are counted, and the statistical results are recorded in the accuracy evaluation result, including the mean, standard deviation, and maximum error. The accuracy evaluation result shows that the standard deviation of the prediction error of a certain prediction is 8 minutes, and 95% of the prediction errors are within 25 minutes. The prediction accuracy differences under different weather conditions and load types are analyzed. The accuracy evaluation result shows that the prediction error is smaller in sunny weather and larger in extreme weather; the prediction accuracy is higher for regular loads and lower for random loads. The prediction accuracy score is calculated, which considers the size and stability of the prediction error. The prediction accuracy score in the accuracy evaluation result is 85, which belongs to a high level. The confidence interval of the prediction accuracy is established, which reflects the credibility of the prediction result. The accuracy evaluation result shows that the 95% confidence interval of a certain prediction time is ±20 minutes, indicating that the actual peak has a 95% probability of arriving within 20 minutes before and after the prediction time.

[0045] The peak starting position is generated according to the peak prediction time and the accuracy evaluation result. First, the peak starting position time range is determined by integrating the peak prediction time and the accuracy evaluation result. A certain system determines the peak starting position time range to be 8:45-9:15 according to the prediction time of 9:00 and the confidence interval of ±15 minutes in the accuracy evaluation result. A dynamic monitoring mechanism is set up within the peak starting position time range to track load changes in real time to correct the prediction result. A certain system starts high-frequency monitoring at the peak starting position of 8:45, updates the load forecast every 5 minutes, and adjusts the peak starting position judgment according to the latest data. Different response strategies are set according to the accuracy evaluation result, accurate regulation is used when the accuracy is high, and conservative strategies are used when the accuracy is low. A certain system starts pre-regulation according to the peak starting position when the score in the accuracy evaluation result is higher than 80; when the score is lower than 60, pre-regulation is started 30 minutes in advance to cope with uncertainty. The peak starting position is associated with the pre-regulation trigger condition, and the corresponding control measures are automatically started when the load reaches the peak starting position.

[0046] The pre-peak pre-regulation adjustable load capacity is obtained by the peak starting position. In the period before the peak starting position, the load resources available for pre-regulation are identified. Pre-peak pre-regulation shifts part of the load from the peak period to the pre-peak period by starting the adjustable load in advance. A certain commercial building starts air conditioning pre-cooling one hour before the peak according to the peak starting position, reducing the cooling load demand in the peak period. The response capacity of the adjustable load is evaluated, including the regulation capacity, response speed, and duration. A certain energy storage has a rated power of 1 MW and can be started within 5 minutes after receiving the dispatching instruction, and can discharge for 2 hours. The pre-peak pre-regulation adjustable load capacity is calculated, and the user acceptance and constraints need to be considered. The heat storage device of a certain residential area can be pre-heated before the peak, but it needs to ensure that it does not affect the normal heating demand of the user, and the pre-peak pre-regulation adjustable load capacity is about 70% of the rated capacity. The contribution of pre-peak pre-regulation to peak load reduction is analyzed, and the pre-peak pre-regulation can reduce the peak load by about 8%, effectively alleviating the supply and demand pressure in the peak period.

[0047] According to the pre-peak pre-regulation adjustable load capacity, the peak load shifting potential area is divided. According to the size of the pre-peak pre-regulation adjustable load capacity, the area is divided into different peak load shifting potential levels. The area with pre-peak pre-regulation adjustable load capacity greater than 15% of the total load is classified as a high potential area, the area with capacity between 5%-15% is classified as a medium potential area, and the area with capacity less than 5% is classified as a low potential area. According to the pre-peak pre-regulation adjustable load capacity, a certain city identifies three high potential areas, which are equipped with a large number of energy storage devices and adjustable loads. The spatial distribution characteristics of the peak load shifting potential area are analyzed, and the aggregation mode of the potential area is identified. The high potential peak load shifting potential area in a certain area is mainly distributed around industrial parks and large commercial complexes, and these areas have rich load regulation resources. The regulation priority of different peak load shifting potential areas is evaluated, considering the adjustable load capacity, regulation cost, and user acceptance. A high potential peak load shifting potential area has low regulation cost and high user acceptance, and is set as a priority regulation area.

[0048] The transferable load data is derived from the peak shaving potential area. The type and scale of the transferable load are identified in the peak shaving potential area, including electric vehicle charging, energy storage device charging and discharging, industrial production load, and building air conditioning load. The electric vehicle charging load of a certain parking lot in the peak shaving potential area reaches 500 kW during the evening peak period, which can be transferred to the night valley period through intelligent charging management. The time characteristics of the transferable load are extracted and recorded in the transferable load data, including the start time, end time, and transfer duration. The standby cold machine of a certain data center in the peak shaving potential area can start pre-cooling 2 hours before the peak, transferring part of the cold load from the peak period. These time characteristics are recorded in the transferable load data. The spatial characteristics of the transferable load are extracted and recorded in the transferable load data, including the node where the load is located, the connection relationship with the energy supply network, and the transfer feasibility. The transferable load of a certain industrial park is mainly concentrated in three production workshops, and the transferable load data shows that these workshops are connected to the same substation, facilitating unified scheduling. The constraint conditions of the transferable load are extracted and recorded in the transferable load data, including user comfort requirements, production process limitations, and equipment performance constraints. The transfer of the residential air conditioning load requires that the indoor temperature fluctuation within the set range does not exceed 2°C, and the transfer of the production equipment load cannot affect the delivery period. The transferable load data includes load type, scale, time characteristics, spatial characteristics, and constraint conditions.

[0049] The regional supply and demand distribution map is formed according to the transferable load data. The transferable load data is mapped to the energy supply network topology, and the transferable load resources of each node are marked on the regional supply and demand distribution map. The regional supply and demand distribution map shows that the transferable load in the northern area is mainly industrial production load, and the transferable load in the southern area is mainly commercial air conditioning load and electric vehicle charging load. The supply and demand distribution changes before and after load transfer are calculated, and load transfer will change the original supply and demand matching state. The regional supply and demand distribution map shows that the supply and demand matching degree of a certain node increases from 0.75 to 0.92 by transferring part of the load in the transferable load data to the valley period. The regional supply and demand distribution map of different time periods is drawn to show the changes in supply and demand situation before and after peak shaving. The regional supply and demand distribution map of a certain city shows that after implementing peak shaving, the number of supply and demand imbalance nodes during the peak period decreases from 8 to 3, and the supply surplus phenomenon during the valley period is alleviated. The spatial imbalance of the regional supply and demand distribution map is analyzed to identify the areas where the supply and demand contradiction still exists. The regional supply and demand distribution map shows that a certain area still has a supply shortage even after load transfer, which needs to be addressed through other means such as cross-regional energy allocation or increasing supply capacity.

[0050] Step S130: Obtain the user response delay duration and the device start-stop cycle; identify the time elasticity range through the response delay duration; capture the adjustment switching time according to the device start-stop cycle; and perform a time-series comparison between the time elasticity range and the adjustment switching time to generate a peak-shaving adjustment space.

[0051] Specifically, user response latency and equipment start-up / shutdown cycles are collected. User response latency for various user types is collected from the demand response platform, and the time interval from the issuance of a dispatch command to the actual load change is statistically analyzed. After receiving a peak-shaving command, a factory needs to coordinate the shutdown sequence of its production lines; the user response latency is 18 minutes after the command is issued. Analyzing the response characteristics of different user types, commercial building air conditioning users have shorter response latency, industrial users have longer response latency due to production process constraints, and residential users have more fluctuating response latency due to behavioral habits. A shopping mall can adjust its central air conditioning power within 5 minutes of receiving a dispatch command; the user response latency is 5 minutes. However, a chemical company needs to complete its current batch of production before reducing the load; the user response latency may extend to 40 minutes. Equipment start-up / shutdown cycle characteristics are collected, including equipment start-up time, normal operating period, shutdown time, and restart preparation time. A boiler's equipment start-up / shutdown cycle is as follows: preheating at 6:00 AM, normal operation at 7:00 AM, load reduction starting at 10:00 PM, complete shutdown at 11:00 PM, and low-power standby starting in the early morning of the next day. Extract the regularity characteristics of equipment start-up and shutdown to identify the matching relationship between equipment start-up and shutdown cycles and load demand cycles. The start-up and shutdown cycle of a data center cooling unit adjusts with changes in server load. During off-peak hours at night, some cooling units can be shut down. The equipment start-up and shutdown cycle is highly correlated with the computing load cycle.

[0052] In some embodiments, identifying the time elasticity interval by the response delay duration includes: generating a delay tolerance threshold based on the response delay duration; applying an upper limit constraint to the response delay duration based on the delay tolerance threshold to generate an elastic reserve constraint delay segment; performing a stability test on the elastic reserve constraint delay segment to determine a stable delay period; and extracting the maximum continuous duration from the stable delay period as the time elasticity interval.

[0053] The delay tolerance threshold is generated according to the response delay duration. The distribution rule of the user response delay duration in the historical scheduling events is counted, and the delay duration corresponding to different percentiles is calculated. The user response delay duration data of a certain area shows that 50% of the responses are completed within 15 minutes, 80% of the responses are completed within 25 minutes, and 95% of the responses are completed within 40 minutes. According to the urgency of the scheduling demand, the maximum response delay that can be tolerated is determined from the statistical data of the user response delay duration, and the delay tolerance threshold is generated. For regular peak shaving demand, the delay tolerance threshold can be set longer; for load control in emergency situations, the delay tolerance threshold is set very low. The delay tolerance threshold of a certain power grid is 30 minutes in regular peak shaving, but the delay tolerance threshold requirement is 5 minutes in frequency anomaly. Different categories of users are set with differentiated delay tolerance thresholds in combination with user types and load characteristics. Industrial users have complex production processes, and the delay tolerance threshold extracted from their user response delay duration data is set to 45 minutes; commercial users respond flexibly, and the delay tolerance threshold is set to 20 minutes; critical load users must respond quickly, and the delay tolerance threshold is set to 10 minutes. The delay tolerance threshold integrates statistical analysis of user response delay duration, emergency level evaluation and user classification.

[0054] The elastic reserve constraint delay section is generated by upper limiting the response delay duration based on the delay tolerance threshold. The actual user response delay duration is compared with the delay tolerance threshold to identify the delay section that meets the threshold requirement. The actual user response delay duration of a certain user ranges from 10 to 35 minutes, and the delay tolerance threshold is 40 minutes. All responses of this user are within the tolerance range. Abnormal response data that exceeds the delay tolerance threshold in the user response delay duration is excluded, which may be caused by equipment failure or communication interruption and should not be considered in regular scheduling. In a certain scheduling event, a user's response delay duration reached 75 minutes due to equipment failure, and this data was marked as abnormal and excluded from the elastic reserve calculation. Under the constraint of the delay tolerance threshold, the elastic reserve constraint delay section available for scheduling is calculated, which represents the available time window from the issuance of the instruction to the completion of the response. A scheduling instruction requires load reduction before 18:00. Considering that the average user response delay duration of the user is 20 minutes and the delay tolerance threshold is 30 minutes, the latest time of the instruction issuance is 17:30, and the elastic reserve constraint delay section is 17:00-17:30. The elastic reserve constraint delay section of a certain area is 30 minutes, which means that the issuance time of the scheduling instruction can be flexibly selected within 30 minutes.

[0055] The stability test of the elastic reserve constraint delay section determines the stable delay period. First, analyze the fluctuation characteristics of the elastic reserve constraint delay section under different times and conditions, and evaluate its stability. Calculate the stability index of the elastic reserve constraint delay section, which is quantified by the coefficient of variation CV, and the calculation formula is CV=σ / μ, where σ is the standard deviation of the elastic reserve constraint delay section, and μ is the mean. Identify the key factors affecting the stability of the elastic reserve constraint delay section, including weather conditions, user behavior pattern changes, and equipment health status. In a certain area, under extreme high temperature weather, the air conditioning load surges, causing the response capacity to decrease, the elastic reserve constraint delay section to shorten by 20%, and the coefficient of variation CV to rise to 0.28, affecting the stability. Set the stability threshold CV_threshold, when CV<CV_threshold, it is determined as a stable delay period, and the periods with insufficient stability are removed, and the periods with stable stability meeting the requirements are retained as stable delay periods. A certain dispatching platform requires that the fluctuation amplitude of the elastic reserve constraint delay section be less than 25%, and sets the stability threshold CV_threshold=0.25, when the coefficient of variation CV is higher than 0.25, the period is marked as unstable and not included in the regular dispatching plan. Perform stability screening on the elastic reserve constraint delay section, and determine the stable delay period by the period that passes the stability test.

[0056] Extract the maximum continuous duration from the stable delay period as the time elasticity interval. First, identify the continuous available time segment in the stable delay period, which requires that the delay characteristics remain stable and uninterrupted within the period. A certain user responds to the delay stably and without production interference from 14:00 to 16:00, forming a continuous 2-hour stable delay period. Extract the maximum continuous duration in the stable delay period, which is the time elasticity interval. The time elasticity interval reflects the optimal dispatching window. The time elasticity interval extracted from a certain industrial park is 3.5 hours, occurring at night from 22:00 to 1:30 the next day, which can be used as the optimal window for load regulation in the park. Consider the length requirement of the dispatching demand, if the extracted time elasticity interval is not enough to meet the dispatching demand, select a suboptimal continuous period or combine multiple short periods. A certain dispatching task requires 4 hours of continuous regulation, but the user extracts a time elasticity interval of only 3 hours, so it needs to select two shorter periods or consider other users to participate.

[0057] The adjustment switching time is captured according to the device start-stop cycle. The key time nodes of the device state transition are extracted from the device start-stop cycle data, including the start time, the time when the stable operation is reached, the time when the load reduction starts, and the time when the complete shutdown is reached. The device start-stop cycle of a certain production device shows that it starts at 8:00, reaches the rated power at 8:30, starts to reduce the load at 17:00, and completely stops at 17:30. These four times are all potential adjustment switching times. The adjustment switching times that match the load scheduling requirements are identified, and the natural start-stop times in the device start-stop cycle are compared with the peak shaving and valley filling requirements of the power grid. The air conditioner of a certain commercial complex starts at 9:00 in the morning during the device start-stop cycle, which coincides with the morning peak of the power grid. If the start time is advanced to 7:30, both the comfort requirement and the peak load can be avoided. The feasibility and cost of implementing load adjustment at the adjustment switching time are evaluated. The adjustment cost of certain adjustment switching times is low and the user perception is weak, and the adjustment of certain adjustment switching times may affect normal business. It is difficult for a hotel to adjust the air conditioning load during the peak period of room check-out, but it can achieve 10% load reduction during the afternoon tea period without increasing customer complaint rate. By comparing each time node in the device start-stop cycle with the load scheduling requirements, the adjustment switching times that meet both the device operation requirements and the scheduling requirements are selected.

[0058] The time flexibility interval and the adjustment switching time are compared in time sequence to generate the peak shifting adjustment space. Based on the time flexibility interval and the adjustment switching time, the peak shifting adjustment space is generated. First, the start and end times of the time flexibility interval and the positions of the adjustment switching times are marked on the time axis, and the overlapping area of the time flexibility interval and the adjustment switching time is identified. The time flexibility interval of a certain user is 17:00-17:30, and the adjustment switching times of the device include 17:00 load reduction start and 17:30 complete shutdown. Both adjustment switching times are located within the time flexibility interval, forming an effective peak shifting adjustment window. The matching degree of the time flexibility interval and the adjustment switching time is analyzed. The combination with high matching degree can achieve low-cost load adjustment, and the combination with low matching degree needs manual intervention or additional incentives. The natural shutdown time of a certain production enterprise is 23:00, while the power grid needs to shave the peak between 21:00 and 22:00. The time flexibility interval and the adjustment switching time do not match, and the peak shifting needs to be achieved by early shutdown, but economic compensation is needed. A multi-dimensional representation of the peak shifting adjustment space is constructed, including the adjustment window in the time dimension, the adjustable load distribution in the space dimension, and the adjustable load capacity in the capacity dimension. The peak shifting adjustment space shows that there are 12 users in a certain region that can participate in adjustment during the 18:00-19:00 period, with a total adjustable capacity of 3.5MW, and the spatial distribution is concentrated in the northern part of the industrial park.

[0059] Step S140, gap matching analysis is performed on the regional supply and demand distribution map and the peak-shaving regulation space to generate gap compensation parameters, a regulation target node is determined through the gap compensation parameters, and a hierarchical regulation instruction set is created based on the regulation target node.

[0060] In some embodiments, the gap matching analysis performed on the regional supply and demand distribution map and the peak-shaving regulation space to generate gap compensation parameters includes: gap interaction transmission tracking is performed on the regional supply and demand distribution map to identify a gap trigger chain; the gap trigger chain is subjected to compensation loss evaluation with the peak-shaving regulation space to generate a compensation loss coefficient; the compensation loss coefficient is used to identify a compensation dissipation boundary; and the compensation dissipation boundary is used to determine the gap compensation parameters.

[0061] Gap interaction transmission tracking is performed on the regional supply and demand distribution map to identify a gap trigger chain. The energy transmission relationship between nodes in the regional supply and demand distribution map is analyzed, and the propagation path of the gap in the network is tracked. The supply gap of a certain node will be transmitted to adjacent nodes through the energy network, which may trigger a chain reaction. This chain transmission process is the core feature of the gap trigger chain. The connection relationship and energy flow direction between nodes are extracted from the regional supply and demand distribution map. In a certain heat supply network, the heat source station supplies heat to the heat exchange station through the main pipeline, and the heat exchange station supplies heat to the downstream users. When the heat supply capacity of the heat source station is insufficient, the gap first affects the directly connected heat exchange station, which cannot meet the downstream demand, and the gap continues to transmit to the user end, forming a complete gap trigger chain. The propagation process of the gap trigger chain is tracked, and the starting position, propagation path and influence range of the gap are recorded. After receiving the upstream gap, a certain energy storage node compensates part of the gap by releasing stored energy, slowing down the transmission speed of the gap trigger chain to the downstream. After receiving the gap, a certain load-intensive node has a large demand, and the gap appears to be amplified at this node, and the gap trigger chain shows a diffusion feature. The response behavior of each node in the gap trigger chain is analyzed, and some nodes have buffering capacity to delay the propagation of the gap trigger chain, and some nodes will accelerate the diffusion of the gap trigger chain due to insufficient regulation capacity.

[0062] A compensation loss coefficient is generated by implementing a compensation loss evaluation on the gap trigger chain and the staggered regulation space. The compensation demand of each node in the gap trigger chain is extracted, and the position and time of the compensation resource in the staggered regulation space are identified. A gap trigger chain contains several nodes, and the gap size and occurrence time of each node are known. The available compensation resources are extracted from the staggered regulation space, and these resources have certain capacity, response speed, and available period. The process loss of the compensation resource from the regulation node to the gap node is evaluated. When the compensation resource is transmitted from the regulation node to the gap node, it will experience response delay and transmission loss. A certain compensation resource needs to pass through several intermediate nodes from the regulation node to the gap node, and each node will cause transmission loss. Response delay also causes lag in compensation time. The compensation loss coefficient is calculated by comparing the theoretical compensation capacity with the actual effective compensation capacity. The formula is η = C_eff / C_theo, where η is the compensation loss coefficient, C_eff is the actual effective compensation capacity considering transmission loss and response delay, and C_theo is the theoretical compensation capacity. A higher compensation loss coefficient indicates better compensation efficiency, and a lower compensation loss coefficient indicates greater loss and lower efficiency in the compensation process.

[0063] The compensation dissipation boundary is identified according to the compensation loss coefficient. The spatial distribution characteristics of the compensation loss coefficient are analyzed, and the areas with high loss are identified. Some areas have low compensation loss coefficients due to their distance from the regulation resources or the number of intermediate links. These areas are considered to be difficult to compensate. The compensation loss coefficient of a remote industrial area is 0.55, indicating that the staggered regulation resources are difficult to effectively compensate for the gap in this area. The compensation loss coefficient of a central urban area is 0.88, indicating that the staggered regulation effect is good. A threshold value of the compensation loss coefficient is set as the judgment standard, and areas below the threshold value are excluded from the compensation dissipation boundary. The gaps in these areas are difficult to compensate effectively through staggered regulation, and other regulation methods such as increasing local supply capacity or implementing demand-side management are needed. A city sets the compensation loss coefficient threshold value to 0.70, and identifies several nodes with compensation loss coefficients below 0.70. These nodes are outside the compensation dissipation boundary, and special control schemes are developed for these nodes. The identification of the compensation dissipation boundary clearly defines the effective range of the staggered regulation, allowing the regulation resources to be accurately placed in areas with high compensation efficiency.

[0064] The notch compensation parameters are determined according to the compensation dissipation boundary. In the area within the compensation dissipation boundary, the effective compensation capacity of each node is calculated. The notch data and the corresponding compensation loss coefficient are extracted from the nodes included in the notch compensation parameters, and the effective compensation capacity is obtained by multiplying the theoretical notch by the compensation loss coefficient. The theoretical notch of a certain node is 600 kW, and the compensation loss coefficient is 0.82. The effective compensation capacity of the node is calculated to be 492 kW, which is included in the notch compensation parameters, and the remaining 108 kW notch needs to be solved by other means. The compensation start time in the notch compensation parameters needs to consider the response delay and transmission time. The notch of a certain node appears at 18:00 in the evening, and the response delay of the compensation resource is 15 minutes and the transmission time is 8 minutes. In order to ensure that the compensation resource arrives in time, the compensation start time in the notch compensation parameters is specified as 17:37. For the area outside the compensation dissipation boundary, the compensation capacity of these areas in the notch compensation parameters is marked as zero, indicating that peak shifting regulation cannot be covered. A certain suburban industrial park is located outside the compensation dissipation boundary, and the available compensation capacity in its notch compensation parameters is 0, which needs to rely on local energy storage or temporary power generation equipment. The notch compensation parameters cover key information such as effective compensation capacity, compensation start time, compensation efficiency, and compensable area range, and quantify the compensation demand and timing requirements of each node.

[0065] The control target nodes are established through the notch compensation parameters. By analyzing the compensation capacity distribution in the notch compensation parameters, the areas and nodes that need to be controlled are identified. From the notch compensation parameter data, it can be seen that the compensation demand of some nodes is large, and the compensation demand of some nodes is moderate. Nodes with large compensation capacity are marked as key control nodes, and these control target nodes have a significant impact on overall supply and demand balance. A certain city identifies several key control target nodes in industrial areas, commercial areas, and residential areas from the notch compensation parameters, and the compensation demand of these control target nodes accounts for a major part of the total notch. By analyzing the time characteristics in the notch compensation parameters, the time period when the notch is concentrated is identified. The morning and evening peak periods are the time periods when the notch is concentrated, and the control preparation needs to be started before these periods. The control priority of each control target node is evaluated, and although a certain substation node has a small notch, it supplies power to hospitals and data centers, which is extremely important, and the control priority in the control target node list is set to the highest. A certain industrial node has a large notch but the user has strong flexibility, and the control priority is set to moderate. By comprehensively analyzing the spatial distribution, time characteristics, and node importance of the notch compensation parameters, a list of control target nodes that need to be implemented is established, and these control target nodes constitute the execution object of the subsequent control instructions.

[0066] The hierarchical regulation instruction set is created based on the regulation target nodes. The regulation target nodes are layered according to response speed. The regulation target nodes with short response time, including energy storage devices and interruptible loads, can quickly respond to regulation requirements. The regulation target nodes with long response time, including industrial loads and building air conditioners, need a longer time to complete the response. Among the regulation target nodes of a certain city, the energy storage stations and large shopping mall air conditioners can respond within a short time, and the chemical plants and printing workshops need a longer time to adjust production arrangements. For the regulation target nodes of different response levels, corresponding regulation instructions are created in the hierarchical regulation instruction set. The fast response layer instructions of the hierarchical regulation instruction set require a short time to complete the response, and the slow response layer instructions of the hierarchical regulation instruction set allow a longer time to complete the response. The hierarchical regulation instruction set of a certain industrial park node stipulates that it is started in the evening period and uses production shift adjustment to reduce load, and the response is completed before the specified time. The hierarchical regulation instruction set of a certain energy storage station node requires discharging immediately after receiving the instruction to quickly compensate for the gap. A priority mechanism of the hierarchical regulation instruction set is established. When multiple regulation target nodes need to be regulated at the same time, the hierarchical regulation instruction set gives priority to the instructions of important nodes and fast response nodes. This hierarchical design can meet the requirements of fast response and fully utilize slow response resources.

[0067] In step S150, the hierarchical regulation instruction set is divided into immediate response instructions and delayed response instructions according to the response time difference. The fluctuation adjustment income recognition is performed on the delayed response instructions to obtain the peak shaving income coefficient. The peak shaving income coefficient is cooperatively configured with the immediate response instructions to form the regulation coordination parameter.

[0068] In some embodiments, the hierarchical regulation instruction set is divided into immediate response instructions and delayed response instructions according to the response time difference, including: establishing a time response axis according to the response time difference; mapping the hierarchical regulation instruction set to the time response axis to form a response time label; dividing the hierarchical regulation instruction set into a fast response section and a peak shaving resource slow response section according to the response time label; and generating immediate response instructions and delayed response instructions by classifying the time difference of the fast response section and the peak shaving resource slow response section.

[0069] According to the response time difference, a time response axis is established. The response time difference data of all instructions in the hierarchical control instruction set is counted, and the response time difference ranges from several minutes to tens of minutes. A certain hierarchical control instruction set contains several instructions, and the response time difference of the shortest energy storage discharge can be completed in several minutes; the longest response time difference of industrial production adjustment needs to wait for the end of the current production batch. The time response axis is established as an analysis tool, the horizontal axis represents the response time extending backward from zero, and the vertical axis represents the number of instructions or the cumulative control capacity corresponding to the response time. On the horizontal axis of the time response axis, a reasonable time interval is marked on the scale, and the capacity scale is marked on the vertical axis to form a complete two-dimensional coordinate. The horizontal axis range of the time response axis of a certain city is set to 0-60 minutes, and the vertical axis range is set according to the actual control capacity. Through the time response axis, the distribution density of instructions in different response time periods can be observed intuitively. The instruction density in a short time period indicates that the fast response resources are rich, and the instruction dispersion in a long time period indicates that the slow response resources are mainly used.

[0070] The hierarchical control instruction set is mapped to the time response axis to form a response time marker. The response time and control capacity of each instruction in the hierarchical control instruction set are extracted, and these instructions are mapped one by one to the corresponding position of the time response axis. A certain energy storage discharge instruction has a short response time and moderate capacity, and is marked at a short time position on the time response axis; a certain industrial control instruction has a long response time and a large capacity, and is marked at a long time position. After all instructions are mapped, a distribution diagram of the response time marker is formed on the time response axis. The density of the marker reflects the concentration of instructions in different response time intervals, and the height of the marker reflects the control capacity. The response time marker of a certain city shows that there are intensive instruction markers in a short time interval, and the instruction markers are relatively dispersed in a longer time interval. By analyzing the aggregation characteristics of the response time marker, the response time cluster in the instruction set can be identified. The response time marker forms obvious aggregation near certain time points, and these aggregations correspond to the typical response time of different types of resources.

[0071] The layered regulation instruction set is divided into a fast response section and a slow response section of peak-shaving resources according to the response time mark. The distribution rule of the response time mark on the time response axis is analyzed to find the natural boundary between fast response and slow response. It is observed from the distribution of the response time mark that the density of instructions changes obviously near a certain time point. Before the time point, the instructions are densely distributed, and after the time point, the instructions are sparsely distributed. The time point reflects the transition area where fast response resources are exhausted and slow response resources start to play a role. The time point is taken as the boundary point between the fast response section and the slow response section of peak-shaving resources. The instructions within the time point are classified into the fast response section, and the instructions outside the time point are classified into the slow response section of peak-shaving resources. The fast response section mainly includes instructions of energy storage, interruptible load and standby unit, and the slow response section of peak-shaving resources mainly includes instructions of industrial load, building regulation and charging management. The instructions in the fast response section are characterized by fast response speed but short duration, and the instructions in the slow response section of peak-shaving resources are characterized by slow response speed but large regulation capacity and long duration.

[0072] The time difference classification of the fast response section and the slow response section of peak-shaving resources generates immediate response instructions and delayed response instructions. Specifically, the fast response section is converted into immediate execution timing, the slow response section of peak-shaving resources is superimposed on the immediate execution timing to form a response difference distribution, a time difference boundary threshold of a delay benefit window is derived from the response difference distribution, and immediate response instructions and delayed response instructions are generated according to the distinguishing significance of the time difference boundary threshold.

[0073] The fast response section is converted into immediate execution timing. The execution time and regulation capacity of all instructions in the fast response section are extracted and arranged in chronological order. The execution time of the instructions in the fast response section is concentrated in a short time after the dispatching command is issued. The fast response section of a certain city includes instructions such as energy storage discharge, interruptible load removal and standby unit startup, which are executed quickly after receiving the dispatching command. An immediate execution timing is constructed to depict the release process of the regulation capacity in the fast response section, with time as the horizontal axis and cumulative regulation capacity as the vertical axis. The immediate execution timing shows that the energy storage starts to discharge at the 2nd minute after the dispatching command is issued, the interruptible load is removed at the 5th minute, and the standby unit starts at the 8th minute. The cumulative regulation capacity increases step by step as time goes on. The slope of the curve of the immediate execution timing reflects the response speed. The larger the slope, the more regulation capacity is put into operation per unit time. The immediate execution timing of a certain industrial park has a larger slope in the first few minutes, indicating that fast response resources are intensively put into operation, and then the curve gradually flattens out, indicating that the fast response resources are basically released.

[0074] The response difference distribution is obtained by superimposing the capacity variation of the slow response resources on the immediate execution time sequence. The instructions in the slow response segment of a certain city start to take effect after 15 minutes, the building pre-cooling control takes effect after 25 minutes, and the charging transfer control takes effect after 40 minutes. The capacity variation curve of the slow response segment is superimposed on the immediate execution time sequence to obtain the response difference distribution containing the whole process of fast and slow response. The response difference distribution shows the complete transition from fast response to slow response. The curve rises steeply in the early stage due to the fast response resources, the slope gradually slows down in the middle stage, and continues to grow in the later stage due to the continuous input of slow response resources. The response difference distribution of a certain region is mainly contributed by the fast response in the first 10 minutes, is in the superimposition stage of fast and slow response in 10-30 minutes, and is mainly contributed by the slow response in 30-60 minutes.

[0075] The time difference threshold of the delay benefit window is derived from the response difference distribution. The slope variation law of the response difference distribution is analyzed to find the time when the slope changes significantly. The response difference distribution curve of a certain city has an obvious slope turning point at 12 minutes. The curve slope is larger before 12 minutes, and the curve slope is significantly reduced after 12 minutes. This slope turning point corresponds to the transition time when the fast response resources are gradually exhausted and the slow response resources begin to play a dominant role. The time when the slope changes significantly is taken as the candidate point of the time difference threshold. On the left side of the candidate point, the instructions are mainly fast response, although the response speed is fast, but the duration is short, and the peak shaving benefit is limited; on the right side of the candidate point, the instructions are mainly slow response, although the response speed is slow, but the peak shaving benefit brought by load transfer can be fully utilized. The average peak shaving benefit of the instructions on both sides of the candidate point is evaluated, and the benefit difference of both sides is analyzed. The 12 minutes of a certain city is taken as the candidate point. The average peak shaving benefit coefficient of the instructions on the left side of the candidate point is low, the average peak shaving benefit coefficient of the instructions on the right side is high, and the benefit difference is significant. The candidate point with the largest benefit difference is selected as the final time difference threshold.

[0076] The instant response instruction and the delayed response instruction are generated according to the distinguishing degree of the time difference threshold. The distinguishing ability of the time difference threshold is evaluated, and the response feature difference on both sides of the threshold is analyzed. The average response time of the instructions on the left side of the time difference threshold is short, the average response time of the instructions on the right side of the time difference threshold is long, and the response time difference on both sides is obvious. Meanwhile, the difference characteristics of the instructions on both sides in capacity distribution and off-peak income are analyzed. The single capacity of the instructions on the left side is small, but the response speed is fast, and the single capacity of the instructions on the right side is large, and the off-peak income is high. The distinguishing degree of the time difference threshold is evaluated by comprehensively analyzing the response time difference and other feature differences. The distinguishing degree score of a certain threshold is high, indicating that the threshold can effectively distinguish the two types of instructions of fast response and slow response. The layered regulation instruction set is classified according to the time difference threshold, and the instant response instruction and the delayed response instruction are generated. A certain city generates the instructions with a response time of less than 12 minutes as instant response instructions, and the instant response instructions are characterized by fast response and covering initial gaps. The instructions with a response time of more than 12 minutes are generated as delayed response instructions, and the delayed response instructions are characterized by large capacity and high off-peak income.

[0077] In some embodiments, the fluctuation regulation income recognition performed on the delayed response instruction includes: converting the delayed response instruction into a load transfer vector; locating a position with the largest peak-valley difference in the load transfer vector; taking the position with the largest peak-valley difference as a yield conversion reference to accumulate yields to form a yield distribution curve; and performing peak extraction on the yield distribution curve to determine an off-peak income coefficient.

[0078] The delayed response instruction is converted into a load transfer vector. The load transfer data in the delayed response instruction is extracted, including transfer capacity, transfer source time, and transfer target time. A certain delayed response instruction transfers industrial load from the evening peak to the night valley, and the transfer source time is the evening peak period and the transfer target time is the night valley period. A load transfer vector is constructed, and the dimension of the vector is the time axis of a day. The elements of the vector represent the load change at each time. In the load transfer vector, the element value of the transfer source time is negative, indicating that the load decreases, the element value of the transfer target time is positive, indicating that the load increases, and the element values of other times are zero. The element value of the load transfer vector of a certain delayed response instruction at 19:00 in the evening peak is negative, indicating that the load decreases at this time; the element value at 23:00 in the night valley is positive, indicating that the load increases at this time. The load transfer effects of multiple delayed response instructions are superimposed to form a comprehensive load transfer vector. The load transfer vector after superimposition of all delayed response instructions in a certain area shows that a certain load is reduced in the evening peak period, and the corresponding load is increased in the night valley period.

[0079] Locate the position of the maximum peak-valley difference in the load transfer vector. Superimpose the load transfer vector on the original load curve to obtain the adjusted load curve. The original load curve is at its peak during the evening peak and at its valley during the night valley, with a large peak-valley difference. After performing load transfer, the load during the evening peak is reduced, the load during the night valley is increased, and the peak-valley difference is reduced. Analyze the changes in peak-valley difference at each time before and after adjustment to identify the time with the largest change in peak-valley difference. The peak shaving effect is significant at a certain time, the valley filling effect is significant at a certain time, and the peak-valley difference decreases the most between the two times. The time is the position of the maximum peak-valley difference. The position of the maximum peak-valley difference in a certain area corresponds to the evening peak and the night valley, and the original peak-valley difference is large. After load transfer, the peak-valley difference is significantly reduced, and the peak shaving and valley filling effect is best. Mark the position of the maximum peak-valley difference as the income conversion reference point. This reference point corresponds to the maximum peak shaving adjustment value. The position of the maximum peak-valley difference in a certain city is between the evening peak reduction and the night valley filling. This position becomes the reference benchmark for calculating the peak shaving income.

[0080] Construct the income distribution curve by accumulating income with the maximum peak-valley difference position as the income conversion reference. Starting from the position of the maximum peak-valley difference, calculate the economic income brought by the time delay response instruction at different times. Reducing load during the evening peak avoids starting high-cost gas peak shaving units, saving fuel costs. Increasing load during the night valley improves the utilization rate of base load units and increases power generation income. In a certain city, reducing load during the evening peak avoids starting high-cost gas units, saving operating costs per hour; increasing load during the night valley improves the utilization rate of base load units, increasing income per hour. Accumulate the economic income at each time in chronological order to construct the income distribution curve. The income distribution curve takes time as the horizontal axis and cumulative income as the vertical axis, showing the accumulation process of peak shaving income over time. A certain income distribution curve shows that peak shaving income starts to accumulate in the late afternoon, reaches a peak at a certain time, then slows down, and finally reaches the total amount of daily cumulative income. The peak position of the income distribution curve reflects the period of the most concentrated peak shaving income, and the slope of the curve reflects the growth rate of income.

[0081] The peak value extraction is performed on the income distribution curve to determine the off-peak income coefficient. The peak value income is extracted from the income distribution curve, and the peak value income represents the maximum single-period economic benefit brought by the time-delay response instruction. The peak value of a certain income distribution curve appears at a certain time, which is the time period when the off-peak income is most concentrated, and the peak value size reflects the maximum economic value of off-peak regulation. The off-peak income coefficient is obtained by the ratio of the peak value income to the execution cost, and the calculation formula is off-peak income coefficient = R_peak / C_total, wherein R_peak is the peak value income, and C_total is the execution cost, which includes user compensation fees, device start-stop cost and dispatching operation cost. The execution cost C_total of a certain time-delay response instruction is certain, the peak value income R_peak is obviously higher than the cost, the off-peak income coefficient reaches a high level, indicating that the instruction has good economy. The off-peak income coefficients of multiple time-delay response instructions are counted, and the income characteristics of different types of instructions are analyzed. The off-peak income coefficients of industrial load transfer instructions are generally high, the off-peak income coefficients of building pre-cooling instructions are moderate, and the off-peak income coefficients of charging transfer instructions are also high.

[0082] According to the off-peak income coefficient and the instant response instruction, the regulation and control coordination parameters are formed. The time-delay response instructions with high off-peak income coefficients are selected as the priority regulation and control objects, and high-yield time-delay response instructions such as industrial park shift optimization and large-scale mall pre-cooling regulation are identified in a certain city. A cooperative configuration mechanism of instant response instructions and time-delay response instructions is established to determine the capacity ratio of regulation and control coordination parameters. According to the gap size and response time requirement, the regulation and control coordination parameters stipulate that the instant response instruction undertakes 30-40% of the regulation and control task for rapid response, and the time-delay response instruction undertakes 60-70% of the regulation and control task for sustained regulation and off-peak optimization. A certain evening peak gap is expected to be 5 MW, and the regulation and control coordination parameters stipulate that the instant response instruction provides 1.8 MW of rapid support, and the time-delay response instruction provides 3.2 MW of sustained regulation. The time sequence configuration of the regulation and control coordination parameters is determined, the instant response instruction is executed immediately after the gap occurs, and the time-delay response instruction is started in advance 15-30 minutes before the gap occurs to ensure that the regulation and control capacity is fully released when the peak comes. The regulation and control coordination parameters clearly define the capacity ratio, execution time sequence, priority order and income expectation of the instant response instruction and the time-delay response instruction, providing a detailed operation scheme for actual regulation and control execution.

[0083] In step S160, the regulation and control execution range is determined based on the regulation and control coordination parameters and the regional supply and demand distribution map, and the cost and income evaluation is performed on the regulation and control execution range to form a city-level energy dispatching execution scheme.

[0084] Specifically, the regulation execution range is determined based on the regulation coordination parameter and the regional supply-demand distribution map. The coverage area of the immediate response instruction and the delayed response instruction is extracted from the regulation coordination parameter to determine which nodes and time periods need to execute the regulation. The regulation coordination parameter shows that the immediate response instruction covers several nodes in the central business district and industrial park, and the delayed response instruction covers a wider range, covering several nodes in the city. The spatial and temporal distribution characteristics of the supply-demand gap are extracted from the regional supply-demand distribution map. The regional supply-demand distribution map shows that the northern industrial zone, central business district and southern residential area have obvious supply-demand gaps during morning and evening peak hours, and the gap location and time have been clearly marked. The coverage range of the regulation coordination parameter is spatially superimposed with the gap distribution of the regional supply-demand distribution map. Through superposition, it is found that the regulation nodes in the regulation coordination parameter can cover most of the gap areas in the regional supply-demand distribution map, and the gap in some remote areas cannot be covered due to the difficulty of regulation resources to reach. The regulation execution range is determined, and the areas that can effectively cover the supply-demand gap and have sufficient regulation resources are included in the execution range. The regulation execution range of a certain city is finally determined as the central business district, northern industrial park and southern residential area. The regulation nodes in these areas can cover the main supply-demand gap, the regulation resources are reasonably allocated, and the execution conditions are mature.

[0085] The cost-benefit evaluation of the regulation execution range forms the city-level energy dispatch execution scheme. In the regulation execution range, the cost-benefit framework of the city-level energy dispatch execution scheme is constructed. The total regulation cost C_total in the regulation execution range is counted, and the cost composition includes user compensation fees, device start-stop costs and dispatch operation and maintenance fees. The total regulation benefit R_total in the regulation execution range is counted, and the benefit sources include reduced fuel costs, reduced device investment needs and improved energy utilization efficiency. By executing the immediate response instructions and the delayed response instructions, the gas peak shaving units with high start-up costs are avoided, the peak load of the power grid is reduced, and the utilization rate of the base load units is improved. The economic feasibility of the city-level energy dispatch execution scheme is evaluated through the benefit-cost ratio, and the calculation formula is benefit-cost ratio = R_total / C_total, wherein R_total is the total regulation benefit, and C_total is the total regulation cost. The total regulation benefit R_total of a certain city is significantly higher than the total cost C_total, the benefit-cost ratio is high, and it is indicated that the economic benefit of the city-level energy dispatch execution scheme is significant. The technical feasibility, user acceptance and execution reliability of the city-level energy dispatch execution scheme are evaluated. The technical feasibility evaluation confirms that the devices and communication conditions in the regulation execution range meet the regulation requirements, the user acceptance survey shows that most users are willing to participate in the regulation plan, and the execution reliability analysis shows that the success execution rate of the regulation instructions is high. The results of the cost-benefit evaluation and the feasibility evaluation are combined to perfect the contents of the city-level energy dispatch execution scheme. The city-level energy dispatch execution scheme clearly defines the regulation execution range, the instruction configuration, the cost-benefit expectation and the execution guarantee measures, and provides complete implementation basis for the accurate dispatch of city energy.

[0086] In order to implement the corresponding functions and technical effects of the above-mentioned method embodiment, a city energy network intelligent allocation method is executed. Referring to Figure 2 , Figure 2 The structure block diagram of a city energy network intelligent allocation system 200 provided by the embodiment of the application is shown. For ease of illustration, only the parts related to the embodiment are shown. The city energy network intelligent allocation system 200 provided by the embodiment of the application comprises:

[0087] The energy consumption collection module 201 is configured to acquire building energy consumption data and energy supply network topology information, identify load fluctuation characteristics according to the energy consumption data, and establish a supply-demand matching benchmark between the load fluctuation characteristics and the energy supply network topology information.

[0088] The load analysis module 202 is configured to detect a load peak value intensive period based on the supply-demand matching benchmark, extract a peak load shifting and valley filling potential region from the load peak value intensive period, derive transferable load data through the peak load shifting and valley filling potential region, and form a regional supply-demand distribution map according to the transferable load data.

[0089] The timing matching module 203 is configured to acquire a user response delay duration and a device start-stop cycle, identify a time elasticity interval through the response delay duration, capture an adjustment switching time according to the device start-stop cycle, perform timing comparison between the time elasticity interval and the adjustment switching time to generate a peak-shaving adjustment space;

[0090] The regulation strategy module 204 is configured to perform gap matching analysis on the regional supply-demand distribution map and the peak-shaving adjustment space to generate a gap compensation parameter, determine a regulation target node through the gap compensation parameter, and create a hierarchical regulation instruction set based on the regulation target node.

[0091] The instruction coordination module 205 is configured to distinguish an instant response instruction and a delay response instruction according to a response time difference based on the hierarchical regulation instruction set, identify a peak-shaving benefit coefficient through fluctuation adjustment benefit recognition of the delay response instruction, and form a regulation coordination parameter based on the peak-shaving benefit coefficient and the instant response instruction.

[0092] The scheme generation module 206 is configured to determine a regulation execution range based on the regulation coordination parameter and the regional supply-demand distribution map, and form a city-level energy dispatch execution scheme through cost-benefit evaluation of the regulation execution range.

[0093] The above-described city energy network intelligent deployment system 200 can implement a city energy network intelligent deployment method according to the above-described method embodiment. The optional items in the above-described method embodiment are also applicable to the present embodiment, which will not be described in detail herein. The remaining content of the present embodiment can be referred to the content of the above-described method embodiment, which will not be described in detail herein.

[0094] The purpose of the above embodiments is to exemplarily reproduce and deduce the technical solutions of the present application, and to completely describe the technical solutions, purposes and effects of the present application. The purpose is to make the public understand the disclosure of the present application more thoroughly and comprehensively, and does not limit the protection scope of the present application.

Claims

1. A method for intelligent dispatching of urban energy networks, characterized in that, include: Acquiring building energy consumption data and energy supply network topology information, identifying load fluctuation characteristics based on the energy consumption data, and establishing a supply-demand matching benchmark by combining the load fluctuation characteristics with the energy supply network topology information, includes: identifying supply-demand balance nodes and supply-demand imbalance nodes based on the matching relationship between the load fluctuation characteristics and the energy supply network topology information; generating a supply-demand deviation value by performing capacity allocation on the surplus of the supply-demand balance nodes and the gap of the supply-demand imbalance nodes; determining a capacity allocation threshold based on the supply-demand deviation value; and applying the capacity allocation threshold to the energy supply network topology information to establish a supply-demand matching benchmark. Based on the supply and demand matching benchmark, the peak load period is detected, the peak shaving and valley filling potential area is extracted from the peak load period, the transferable load data is exported from the peak shaving and valley filling potential area, and a regional supply and demand distribution map is formed based on the transferable load data; The process involves acquiring user response latency and device start-stop cycle, and identifying a time elasticity interval based on the response latency, including: generating a latency tolerance threshold based on the response latency; imposing an upper limit constraint on the response latency based on the latency tolerance threshold to generate an elastic reserve constraint latency segment; performing a stability test on the elastic reserve constraint latency segment to determine a stable latency period; extracting the maximum continuous duration from the stable latency period as the time elasticity interval; capturing the adjustment switching time based on the device start-stop cycle; and performing a time-series comparison between the time elasticity interval and the adjustment switching time to generate a peak-shaving adjustment space. Gap matching analysis is performed on the regional supply and demand distribution map and the peak-shifting adjustment space to generate gap compensation parameters. The control target node is established through the gap compensation parameters, and a hierarchical control instruction set is created based on the control target node. The hierarchical control instruction set is differentiated into immediate response instructions and delayed response instructions based on the response time difference, including: establishing a time response axis based on the response time difference; mapping the hierarchical control instruction set to the time response axis to form response time markers; dividing the hierarchical control instruction set into a fast response segment and a slow response segment for peak-shaving resources based on the response time markers; classifying the fast response segment and the slow response segment for peak-shaving resources based on time difference to generate immediate response instructions and delayed response instructions; and performing fluctuation regulation benefit identification on the delayed response instructions to obtain peak-shaving benefit coefficients, including: converting the delayed response instructions into a load transfer vector; locating the position with the maximum peak-to-valley difference in the load transfer vector; using the position with the maximum peak-to-valley difference as the benefit conversion benchmark to accumulate benefits and form a benefit distribution curve; performing peak extraction on the benefit distribution curve to determine the peak-shaving benefit coefficients; and configuring the peak-shaving benefit coefficients in conjunction with the immediate response instructions to form control coordination parameters. Based on the aforementioned regulation coordination parameters and the regional supply and demand distribution map, the scope of regulation implementation is determined, and a cost-benefit assessment is conducted on the scope of regulation implementation to form a city-level energy dispatch implementation plan.

2. The method according to claim 1, characterized in that, The extraction of peak-shaving and valley-filling potential areas from the peak load periods includes: Peak duration data is generated by identifying peak duration during the peak load periods. The peak start position is determined based on the peak duration data; The pre-peak adjustable load capacity is obtained by using the peak start position; Based on the pre-peak adjustable load capacity, the potential areas for peak shaving and valley filling are divided.

3. The method according to claim 1, characterized in that, The step of performing gap matching analysis on the regional supply and demand distribution map and the peak-shifting adjustment space to generate gap compensation parameters includes: The gap triggering chain is identified by inter-gap communication and tracking on the supply and demand distribution map of the region. The gap trigger chain and the peak shifting adjustment space are subjected to a compensation loss assessment to generate a compensation loss coefficient. The compensation dissipation boundary is identified based on the aforementioned compensation loss coefficient; The gap compensation parameters are determined based on the compensation dissipation boundary.

4. The method according to claim 2, characterized in that, Determining the peak start position based on the peak duration data includes: Analyze the key points of duration variation in the peak duration data; Determine the available peak prediction time at the key points of the duration change; The accuracy of the predicted peak time is evaluated to generate an accuracy evaluation result; The peak start position is generated based on the peak prediction time and the accuracy evaluation result.

5. The method according to claim 1, characterized in that, The step of classifying the fast response segment and the slow response segment for off-peak resources by time difference to generate immediate response commands and delayed response commands includes: Convert the fast response segment into an instant execution sequence; The slow response segment of the off-peak resource is superimposed on the real-time execution sequence to form a response difference distribution; Derive the time difference threshold of the delay benefit window from the response difference distribution; Immediate response commands and delayed response commands are generated based on the distinguishability of the time difference threshold.

6. A smart dispatching system for urban energy networks, characterized in that, include: An energy consumption acquisition module is used to acquire building energy consumption data and energy supply network topology information, identify load fluctuation characteristics based on the energy consumption data, and establish a supply-demand matching benchmark by comparing the load fluctuation characteristics with the energy supply network topology information. This includes: identifying supply-demand balance nodes and supply-demand imbalance nodes based on the matching relationship between the load fluctuation characteristics and the energy supply network topology information; generating a supply-demand deviation value by performing capacity matching between the surplus of the supply-demand balance nodes and the gap of the supply-demand imbalance nodes; determining a capacity allocation threshold based on the supply-demand deviation value; and applying the capacity allocation threshold to the energy supply network topology information to establish a supply-demand matching benchmark. The load analysis module is used to detect periods of concentrated load peaks based on the supply and demand matching benchmark, extract potential areas for peak shaving and valley filling from the periods of concentrated load peaks, export transferable load data from the potential areas for peak shaving and valley filling, and form a regional supply and demand distribution map based on the transferable load data. The timing matching module is used to acquire the user response latency duration and the device start-stop cycle, and to identify a time elasticity interval through the response latency duration. This includes: generating a latency tolerance threshold based on the response latency duration; applying an upper limit constraint to the response latency duration based on the latency tolerance threshold to generate an elastic reserve constraint latency segment; performing a stability test on the elastic reserve constraint latency segment to determine a stable latency period; extracting the maximum continuous duration from the stable latency period as the time elasticity interval; capturing the adjustment switching time according to the device start-stop cycle; and performing a timing comparison between the time elasticity interval and the adjustment switching time to generate a peak-shifting adjustment space. The regulation strategy module is used to perform gap matching analysis on the regional supply and demand distribution map and the peak-shifting regulation space to generate gap compensation parameters, establish regulation target nodes through the gap compensation parameters, and create a hierarchical regulation instruction set based on the regulation target nodes; The instruction coordination module is used to generate immediate response instructions and delayed response instructions by distinguishing the hierarchical control instruction set according to the response time difference. This includes: establishing a time response axis based on the response time difference; mapping the hierarchical control instruction set to the time response axis to form response time markers; dividing the hierarchical control instruction set into a fast response segment and a slow response segment for peak-shaving resources based on the response time markers; classifying the fast response segment and the slow response segment for peak-shaving resources by time difference to generate immediate response instructions and delayed response instructions; and performing fluctuation regulation benefit identification on the delayed response instructions to obtain peak-shaving benefit coefficients. This includes: converting the delayed response instructions into a load transfer vector; locating the position with the maximum peak-to-valley difference in the load transfer vector; accumulating benefits using the position with the maximum peak-to-valley difference as a benefit conversion benchmark to form a benefit distribution curve; extracting peak values ​​from the benefit distribution curve to determine the peak-shaving benefit coefficients; and configuring the peak-shaving benefit coefficients in conjunction with the immediate response instructions to form control coordination parameters. The scheme generation module is used to determine the scope of regulation execution based on the regulation coordination parameters and the regional supply and demand distribution map, and to conduct a cost-benefit assessment on the scope of regulation execution to form a city-level energy dispatch execution scheme.

Images