A method for improving power grid frequency safety based on a virtual power plant

Abstract

The application discloses a kind of power grid frequency safety promotion methods based on virtual power plant, including in hierarchical virtual power plant architecture, using the minimum probability scene method based on Lyapunov criterion and failure mode and effects analysis method, considering information-physical uncertainty virtual power plant reliable capacity mathematical model is established for selected area;With investment minimization as target, with information-physical uncertain scene normal capacity, reliable capacity and delay as constraint, the virtual power plant terminal-communication system reconstruction planning model is constructed;Based on mathematical transformation, the planning model is converted into mixed integer linear programming model, and the reliable frequency modulation capacity of virtual power plant and terminal-communication system reconstruction upgrade scheme are obtained by using CPLEX software solution;Whether the reliable capacity supply and climbing rate obtained meet the needs of power grid frequency safety is verified.The application helps virtual power plant to provide reliable frequency modulation capacity economically and efficiently, assists power grid frequency regulation and improves power grid frequency safety.

Description

Technical Field

[0001] This invention relates to the field of power grids, and more particularly to a method for improving power grid frequency security based on virtual power plants. Background Technology

[0002] In many regions, the trend towards building power systems heavily reliant on renewable energy is increasingly evident, representing part of the effort to transition to low-carbon electricity supply. However, the integration of large amounts of renewable energy into the grid reduces system rotational inertia and increases the uncertainty of power supply. Consequently, the pressure on grid frequency regulation is mounting. Traditional thermal power units, due to their long response time delays and low slopes, cannot maintain grid frequency stability using conventional regulation methods alone. Therefore, innovative solutions are needed. One effective approach is to construct virtual power plants to support frequency regulation. Virtual power plants utilize advanced measurement and communication technologies, making full use of flexible resources such as electric vehicles and air conditioning units. This allows flexible resources to be rapidly adjusted and participate in power dispatch and frequency regulation.

[0003] Despite the implementation of some virtual power plant demonstration projects, the number of virtual power plants supporting grid frequency regulation remains relatively small, and their frequency regulation capabilities are often unreliable and unsatisfactory. One reason is that the controllable power of most individual resources is small and random, and although they are numerous, their access is scattered, making it difficult to directly utilize their regulation capabilities. Another reason is that information systems often lack real-time performance and reliability, failing to meet the efficient interactive requirements of frequency regulation. To address these challenges, it is necessary to coordinate and plan to form a large resource pool and an information system to support these resources' participation in frequency regulation. Resources in the pool can then be further selected during system operation to achieve short-term control objectives. This coordination essentially involves planning a terminal-communication system for the virtual power plants to coordinate with the frequency regulation resource pool.

[0004] Despite extensive research on virtual power plant construction, limitations remain in several areas. Current research suffers from insufficient coordination between the physical and information layers of virtual power plants, and existing terminal-communication optimization methods have not effectively quantified the joint risks arising from information-physical uncertainties in the hierarchical architecture of virtual power plants based on edge energy management units. These issues make it difficult to reliably achieve the expected frequency regulation targets in the hierarchical architecture of virtual power plants, thus failing to meet the needs of grid frequency security. Summary of the Invention

[0005] This invention provides a method for improving power grid frequency security based on virtual power plants. This invention fully utilizes historical information from flexible resources to optimize and modify the virtual power plant terminal-communication system, achieving the goals of assisting power grid frequency regulation, ensuring power grid frequency security, and reliable operation. Details are described below:

[0006] A method for improving power grid frequency security based on virtual power plants, the method comprising:

[0007] In the hierarchical virtual power plant architecture, the minimum probability scenario method based on the Raida criterion and the failure mode and effects analysis method are used to establish a reliable capacity mathematical model of the virtual power plant considering information-physical uncertainty for the selected area.

[0008] With the goal of minimizing investment, and taking normal capacity, reliable capacity and latency under cyber-physical uncertainty as constraints, a virtual power plant terminal-communication system transformation planning model is constructed.

[0009] Based on mathematical transformation, the planning model is converted into a mixed integer linear programming model, and the CPLEX software is used to solve for the reliable frequency regulation capacity of the virtual power plant and the upgrade scheme for the terminal-communication system.

[0010] By comparing the estimated frequency regulation capacity demand with the reliable capacity supply of the virtual power plant, we can verify whether the obtained reliable capacity supply and ramp rate meet the grid frequency security requirements.

[0011] The modeling process for establishing a mathematical model of the reliable capacity of a virtual power plant considering information-physical uncertainties, using the minimum probability scenario method based on the Raida criterion and the failure mode and effects analysis method, is as follows:

[0012] The scenarios are classified according to the temporal characteristics of resource potential; for each feature class, the minimum adjustable capacity of virtual power plant resources under potential uncertainty is estimated using the Raida criterion.

[0013] Based on the evaluation results of the Raida criterion, the minimum adjustable capacity of virtual power plant resources under information system failure is estimated using the failure mode and effects analysis method, thus forming a reliable capacity model for virtual power plant.

[0014] Furthermore, the final mathematical model for the reliable capacity of the virtual power plant is as follows:

[0015]

[0016]

[0017]

[0018]

[0019] in, It is a scene class s k The reliable capacity of the virtual power plant is given, where ρ1 is the conservative preference coefficient. It is a scene class s k Virtual power plant capacity expectation after information link failure. It is a scene class s kMinimum capacity of the virtual power plant after a downlink failure, where N is the number of nodes. It is a scene class s k The average historical capacity of the virtual power plant. z is the historical average adjustable capacity of resource j under node i. i,j y is the decision variable for whether or not the resource is selected. i These are the decision variables for whether to establish backup communication, where α0 is the annual failure rate of the edge terminal, β0 is the annual failure rate per unit length of optical fiber, and D... d It is the length of line d. It is the capacity after a fault in terminal b in the fault number set K. Minimum value, After fiber optic cable D failure Minimum value.

[0020] The virtual power plant terminal-communication system upgrade planning model for improving power grid frequency security is as follows:

[0021] min C inv =C edge +(C of +C wir +C lo )+C sf

[0022] stz i,j ≤x i

[0023] y i ≤x i

[0024]

[0025]

[0026]

[0027] T c,i ≤T0

[0028] Among them, C edge It is the cost of edge devices; (C) of +C wir +C lo C is the cost of communication upgrades; sf It is the incentive cost of frequency-adjustable contracts; These are various capacity target constraints, T c,i The delay is controlled for node i, and T0 is the delay requirement.

[0029] Furthermore, the mathematical transformation method for converting the programming model into a mixed-integer linear programming model involves the product of multiple 0-1 variables, with n 0-1 variables h. i The product H is transformed by the following formula:

[0030]

[0031] The verification of whether reliable capacity supply and ramp-up rate meet the requirements of power grid frequency security is as follows:

[0032] Calculate the pre-planning load for each scenario based on 5% of the maximum predicted annual load for each scenario. k The frequency modulation capacity demand that appears below And the reliable capacity supply of the virtual power plant planned under the corresponding scenario. Compare to determine whether frequency security has been improved to meet the requirements: if for any s k Reliable capacity obtained from virtual power plant planning All exceed the frequency modulation capacity requirement. This indicates that frequency security is guaranteed; conversely, more flexible resources from other regions should be incorporated to improve the frequency security of the power grid. Regarding the ramp rate, the ramp rate of each controlled resource is tested through actual regulation to see if it meets the requirements. If each individual meets the requirements, the ramp rate test is passed.

[0033] The beneficial effects of the technical solution provided by this invention are:

[0034] (1) By combining the Raida criterion, the minimum probability scenario method, and the failure mode and effects analysis method, this invention establishes a reliable capacity mathematical model for virtual power plants that considers information-physical uncertainty. This model can better assist virtual power plants in conducting reliable frequency regulation capacity analysis during planning and operation, and thus assist in analyzing system frequency security.

[0035] (2) The virtual power plant terminal-communication system collaborative planning model proposed in this invention can effectively reduce investment waste, support virtual power plant frequency regulation more economically and efficiently, and improve the frequency security of the power grid.

[0036] (3) Compared with other heuristic solution methods for uncertainty models, this invention uses a proposed mathematical method to convert the planning model into a mixed integer linear programming model, so that the model can be solved by CPLEX software, and can provide a globally optimal transformation scheme for virtual power plant terminals and communications to improve frequency security.

[0037] Therefore, this invention can make full use of existing historical potential information on resources to provide a reliable and more economical virtual power plant terminal-communication transformation scheme for large-scale virtual power plants, so that the power grid frequency regulation has sufficient reliable capacity, thereby improving the power grid frequency security. Attached Figure Description

[0038] Figure 1 The flowchart shows a method for improving grid frequency security based on virtual power plants.

[0039] Figure 2 This is a schematic diagram of the hierarchical architecture of the virtual power plant for power grid frequency regulation function of the present invention;

[0040] Figure 3 This is a schematic diagram of the virtual power plant planning architecture for the power grid frequency regulation function of this invention;

[0041] Figure 4 A schematic diagram illustrating the frequency modulation capacity requirements of the IEEE-33 node system;

[0042] Figure 5 This is a schematic diagram of the terminal and communication planning results of the method of the present invention;

[0043] Figure 6 This is a schematic diagram of the capacity planning results of the method of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below.

[0045] To address the problems existing in the background technology, this invention utilizes a stochastic method based on the Laida criterion and failure mode and effects analysis to establish a virtual power plant reliability capacity constraint and terminal-communication system transformation planning model. This model can effectively quantify the information-physical joint risks in the hierarchical structure of the virtual power plant, forming a virtual power plant with a certain reliable capacity to support frequency regulation, thus providing reliable support for power grid frequency security.

[0046] Example 1

[0047] A method for improving grid frequency security based on virtual power plants, see [link to relevant documentation]. Figure 1 The method includes the following steps:

[0048] 101: Estimate the frequency regulation capacity demand of the selected area feeder caused by its own source load fluctuations, and use it as the capacity constraint value for subsequent planning models;

[0049] Specifically, based on 5% of the maximum predicted annual load for different periods of each season in the region, the potential frequency regulation capacity demand for different periods of each season before planning is calculated.

[0050] 102: Group similar flexible resources with similar geographical locations into the same cluster, calculate the historical adjustable potential and statistical information of each cluster based on the historical electricity consumption data of each resource, and form a planning dataset;

[0051] Individual flexible resources of the same type that are geographically close are grouped into the same cluster, such as an electric vehicle charging station or an HVAC cluster in a smart building. Based on frequency regulation market requirements and statistical formulas, the historical adjustable potential ΔP of each cluster is calculated. i,j,t In addition to statistical information, which mainly considers the mean and standard deviation, when the historical potential of a resource cluster unit in a certain scenario does not meet the normal distribution, the minimum unit potential in subsequent reliable capacity calculations is taken as the mean and the standard deviation is taken as 0.

[0052] 103: In the hierarchical virtual power plant architecture, the minimum probability scenario method and failure mode and effects analysis method based on the Raida criterion are used to establish a reliable capacity mathematical model of the virtual power plant considering information-physical uncertainty for the selected area, providing a basis for establishing reliable capacity constraints in the planning model;

[0053] Among them, a fast and reliable layered virtual power plant architecture, such as Figure 2 As shown, the reliable capacity of a virtual power plant is defined as the weighted average of the conservative capacity and the expected capacity under information-physical uncertainty. The conservative capacity is calculated by quantifying the maximum impact of information-physical uncertainty in the capacity model, while the expected capacity is calculated by substituting the mean of each input within the scenario class into the capacity model.

[0054] Furthermore, in step 102, the reliable capacity considering information-physical uncertainty is quantitatively modeled using the minimum probability scenario method and failure mode and effects analysis method based on the Laida criterion. The basic steps are as follows:

[0055] 1031: Classify scenarios based on the temporal characteristics of resource potential, such as season, time period, weekday, etc.

[0056] This involves combining seasonal features, time-of-day features, and weekday features, with each combination considered as a scenario category, and each category of features constituting a feature class s. k That is, scene-based.

[0057] 1032: For each feature class s obtained in the previous step k In this study, the adjustable capacity of virtual power plant resources under potential uncertainty is estimated using the Raida criterion. The minimum value;

[0058] 1033: Based on the evaluation results of the Raida criterion, the adjustable capacity of virtual power plant resources under information system failure is estimated using Failure Mode and Effects Analysis. The minimum value is used to form a virtual power plant reliable capacity model.

[0059] 104: With the goal of minimizing investment, and taking normal capacity, reliable capacity, and latency under cyber-physical uncertainty as constraints, a virtual power plant terminal-communication system transformation planning model is constructed to improve the frequency security of the power grid.

[0060] Among them, the system planning architecture is as follows Figure 3 As shown, the investment consists of three parts: communication upgrade cost, terminal configuration cost, and resource mobilization incentive, which constitute the objective function; various capacity constraints are established based on the capacity model and capacity demand forecast; finally, latency is modeled and constraints are established.

[0061] 105: Based on mathematical transformation, the planning model is converted into a mixed integer linear programming model, and CPLEX software is used to solve it. The solution yields the reliable frequency regulation capacity of the virtual power plant and the upgrade scheme for the terminal-communication system.

[0062] After linearization, the model can be directly solved using the CPLEX software.

[0063] 106: Compare the predicted frequency modulation capacity demand Reliable capacity supply with virtual power plants The test verifies whether the reliable capacity supply and ramp rate meet the grid frequency security requirements.

[0064] Frequency security can be assessed by comparing the difference between capacity demand and planned supply in various scenarios. The pre-planning load (s) for each scenario is calculated based on 5% of the maximum predicted annual load for each scenario. k The following may result in frequency modulation capacity requirements And the reliable capacity supply of the virtual power plant planned under the corresponding scenario. Compare to determine whether frequency security has been improved to meet the requirements: if for any s k Reliable capacity obtained from virtual power plant planning All exceed the frequency modulation capacity requirement. This indicates that frequency security is guaranteed; conversely, it indicates that although the plan has improved frequency security, it still cannot meet the power grid's frequency security requirements, necessitating an expansion of the planning area and the selection of regions with more flexible resources to improve power grid frequency security. Regarding ramp rate, actual control is used to verify whether the ramp rate of each controlled resource meets the requirements. If each individual meets the requirements, the ramp rate test passes; otherwise, these resources are excluded and replanned.

[0065] In summary, by combining the Raida criterion, the minimum probability scenario method, and the fault mode and effects analysis method, the embodiments of the present invention reduce the conservatism of the optimization method, which can help plan the virtual power plant terminal-communication system, effectively increase the proportion of reliable frequency regulation capacity of the virtual power plant, and improve the frequency security of the power grid.

[0066] Example 2

[0067] The method proposed in this invention embodiment is used to plan a virtual power plant terminal-communication system supporting frequency modulation at the feeder level and above. The following is a combination of... Figure 1 , Figure 2 , Figure 3 The specific calculation formulas for the scheme in Example 1 are further described below:

[0068] 201: Estimate the frequency regulation capacity demand of the selected area feeder caused by its own source load fluctuations, and use it as the capacity constraint value for the subsequent planning model;

[0069] In various scenarios, the frequency regulation capacity requirement is generally taken as a percentage of the annual maximum predicted load for the corresponding scenario. This invention calculates the possible frequency regulation capacity requirement for different periods of each season in the region before planning, based on 5% of the annual maximum predicted load for different periods of each season.

[0070] 202: Group similar flexible resources with similar geographical locations into the same cluster, calculate the historical adjustable potential and statistical information of each cluster based on the historical electricity consumption data of each resource, and form a planning dataset;

[0071] Calculate the historical potential ΔP of each resource i,j,t :

[0072] △P i,j,t =(P max,i,j,t -P min,i,j,t ) / twenty one)

[0073] Among them, P max,i,j,t and P min,i,j,t These are the maximum and minimum power of cluster j that satisfies the minimum satisfaction at time t, respectively. The statistical information mainly considers the mean and standard deviation. When the historical potential of a resource cluster unit in a certain scenario does not meet the normal distribution, the minimum unit potential in the subsequent reliable capacity calculation is taken as the mean and the standard deviation is taken as 0.

[0074] 203: In the hierarchical virtual power plant architecture, the minimum probability scenario method and failure mode and effects analysis method based on the Raida criterion are used to establish a reliable capacity mathematical model of the virtual power plant considering information-physical uncertainty for the selected area, providing a basis for establishing reliable capacity constraints in the planning model;

[0075] Figure 2The layered virtual power plant architecture consists of a physical resource layer, a bottom terminal layer, a downlink communication layer, an edge layer, an uplink communication layer, and a cloud platform. Physical resources available for frequency regulation include various distributed power sources, energy storage devices, electric vehicles, and air conditioners. Local terminals are measurement and control devices, such as smart inverters, smart switches, and energy management terminals. The downlink communication layer handles short-range communication between node edge devices and local terminals, using both wired and wireless methods. For resources with high reliability and real-time requirements, a dual-mode communication method combining wired and wireless can be used. Node edge devices, such as energy management units, collect data from local terminals, assess the node's regulation capabilities, and report to the superior substation via the uplink communication layer. The substation interacts with the grid dispatch platform, distributing regulation tasks from each node to the edge terminals. The edge terminals quickly decompose and allocate tasks to various flexible resources. Furthermore, edge terminals can transmit information to the virtual power plant cloud platform via low-speed communication to assist in regulation capacity prediction and reporting. The uplink communication layer mainly includes Ethernet Passive Optical Network (EPON), Industrial Ethernet, and 4G / 5G communication.

[0076] The reliable capacity of a virtual power plant is defined as the weighted average of the conservative capacity and the expected capacity of the virtual power plant under information-physical uncertainty.

[0077]

[0078]

[0079]

[0080]

[0081] in, It is a scene class s k The reliable capacity of the virtual power plant is given, where ρ1 is the conservative preference coefficient. It is a scene class s k Virtual power plant capacity expectation after information link failure. It is a scene class s k Minimum capacity of the virtual power plant after a downlink failure, where N is the number of nodes. It is a scene class s k The average historical capacity of the virtual power plant. z is the historical average adjustable capacity of resource j under node i. i,j y is the decision variable for whether or not the resource is selected. i These are the decision variables for whether to establish backup communication, where α0 is the annual failure rate of the edge terminal, β0 is the annual failure rate per unit length of optical fiber, and D... d It is the length of line d. It is the capacity after a fault in terminal b in the fault number set K. Minimum value, After fiber optic cable D failure Minimum value.

[0082] The impact of cyber-physical uncertainty in the capacity model is quantified using the minimum probability scenario method based on the Laida criterion and failure mode and effects analysis. The basic steps are as follows:

[0083] 1) Classify scenarios based on the temporal characteristics of resource potential, such as season, time period, and weekday. The normal adjustment capability of a virtual power plant refers to its adjustment capability under fault-free conditions, applicable to frequency regulation services, i.e., frequency regulation capability. The adjustment capability of flexible resources is greatly affected by the environment and has temporal characteristics; therefore, it is necessary to construct scenario classifications to reduce system errors at various levels. Taking electric vehicle charging stations and HVAC systems as examples, charging station scenario classification generally considers time period labels. Time periods are generally in 1-hour intervals. Building HVAC systems can also include seasonal labels.

[0084] In the planning process, each time period can be divided into a scenario category. This classification reduces the impact of the system environment within a category. Assume the final scenario category set is S = {s1, s2, ..., s...} k ,…,s d}. s k This is the set of historical scenes under the k-th tag combination, representing a scene category. Based on the time tags mentioned above, d = 4 × 24 = 96. If we add the month and weekday tags, then d = 12 × 2 × 24 = 576. It is a scenario category s k The total regulating capacity of the virtual power plant can be expressed in the following form:

[0085]

[0086] Where N is the number of nodes, m i Let i be the number of flexible resource clusters. For the adjustable resource cluster j of node i in scenario category s k Adjustable power.

[0087] 2) Raida Criterion for Estimating the Potential of Virtual Power Plant Resource Clusters under Potential Uncertainty Conservative value:

[0088]

[0089] in, The historical adjustable capacity of resource j under node i Standard deviation.

[0090] 3) Based on step 2), use Failure Mode and Effects Analysis (FMEA) to estimate the conservative capacity of the virtual power plant. The analysis results are as follows:

[0091]

[0092]

[0093] 204: With the goal of minimizing investment, and taking normal capacity, reliable capacity, and latency under cyber-physical uncertainty as constraints, we construct a virtual power plant terminal-communication system transformation planning model to improve power grid frequency security.

[0094] Among them, the terminal-communication system planning architecture is as follows: Figure 3 As shown, the overall planning model is as follows:

[0095]

[0096] Among them, C edge It is the cost of edge devices; (C) of +C wir +C lo C is the cost of communication upgrades; sf It is the incentive cost of frequency adjustment contracts. These are various capacity target constraints, T c,i The delay is controlled for node i, and T0 is the delay requirement. A more detailed explanation of the objective function follows:

[0097] 1) Edge terminal C edge The cost can be expressed as:

[0098]

[0099] Where f1 is the sum of the unit prices of the edge device and the ONU.

[0100] 2) Communication upgrade costs include fiber optic investment C of Wireless communication cost C wir and downlink communication upgrade cost C lo C of It can be represented as:

[0101]

[0102] Where f2 is the price per unit length of optical cable. Z is the set of wire numbers for existing optical fibers.

[0103] The wireless communication cost C of the uplink communication layer wir It can be represented as follows:

[0104]

[0105] Here, f3 represents the sum of the annual fee for the edge device's wireless package and the fee for the 4G / 5G communication module. The underlying device's communication module has been upgraded to dual-mode communication.

[0106] C lo Specifically as follows:

[0107]

[0108] Here, f4 represents the upgrade price for a single communication module. To reduce costs, each group will only have one module installed free of charge for the cluster owner to try out; the cost of the remaining modules will be borne by the cluster owner.

[0109] 3) To incentivize more high-capacity, flexible resources to participate in frequency regulation services and generate revenue, a contract incentive fee can be provided to the selected resources. This embodiment of the invention primarily calculates the one-time investment for frequency regulation in the following year to provide an initial budget plan. Resource contract incentive fee C sf It is expressed as follows:

[0110]

[0111] Here, f5 is the contract compensation for adjustable unit power. It is a scene class s k The percentage of time it occurs throughout the year.

[0112] Based on the capacity model and capacity demand forecast, various capacity constraints are established. According to equation (9), when conservative values ​​are mainly considered, the reliable capacity constraint is transformed into:

[0113]

[0114]

[0115] Finally, the delay is modeled and constraints are established:

[0116] Control latency is also an important metric for virtual power plants. This embodiment of the invention primarily considers the internal latency T of the virtual power plant's terminal-communication system. c,i :

[0117]

[0118] Where τ is the fixed node delay. G 2,i It is the set of nodes that the power path of node i passes through. i This represents the number of communication nodes traversed by the power path. v is the speed of light in the optical fiber. Then, the following additional constraints are added:

[0119]

[0120] Where T0 is a constraint value. Considering that HPLC has a large bandwidth and downlink communication is short-distance communication with relatively small latency, this embodiment of the invention does not specifically consider this latency.

[0121] 205: Based on mathematical transformation, the planning model is converted into a mixed integer linear programming model, and CPLEX software is used to solve it. The solution yields the reliable frequency regulation capacity of the virtual power plant and the upgrade scheme for the terminal-communication system.

[0122] The mathematical method proposed for model transformation is as follows:

[0123] The model involves the product of multiple 0-1 variables, with n 0-1 variables h. i The product H can be linearized using the following formula:

[0124]

[0125] By coordinating resource selection, edge terminal allocation, and communication planning, this planning approach can meet the reliable capacity requirements of virtual power plants and support rapid and reliable adjustments to flexible resources.

[0126] 206: By comparing predicted frequency modulation capacity requirements With the planned virtual power plant reliable capacity supply The test verifies whether the reliable capacity supply and ramp rate meet the grid frequency security requirements.

[0127] In various scenarios, the frequency regulation capacity requirement is generally taken as a percentage of the maximum predicted annual load for the corresponding scenario. In this embodiment of the invention, the pre-planning load s for each scenario is calculated based on 5% of the maximum predicted annual load for each scenario. k The following may result in frequency modulation capacity requirements And the reliable capacity supply of the virtual power plant planned under the corresponding scenario. Compare to determine whether frequency security has been improved to meet the requirements: if for any s k Reliable capacity obtained from virtual power plant planning All exceed the frequency modulation capacity requirement. This indicates that frequency security is guaranteed; conversely, it indicates that although the plan has improved frequency security, it still cannot meet the power grid's frequency security requirements, necessitating an expansion of the planning area and the selection of regions with more flexible resources to re-control power grid security. Regarding ramp rate, actual control is used to verify whether the ramp rate of each controlled resource meets the requirements. If each individual meets the requirements, the ramp rate test passes; otherwise, these resources are excluded and re-planned.

[0128] Example 3

[0129] The following will use specific examples... Figures 3 to 6Tables 1 to 4 verify the feasibility of the schemes in Examples 1 and 2, as detailed below:

[0130] 301: IEEE-33 Node Test System

[0131] 1) System Setup: The flexible resources to be contracted mainly include gas turbines (GT), charging piles (CP), and commercial buildings using heating, ventilation, and air conditioning (HVAC). The distribution and capacity of each type of resource are shown in Table 1. Two types of CP are considered: Category I is designated for companies or shopping malls with frequent daytime use, and Category II is for blocks with frequent nighttime use. The HVAC system has a fixed regulation range in winter and summer.

[0132] Table 1 Resource Distribution

[0133]

[0134]

[0135] 2) Power Configuration: The system's maximum annual load is 3.715 MW. Since this demand is divided by season and time period, the scenario classification in this embodiment of the invention also follows this division, resulting in a total of 24 × 4 = 96 scenarios. For example... Figure 4 As shown, 5% of the total load under each scenario is taken as the frequency regulation capacity construction requirement for the power distribution system.

[0136] The system's backup capacity can already meet 20% of the maximum capacity requirement. The controllable time for the HVAC system is set to 8:00-20:00 in summer and to remain on in winter.

[0137] 3) Other parameter settings: Assuming a node spacing of 0.8km, f1 = ¥5000, f2 = ¥6 / m. Because GT power supply is more stable, a higher contract incentive price can be set. The system is set to operate for 10 years. f3 = ¥1000, f4 = ¥200, f... 5,GT =200¥ / kW,f 5,load =100¥ / kW, M0=50, T0=10ms, τ=0.1ms, v=200000km / s. For system faults, the calculations in this embodiment mainly consider conservative cases, i.e., ρ1=0, λ=0.1, a0=β0=1%. The declared price for frequency regulation capacity is set at 60¥ / MW. The shortage penalty price is set at 10 times the declared price.

[0138] The selection results of the adjustable resource clusters are shown in Table 2, and the terminal-communication network configuration results are as follows: Figure 5 As shown, most nodes containing GT (Ground Controller) are placed at the edge because GT can provide more stable output compared to other resources. Node 33 was not selected because it is too far from 26 and there are no other resource nodes nearby, resulting in wasted communication costs. Although the capacity of node 9 is smaller than that of node 4, increasing the capacity of node 9 can meet the needs, while configuring node 4 would increase economic costs; therefore, node 9 is prioritized over the edge. In addition, appropriately giving up larger power points can reduce losses from edge device failures. CP I is selected significantly more often than CP II because daytime demand is much higher than nighttime demand. Some HVAC systems, despite having large capacities, were not selected because they are unusable during certain periods and seasons. In summary, the optimization results are the result of a combination of factors including regulation capacity, communication costs, and uncertainties.

[0139] Table 2 Resource Selection

[0140]

[0141] Figure 6 This shows the minimum net surplus of adjustment capacity under various scenarios of information system failure. The calculation method is to subtract the corresponding demand from the total capacity in each scenario, representing the reliable capacity surplus. When it is greater than 0, it indicates that the grid frequency security has improved to a level that can meet frequency stability requirements. Experiments revealed that terminal failures have a greater impact in summer and winter, while fiber optic failures have a greater impact in spring and summer. This is because HVAC systems occupy a large proportion of regulation capacity in summer, making node failures more significant. Preparing backup communication can mitigate the impact of fiber optic failures. In summary, all net surplus constraint function values ​​are greater than 0, and many points are close to 0, indicating that the frequency regulation capacity provided by the planned virtual power plant in various scenarios exceeds the capacity gap, demonstrating that frequency regulation has sufficient capacity guarantees and frequency security has improved.

[0142] This invention compares several planning schemes to evaluate the superiority of the proposed method, as detailed in Table 3. In Table 3, Scheme 1 corresponds to the proposed method. Due to the very short delay time, the comparison mainly focuses on the investment and adjustment capabilities under various conditions to highlight the scheme's superior economic efficiency in meeting frequency security requirements.

[0143] Table 3 Description of Different Planning Schemes

[0144]

[0145] Table 4 Comparison Results

[0146]

[0147] The net surplus of several capabilities in Table 4 is the result of taking the minimum value of all schemes. Since schemes 2, 3, and 4 do not simultaneously consider the joint risks caused by cyber-physical uncertainties, the minimum net surplus after a failure is less than 0. This indicates that if cyber-physical uncertainties are not considered simultaneously, the superposition effect may lead to the virtual power plant failing to complete its expected regulation tasks, or even resulting in a huge capacity gap and grid frequency security issues. When extreme weather events occur on multiple virtual power plants on the same day, insufficient backup power will have a significant impact on the safe operation of the grid. Therefore, both cyber-physical and physical uncertainties must be considered. Furthermore, by comparing the investments under these scenarios, it can be found that considering cyber-physical uncertainties with the help of system optimization does not necessarily increase investment significantly. Table 4 also calculates the internal rate of return (IRR) through operation and failure simulations during the operating life. In scheme 4, the IRR is 8% lower than the generally required value, failing to meet normal project investment standards. The proposed method has the highest IRR, indicating that it is highly beneficial for the construction of virtual power plant projects. In summary, the proposed method improves frequency security with better economic efficiency, achieving safe and stable grid operation.

[0148] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0149] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for improving power grid frequency security based on a virtual power plant, characterized by, The method includes: In the hierarchical virtual power plant architecture, the minimum probability scenario method based on the Raida criterion and the failure mode and effects analysis method are used to establish a reliable capacity mathematical model of the virtual power plant considering information-physical uncertainty for the selected area. With the goal of minimizing investment, and using normal capacity, reliable capacity, and latency under cyber-physical uncertainty scenarios as constraints, a virtual power plant terminal-communication system transformation planning model is constructed. Based on mathematical transformation, the planning model is converted into a mixed integer linear programming model, and the CPLEX software is used to solve it to obtain the reliable frequency regulation capacity of the virtual power plant and the terminal-communication system transformation and upgrade scheme. Compare the estimated frequency regulation capacity demand with the reliable capacity supply of the virtual power plant to verify whether the obtained reliable capacity supply and ramp rate meet the grid frequency security requirements; The modeling process for establishing a mathematical model of the reliable capacity of a virtual power plant considering information-physical uncertainties, using the minimum probability scenario method based on the Laida criterion and the failure mode and effects analysis method, is as follows: The scenarios are classified according to the temporal characteristics of resource potential; for each feature class, the minimum adjustable capacity of the virtual power plant resource cluster under potential uncertainty is estimated using the Raida criterion. Based on the evaluation results of the Raida criterion, the minimum adjustable capacity of the virtual power plant under uncertain faults of the information system is estimated using the failure mode and effects analysis method, thus forming a reliable capacity model for the virtual power plant. The final mathematical model for the reliable capacity of the virtual power plant is as follows: ； ； ； ； in, It is a scene class s k Virtual power plant reliable capacity It is the conservative preference coefficient. It is a scene class s k Virtual power plant capacity expectation after information link failure. It is a scene class s k Minimum capacity of the virtual power plant after a downlink failure, where N is the number of nodes. It is a scene class s k The average historical capacity of the virtual power plant. It is the historical average adjustable capacity of resource j under node i. It is the decision variable for whether or not the resource is selected. It is a decision variable for whether or not backup communication is established. It is the annual failure rate of edge terminals. This refers to the annual failure rate per unit length of optical fiber. It is the length of line d. It is the capacity after a fault in terminal b in the fault number set K. Minimum value, After fiber optic cable D failure Minimum value; The virtual power plant terminal-communication system transformation planning model is as follows: ； s.t. ; ； ； ； ； ； wherein, is the edge device cost; is the communication upgrade cost; is the incentive cost of the frequency regulation contract; are the various capacity target constraints, is the control delay for node i, is the delay requirement; The mathematical transformation method for converting the programming model into a mixed-integer linear programming model involves the product of multiple 0-1 variables, with n 0-1 variables h. i The product H is transformed using the following formula: 。 2. The method for improving power grid frequency safety based on virtual power plant according to claim 1, characterized in that, Whether the reliable capacity supply and ramp rate obtained from the test meet the requirements of power grid frequency security is as follows: Calculate the pre-planning load for each scenario based on 5% of the maximum predicted annual load for each scenario. k The frequency modulation capacity demand that appears below And the reliable capacity supply of the virtual power plant planned under the corresponding scenario. Compare to determine whether frequency security has been improved to meet the requirements: if for any s k Reliable capacity obtained from virtual power plant planning All exceed the frequency modulation capacity requirement. This indicates that frequency security is guaranteed; Conversely, more flexible resources from other regions can be incorporated to improve the frequency security of the power grid; in terms of ramp rate, the ramp rate of each controlled resource is tested through actual regulation to see if it meets the requirements, and if each individual meets the requirements, the ramp rate test is passed.

Images