A two-stage AGC scheduling method considering adaptive mode switching

Abstract

This invention discloses a two-stage AGC scheduling method considering adaptive mode switching, belonging to the power field. First, based on the characteristics of proactive and real-time scheduling, a basic framework for two-stage scheduling is constructed. Then, three operating modes are defined by quantifying resource regulation capabilities, and adaptive switching between modes is achieved. Second, a proactive-real-time two-stage scheduling model is constructed: in the proactive stage, a variable objective optimization method is introduced based on the mode determination results to pre-determine the output level of some resources; in the real-time stage, an optimization model is constructed considering dynamic task coefficients and minimizing costs to complete the scheduling of remaining resources. Finally, a two-stage scheduling frequency response model is constructed. Simulation analysis verifies that the proposed strategy can effectively improve the frequency recovery effect and assessment quality of the regional power grid, and maintain a good state of charge for energy storage power stations.

Description

Technical Field

[0001] This invention relates to the field of power, and more specifically to a two-stage AGC scheduling method that considers adaptive mode switching. Background Technology

[0002] The high proportion of renewable energy equipment integrated into the power system is an inevitable trend for future development. However, its low inertia and strong volatility pose a serious threat to power system frequency control. Automatic generation control (AGC) is currently the main means to alleviate frequency problems and improve the quality of control performance standards (CPS) assessments. Energy storage devices, due to their short response time and fast adjustment speed, are gradually becoming commonly used auxiliary devices in AGC regulation. Therefore, considering resource regulation capabilities and system frequency regulation requirements, achieving rational scheduling of multiple types of resources, including energy storage, during the automatic generation control process has become a major issue in the development and construction of new power systems.

[0003] Currently, real-time scheduling is one of the more common scheduling methods in the design of specific AGC scheduling methods. Existing real-time scheduling strategies aim to improve AGC control performance. Existing technologies propose a regional control error mining algorithm to effectively improve system frequency regulation performance and achieve the desired active power allocation. Further, existing technologies propose a method for coordinating frequency regulation using battery energy storage systems to improve the AGC performance of such generators. Simultaneously, the control strategy should consider the capacity constraints of the energy storage itself to extend its operating duration. Existing technologies design controllers to manage the state of charge (SOC) of the energy storage system and avoid overcharging and over-discharging. Existing technologies propose a two-layer optimization strategy for frequency regulation power that considers frequency regulation costs and SOC recovery, preventing extreme charging and discharging levels of energy storage and significantly improving frequency regulation performance. Furthermore, considering the differences in resources participating in the frequency regulation process, existing strategies decompose high- and low-frequency commands to fully utilize the advantages of various resource characteristics. Overall, real-time scheduling has the advantages of short time scale and high regulation accuracy; however, insufficient resource flexibility and errors caused by lag control indicate that there is still considerable room for improvement in the frequency control process.

[0004] To overcome the aforementioned shortcomings, advance scheduling has gradually become another commonly used scheduling method in current AGC scheduling processes. Existing technology proposes a real-time frequency regulation method based on frequency regulation zone scheduling, effectively ensuring the frequency security and stability of the power grid. Existing technology is based on random N... 1. Constraint-based scheduling models are established using resource-constraint networks, providing more effective control strategies. Simultaneously, to improve the economy of the scheduling process, the frequency regulation cost of resources is incorporated into the scheduling model for unified optimization. Existing technologies establish discrete-continuous stochastic look-ahead economic scheduling models to achieve optimal coordination of fast and slow resources, reducing the total frequency regulation cost. Existing technologies propose a power uncertainty decomposition supply model, whose control effect can simultaneously improve system efficiency and frequency reliability. Furthermore, given the continuously expanding scale of AGC control, centralized algorithms have high computational requirements; therefore, distributed solution methods have gradually become a research hotspot in look-ahead scheduling, significantly improving model solution efficiency. Overall, look-ahead scheduling overcomes the lag problem of real-time scheduling to some extent, but the solution effect of look-ahead scheduling depends on the accuracy of prediction and has a long time span, thus scheduling errors still exist.

[0005] In response to this, combining the advantages and disadvantages of both proactive and real-time scheduling, a two-stage coordinated scheduling method has been gradually applied in recent years. Existing technologies, to maintain a good energy storage state, utilize thermal power unit frequency control and restore energy storage SOC in the proactive stage, and further utilize refined frequency regulation of energy storage in the real-time stage, thereby effectively improving system frequency and reducing energy storage degradation. Building upon this, existing technologies consider the potential command conflicts between the two scheduling methods, allocating resources based on historical experience, and then modeling and scheduling each stage separately, fully leveraging the resource regulation capabilities. The existing two-stage scheduling method achieves complementary advantages between the two, further improving frequency regulation performance. However, the current two-stage scheduling method has relatively fixed resource allocation and task assignment, which may lead to insufficient coordination between resources. Summary of the Invention

[0006] To address the aforementioned shortcomings in the prior art, this invention provides a two-stage AGC scheduling method that considers adaptive mode switching.

[0007] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0008] A two-stage AGC scheduling method considering adaptive mode switching includes the following steps:

[0009] S1. Construct a two-stage basic scheduling architecture consisting of a proactive scheduling stage and a real-time scheduling stage;

[0010] S2. Develop the operating mode of the AGC system based on resource regulation capacity and power grid regulation requirements, and perform adaptive switching between modes;

[0011] S3. Based on the operation mode determination results, construct an advanced scheduling optimization model to pre-determine the output of some resources;

[0012] S4. Construct a real-time scheduling optimization model, taking into account dynamic task coefficients, and complete the power allocation of remaining resources;

[0013] S5. Construct a two-stage scheduling frequency response model and verify the effectiveness of the strategy through simulation.

[0014] Furthermore, in S1:

[0015] In the advance scheduling phase, with a time granularity of 1 minute, unbalanced power is obtained based on ultra-short-term prediction, and an optimization model for power balance, resource regulation capability, and CPS assessment constraints is constructed.

[0016] The real-time scheduling phase uses a time granularity of 4 seconds. Based on the advance scheduling of fixed part of the resource output, frequency deviation and tie line deviation are collected to calculate ACE and construct a real-time power allocation model.

[0017] Furthermore, step S2 specifically includes the following steps:

[0018] S21. Quantitative upper and lower limits of conventional unit regulation capacity P RGU,t , P RGL,t :

[0019]

[0020] In the formula, P RGi,t-1 , R rateG,i , P Gi,max , P Gi,min They represent the first i The power value, maximum grade rate, maximum and minimum power of a conventional unit at the previous moment; m Indicates the total number of conventional generating units; The scheduling time interval;

[0021] S22. Introducing 0-1 decision variables T 1. Determine whether conventional generating units can support grid regulation requirements:

[0022] when hour T 1=1, otherwise T 1=0, when T When 1=0, priority is given to allocating energy storage to participate in advanced dispatch until... T 1=1; Introducing 0-1 decision variables T 2. Determine the adjustment margin: When hour T 2 = 1; where, To adjust the margin coefficient, To adjust the deviation, Δ P t This refers to regional power imbalance fluctuations.

[0023] S23, the mode switching command satisfies:

[0024] In the formula, This is a mode switching command;

[0025] S24. Determine the operating mode based on the mode switching instruction, including:

[0026] when Mode 1: Conventional generating units participate in advanced dispatching, while the lowest state of charge energy storage power station is restored to SOC, and the remaining energy storage participates in real-time phased dispatching.

[0027] when Mode 2: Conventional generating units are scheduled in advance, and energy storage power stations participate in real-time scheduling to complete a two-stage scheduling process;

[0028] when Mode 3: Some energy storage assists conventional units in advance scheduling, while the remaining energy storage participates in real-time phase scheduling.

[0029] Furthermore, step S3 specifically includes the following steps:

[0030] S31. Introduce various resource cost functions and define variable objective functions under different operating modes;

[0031] S32, taking into account CPS assessment to determine advance scheduling constraints.

[0032] Furthermore, the variable objective function in different operating modes in S31 is expressed as follows:

[0033] Mode 1: Conventional generating units are sufficient to meet regulation demands and have ample margin. The objective is for conventional generating units to participate in power smoothing during the advanced phase, while simultaneously assisting some energy storage power stations with the lowest state of charge (SOC) to restore their State of Charge (SOC). The objective function is... F Represented as:

[0034]

[0035] In the formula, M The penalty factor for energy storage advance recovery is set to a constant, indicating that it has a higher priority; q The number of energy storage units participating in the advanced phase of SOC recovery; T The scheduling period is set to 15 minutes. c f,i Indicates the first i The auxiliary frequency regulation cost coefficient for each thermal power unit For the first j Energy storage t Timing adjustment deviation, P RGi,t Indicates the first i The power value of a conventional unit at the current moment. m Indicates the total number of conventional generating units;

[0036] Mode 2: When conventional units can meet the frequency regulation requirements during the advanced stage but the margin is insufficient, the objective function is... F Including only the frequency regulation costs of conventional units, it is expressed as:

[0037]

[0038] Mode 3: If conventional generating units are insufficient to support the demand during the advanced adjustment phase, some energy storage needs to be allocated to participate in the advanced phase. The determined number of energy storage units participating in the advanced dispatch is... p The optimization aims to minimize the combined cost of all resources participating in the advance stage scheduling, with the objective function being... F Represented as:

[0039]

[0040] In the formula, This indicates that the unit is in t Frequency tuning costs at any given time F Rbj,t It equals the sum of dynamic capacity cost, dynamic mileage cost, and lifetime depreciation cost.

[0041] Furthermore, the constraints in S32 include:

[0042] Power balance constraints are expressed as:

[0043]

[0044] In the formula, K Gi Indicates the first i The primary frequency regulation unit adjustment coefficient for a conventional generating unit, Δ f t Δ P T,t for t Frequency deviation and tie-line power deviation at specific times; Δ P t This is due to regional power imbalance fluctuations caused by uncertainties in load and new energy sources; R RGi,t Indicates the first i A conventional unit t The rate of increase at any given moment;

[0045] CPS constraints are represented as follows:

[0046]

[0047]

[0048] In the formula, and These are the hard constraint indicators based on the NERC assessment standard, and:

[0049]

[0050]

[0051] In the formula, B This is the frequency deviation coefficient for the control area; ɛ 1min The statistical value of the root mean square of the 1-minute frequency average deviation of the interconnected power grid in the previous year is taken. E ACE,t Indicates the ACE value. ɛ 15min Take the root mean square value of the frequency deviation over 15 minutes in the previous year for the interconnected power grid; B s This refers to the frequency deviation coefficient of the entire interconnected power grid; For the scheduling period, They are respectively The maximum and minimum values;

[0052] Conventional generating units, considering ramp-up constraints and power constraints, are expressed as follows:

[0053]

[0054]

[0055] In the formula, Let t be the output of the conventional unit. P Gi,max , P Gi,min These are the maximum and minimum ramp power for conventional generating units, respectively. Maximum gradeability;

[0056] Considering power and capacity constraints, an energy storage power station can be represented as follows:

[0057]

[0058]

[0059] In the formula, Indicates the first j Energy storage power station t Capacity value at any given time These are the maximum and minimum values ​​of the capacity, respectively. Rated power;

[0060] Power deviation constraint, expressed as:

[0061]

[0062]

[0063] In the formula, This represents the communication power deviation at time t. For the regional power grid frequency deviation, Δ P Tmin Δ P Tmaxx These represent the minimum and maximum power deviations of the tie line, Δ. f min Δ f max These represent the minimum and maximum frequency deviations of the regional power grid, respectively.

[0064] Furthermore, step S4 specifically includes the following steps:

[0065] S41. Define the dynamic task coefficient for the charging and discharging phase at the current moment based on the current state of charge and the regional power grid's charging and discharging requirements. and The specific calculation method is as follows:

[0066]

[0067]

[0068]

[0069] In the formula, and The following are the dynamic task coefficients for the charging and discharging phases, in order. SOC min , SOC max Represents the minimum and maximum values ​​of the state of charge. α jc,t , α jd,t These represent the state-of-charge deviation coefficients during the charging and discharging phases, respectively. w As an adaptive factor, SOC j,t Indicates the first j Energy storage t State of charge at time t, SOC ref This is the reference value for the state of charge;

[0070] S42. Define the maximum adjustable power during the charging and discharging phase of an energy storage power station. and The maximum adjustable power is adjusted based on the dynamic task coefficient change curve with SOC during the charging and discharging phase to meet the charging and discharging requirements.

[0071] S43. Considering minimizing energy storage costs, construct the objective function for the real-time scheduling phase;

[0072] S44. A real-time scheduling model is constructed taking into account constraints such as maximum adjustable power, and the total AGC instructions for the real-time stage are allocated to the remaining energy storage units based on the model scheduling results.

[0073] Furthermore, the maximum adjustable power during the charging and discharging phase of the energy storage power station in S42... and The specific calculation method is as follows:

[0074]

[0075] In the formula, and The following are the maximum adjustable power outputs during the charging and discharging phases of the energy storage power station. and The following are the dynamic task coefficients for the charging and discharging phases, in order. This is the rated power.

[0076] Furthermore, the objective function in S43 is expressed as:

[0077]

[0078] In the formula, The objective function of the real-time scheduling process is... For the first j A function for the cost of AGC (Automatic Guided Service) auxiliary services for an energy storage power station. This represents the total number of energy storage power stations. The number of energy storage units participating in advanced dispatch.

[0079] Furthermore, the real-time scheduling constraints in S44 are expressed as follows:

[0080]

[0081] In the formula, This is the general instruction for real-time stage scheduling. for t Time of the first j The output power of an energy storage power station This represents the total number of energy storage power stations. The number of energy storage units participating in advanced dispatching. These represent the maximum adjustable power during the charging and discharging phases, for t Time of the first j The capacity value of an energy storage unit. They represent the first j The maximum and minimum capacity of each energy storage power station.

[0082] The present invention has the following beneficial effects:

[0083] 1) The strategy of this invention analyzes the dynamic response process of resources under the strategy by constructing a two-stage AGC scheduling frequency response model. The results show that this strategy improves the traditional fixed resource coordination mode of AGC scheduling by setting optimization objectives according to different operating modes during the scheduling process, and fully utilizes the adjustment potential and coordination capabilities of multiple types of resources.

[0084] 2) This invention uses a single scheduling method and a two-stage scheduling method with fixed resource allocation as comparative strategies. Comparative analysis shows that, on the one hand, the strategy of this invention can effectively improve the frequency regulation capability of AGC, maintaining the frequency deviation below 0.08Hz, meeting the CPS assessment requirements. On the other hand, the strategy of this invention effectively improves various SOC indicators of the energy storage cluster during simulation, achieving a balance between system frequency improvement and energy storage SOC recovery. Attached Figure Description

[0085] Figure 1 This is a schematic diagram of the two-stage scheduling framework for AGC that considers mode switching in this invention.

[0086] Figure 2 This is a dynamic task coefficient curve diagram of an embodiment of the present invention.

[0087] Figure 3 This is a schematic diagram of the two-stage AGC scheduling frequency response model according to an embodiment of the present invention.

[0088] Figure 4 This is a graph showing the unbalanced power disturbance in an embodiment of the present invention.

[0089] Figure 5 This is a schematic diagram showing the SOC fluctuation of an energy storage power station in each scheduling cycle according to an embodiment of the present invention.

[0090] Figure 6 This is a schematic diagram of the advance scheduling of frequency modulation power commands for various adjustment resources in an embodiment of the present invention, where (a) is scheduling period 1, (b) is scheduling period 2, (c) is scheduling period 3, and (d) is scheduling period 4.

[0091] Figure 7 This is a schematic diagram illustrating the frequency deviation of various strategies under unbalanced power in an embodiment of the present invention.

[0092] Figure 8This is a schematic diagram showing the response deviation rate of power generation equipment under various strategies in an embodiment of the present invention.

[0093] Figure 9 This is a schematic diagram of the advance scheduling situation of strategy 3 in the scheduling period 3 of the present invention.

[0094] Figure 10 This invention presents a comparison of the SOC index of energy storage clusters under different strategies in the embodiments of the present invention, where (a) is the deviation of the equivalent SOC from the benchmark value, (b) is the maximum range of SOC of the energy storage cluster, and (c) is the SOC balance of the energy storage cluster. Detailed Implementation

[0095] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0096] A two-stage AGC scheduling method that considers adaptive mode switching, such as Figure 1 As shown, it includes the following steps:

[0097] S1. Construct a two-stage basic scheduling architecture consisting of a proactive scheduling stage and a real-time scheduling stage;

[0098] In this embodiment, the scheduling framework design considers both proactive and real-time scheduling, achieving complementary advantages through a two-stage scheme. The core idea is: in the proactive scheduling process, some resources are pre-emptively mobilized based on forecast information to mitigate most power imbalances; in the real-time scheduling stage, the remaining power deviations are further refined and filled. The specific work of the two stages is as follows:

[0099] 1) Advance Scheduling Phase: This phase utilizes ultra-short-term forecasts to obtain unbalanced power caused by uncertainties in resources such as load and renewable energy. Considering constraints such as power balance, resource regulation capacity, and CPS assessment, an advance AGC scheduling optimization model is constructed. To ensure the reliability of ultra-short-term forecasts, the scheduling time granularity cannot be too short; this invention uses 1 minute.

[0100] 2) Real-time scheduling phase: Based on the advance scheduling of fixed resource output, the area control error (ACE) is calculated by acquiring real-time feedback information such as frequency deviation and tie-line deviation. The controller then generates a total AGC command and further constructs a real-time power allocation model to complete the scheduling process of the remaining power generation resources. The real-time phase has a shorter scheduling granularity than the advance phase; this invention uses 4 seconds.

[0101] Advance scheduling and real-time scheduling are complementary. In the advance stage, resources can be allocated in advance based on predictive information to overcome the lag problem of real-time scheduling. In the real-time stage, the time scale is shorter and it is used to further smooth out the power deviation that still exists after advance scheduling, so as to achieve the purpose of precise scheduling.

[0102] S2. Develop the operating mode of the AGC system based on resource regulation capacity and power grid regulation requirements, and perform adaptive switching between modes;

[0103] In actual regulation, to effectively leverage the regulation potential of power generation resources, it is necessary to simultaneously consider both resource regulation capacity and grid regulation needs, and to rationally allocate resources and tasks for the two phases of advance and real-time regulation. Considering that conventional generating units typically possess strong continuous output capabilities, they are suitable for participating in the advance phase to pre-determine a portion of the base output; while energy storage resources have advantages such as fast response speed and high regulation accuracy, making them more suitable for handling rapid frequency fluctuations and short-term power imbalances in the real-time phase. Therefore, the initial resource allocation is as follows: conventional generating units participate in advance scheduling, and energy storage power stations participate in real-time scheduling. Specifically, the steps include the following:

[0104] S21. Quantitative upper and lower limits of conventional unit regulation capacity P RGU,t , P RGL,t ,

[0105] Based on the initial resource allocation, the regulation capacity of power generation resources and the grid regulation requirements are quantified. The regulation capacity of conventional generating units is related to their ramp-up capability and power output boundary. t Upper and lower limits of conventional unit power support P RGU,t , P RGL,t They can be represented as:

[0106] (1)

[0107] In the formula, P RGi,t-1 , R rateG,i , P Gi,max , P Gi,min They represent the first i The power value, maximum ramp rate, maximum and minimum power of each conventional unit at the previous moment; m Indicates the total number of conventional generating units; This is the scheduling time interval.

[0108] S22. Introducing 0-1 decision variables T 1. Determine whether conventional generating units can support grid regulation requirements.

[0109] set up t The power regulation demand of the power grid at any given time is Δ P t The allowable adjustment deviation in the AGC scheduling process is ɛ L To ensure that conventional generating units can adequately support power demand during the advance scheduling phase, the following must be met:

[0110] (2)

[0111] when hour T 1=1, otherwise T 1=0, when T When 1=0, priority is given to allocating energy storage to participate in advanced dispatch until... T 1=1; Introducing 0-1 decision variables T 2. Determine the adjustment margin: When hour T 2 = 1; where, To adjust the margin coefficient, To adjust the deviation;

[0112] When conventional generating units alone cannot support the unbalanced power demand, energy storage is allocated to participate in the advanced dispatch process. Since energy-type energy storage can provide larger energy support first, it is prioritized in the advanced dispatch process. Using the ratio of rated power to rated capacity as a metric, each energy storage station is ranked from lowest to highest ratio and allocated sequentially to the advanced dispatch stage until the grid regulation demand is met. Let the number of energy storage units participating in the advanced dispatch be... p If there are several, then their regulatory capacity has upper and lower limits. P RbU,t , P RbL,t They are represented as follows:

[0113] (3)

[0114] In the formula, E bj,max , E bj,min , P bj,rate , η cj , η dj They represent the first j The maximum capacity, minimum capacity, rated power, and charge / discharge efficiency of each energy storage power station. E bj,t-1 Indicates the previous time / initial time. j The capacity value of an energy storage unit.

[0115] Upper and lower limits of the power generation regulation capacity of all resources participating in advanced dispatch P U,t , P L,t It can be represented as:

[0116] (4)

[0117] After the energy storage equipment is involved, the following should be ultimately met:

[0118] (5)

[0119] When conventional generating units can support the ahead-of-time unbalanced power output, no energy storage is needed for proactive regulation. With sufficient regulation margin, conventional generating units can assist the energy storage power station in SOC recovery. Furthermore, 0-1 decision variables are introduced. T 2. Characterizes whether the regulation margin of conventional units is sufficient, assuming the regulation margin coefficient is... β When the following conditions are met:

[0120] (6)

[0121] This indicates that current conventional generating units have sufficient regulation margin. T Set 2 to 1, otherwise set it to 0.

[0122] S23. Based on resource regulation capacity and power grid regulation needs, the AGC system regulation process can be subdivided into three operating modes, with dynamic coordination switching performed in each scheduling cycle (15 minutes). Switching instructions... T mode The following criteria must be met:

[0123] In the formula, This is a mode switching command;

[0124] S24. Determine the operating mode based on the mode switching instruction, including:

[0125] when Mode 1: Conventional generating units participate in advanced dispatching, while assisting energy storage power stations with poor state of charge to restore their state of charge (SOC), and other energy storage units participate in real-time phased dispatching.

[0126] when Mode 2: Conventional generating units are scheduled in advance, and energy storage power stations participate in real-time scheduling to complete a two-stage scheduling process;

[0127] when Mode 3: Some energy storage assists conventional units in advance scheduling, while the remaining energy storage participates in real-time phase scheduling.

[0128] Furthermore, based on the aforementioned model classification and switching process, automatic resource allocation and correction of the objective function for the advance scheduling stage are completed, thereby determining the power output of resources participating in advance scheduling. On this basis, the remaining adjustment resources not participating in advance scheduling enter the real-time scheduling stage. Considering the adjustment capacity and cost of the resources, a power allocation model is constructed, and the total instructions for the real-time stage are then allocated to the remaining adjustment resources, completing the real-time scheduling process.

[0129] The overall scheduling framework is divided into two parts: mode autonomous determination and two-stage scheduling modeling. The two-stage scheduling framework considering mode switching is as follows: Figure 1 As shown. In the autonomous mode determination process, by collecting ultra-short-term forecast information and power generation resource operating parameters, switching commands are obtained based on the above criteria. T mode This process determines the appropriate mode for the scheduling cycle. In the two-stage scheduling modeling process, a scheduling model is established based on the mode determination results, sequentially completing precise scheduling of different resources. This section integrates the mode switching process into the scheduling framework, enabling the model to adaptively complete the reasonable allocation of resources and task assignment based on actual needs. This ensures full utilization of resource adjustment capabilities and enhances the adaptability of the AGC system to complex scenarios.

[0130] S3. Based on the operation mode determination results, construct an advanced scheduling optimization model to pre-determine the output of some resources;

[0131] This embodiment specifically includes the following steps:

[0132] S31. Introduce various resource cost functions and define variable objective functions under different operating modes;

[0133] For conventional units participating in the AGC frequency control process, let the first... i The initial power of each conventional unit is P RGi,0 ,but t The unit output at any time P geni,t Represented as:

[0134] (8)

[0135] In the formula, R RGi,k express k The rate of ascent at any given time, expressed in MW / min, Δ t =1min.

[0136] in, t Time of the first i Power command values ​​for frequency regulation of conventional generating units P RGi,t for:

[0137] (9)

[0138] The frequency regulation service cost function of a conventional generating unit can then be expressed as:

[0139] (10)

[0140] In the formula, c f,i Indicates the first i Auxiliary frequency regulation cost coefficient for each thermal power unit; F RGi,t This indicates that the unit is in t Frequency adjustment costs at any given time.

[0141] set up t Time of the first j The output power of each energy storage power station is P Rbj,t Then its capacity loss Δ E Rbj,t It can be represented as:

[0142] (11)

[0143] In the formula, the output power of the energy storage power station is positive when it is discharged.

[0144] Energy storage costs include operating costs and investment costs factored into lifetime depreciation. Operating costs further include the dynamic capacity costs associated with participating in frequency regulation ancillary services. C capj,t With dynamic mileage cost C milj,t , respectively represented as:

[0145] (12)

[0146] (13)

[0147] In the formula, c cap,j and c mil,j They represent the first j The capacity cost coefficient and mileage cost coefficient of an energy storage power station.

[0148] Furthermore, considering that the lifespan of energy storage depends on its operating mode, frequent charging and discharging will accelerate its lifespan degradation and shorten its service life. Therefore, it is necessary to address the lifespan loss caused by the AGC (Automatic Generation Control) process. C lifej,t The quantification results are related to the initial total investment cost, single charge-discharge depth, and cycle life of the energy storage power station, and are expressed as follows:

[0149] (14)

[0150] In the formula, C inv,j 、N DOD,j Indicates the first j Initial investment cost and number of cycles at 100% depth of discharge for an energy storage power station; k p The constant is obtained by fitting the number of cycles and the depth of discharge of the energy storage power station, and is generally between 1.1 and 2.2. In this invention, we take 2.

[0151] Therefore, the first j AGC auxiliary service cost function for an energy storage power station F Rbj,t It equals the sum of dynamic capacity cost, dynamic mileage cost, and lifetime depreciation cost, expressed as:

[0152] (15)

[0153] Taking into account both resource adjustment capabilities and system power requirements, the scheduling process can be divided into three operating modes, and the objective functions for advance scheduling under different modes are as follows:

[0154] 1) For Mode 1, where conventional generating units are sufficient to meet regulation demands and have ample margin, the objective is set as follows: conventional generating units participate in power smoothing during the advanced phase, while simultaneously assisting some energy storage power stations with poor state of charge (SOC) to recover their State of Charge (SOC). The objective function is... F Represented as:

[0155] (16)

[0156] In the formula, M The penalty factor for advanced recovery of energy storage is set to a large constant, indicating that it has a higher priority; q The number of energy storage units participating in the advanced phase of SOC recovery; T The scheduling period is set to 15 minutes. ɛ j,t Indicates the first j Energy storage t The difference between the time and the baseline value can be expressed as:

[0157] (17)

[0158] In the formula, SOC j,t Indicates the first j Energy storage t The state of charge at any given moment; SOC refThe reference value for the state of charge is 0.5 in this invention.

[0159] In addition, a deviation threshold is set for energy storage power stations to participate in advanced recovery. γ When the energy storage state of charge satisfies:

[0160] (18)

[0161] This indicates that the energy storage power station has a poor state of charge and needs to participate in the advance recovery phase. To prevent command conflicts, the energy storage devices participating in the advance recovery phase will not participate in the real-time frequency control phase in this dispatch cycle.

[0162] 2) For Mode 2, where conventional generating units can meet the frequency regulation requirements of the advanced stage but the margin is insufficient, the objective function only includes the frequency regulation cost of conventional generating units, i.e.:

[0163] (19)

[0164] 3) For Mode 3, conventional generating units are insufficient to support the advance-stage regulation demand, requiring the allocation of some energy storage to participate in the advance-stage. If the number of energy storage units participating in advance dispatch is determined by the principles in Section 1... p The optimization aims to minimize the combined cost of all resources participating in the advance stage scheduling. The objective function can be expressed as:

[0165] (20)

[0166] S32, taking into account CPS assessment to determine advance scheduling constraints.

[0167] During frequency control in this region, power balance must be maintained within the region at any given time. The sum of the frequency regulation power of generating equipment, the power deviation of tie lines, and the self-generated primary frequency regulation power of conventional units should be kept consistent with the unbalanced power, i.e., satisfying the following relationship:

[0168] (twenty one)

[0169] In the formula, K Gi Indicates the first i The primary frequency regulation unit adjustment coefficient for a conventional generating unit, Δ f t Δ P T,t for t Frequency deviation and tie-line power deviation at specific times; Δ P t This is due to regional power imbalance fluctuations caused by uncertainties in load and new energy sources.

[0170] 2) CPS constraints

[0171] Based on the North American Electric Reliability Council (NERC) assessment standards, hard constraints CPS1 and CPS2 are established to ensure that the optimization results meet the assessment quality requirements. CPS1 primarily measures the relationship between frequency deviation and regional control error, and its index value is expressed as follows:

[0172] (twenty two)

[0173] In the formula, B This is the frequency deviation coefficient for the control area; ɛ 1min The statistical value of the root mean square of the 1-minute frequency average deviation of the interconnected power grid in the previous year is taken. E ACE,t The ACE value is represented by the following formula:

[0174] (twenty three)

[0175] According to the judgment criteria, a CPS1 that is too small will fail to meet the assessment requirements, while a CPS1 that is too large poses a risk of difficulty in quickly correcting a sudden reversal in frequency deviation. Therefore, this invention adopts the minimum value. K CPS1_min 100%, maximum value K CPS1_max To achieve 400%, the CPS1 metric should meet the following requirements:

[0176] (twenty four)

[0177] Similarly, CPS2 demonstrates control over the amplitude of ACE, and the index value can be expressed as:

[0178] (25)

[0179] According to the definition of CPS2, the constraints satisfy:

[0180] (26)

[0181] In the formula, ɛ 15min Take the root mean square value of the frequency deviation over 15 minutes in the previous year for the interconnected power grid; B s This represents the frequency deviation coefficient for the entire interconnected power grid.

[0182] 3) Other constraints

[0183] From the power supply side, the regulation capacity boundaries of energy storage power stations and conventional generating units need to be considered. For conventional generating units, considering ramp-up constraints and power constraints, it can be expressed as:

[0184] (27)

[0185] (28)

[0186] In the formula, R RGi,t Indicates the first i A conventional unit t The rate of increase at any given moment.

[0187] Considering power and capacity constraints, an energy storage power station can be represented as follows:

[0188] (29)

[0189] (30)

[0190] In the formula, E bj,t Indicates the first j Energy storage power station t The capacity value at any given time, with the minimum and maximum capacity being 0.1 and 0.9 times its rated capacity, respectively.

[0191] From the network side perspective, in addition to meeting CPS assessment constraints, it is also necessary to meet relevant constraints such as regional power grid frequency deviation and tie-line power deviation, namely:

[0192] (31)

[0193] (32)

[0194] In the formula, Δ P Tmin Δ P Tmax Δ f min Δ f max These represent the minimum and maximum power deviation of the tie line, and the minimum and maximum frequency deviation, respectively.

[0195] Taking into account the aforementioned variable objective function and multiple constraints, the AGC advance stage optimization scheduling model adopted in this invention is formed. The nonlinear terms are linearized, transforming it into a linear programming model, which is then solved precisely using Yalmip+CPLEX.

[0196] S4. Construct a real-time scheduling optimization model, taking into account dynamic task coefficients, and complete the power allocation of remaining resources;

[0197] This embodiment specifically includes the following steps:

[0198] S41. Define the dynamic task coefficient for the charging and discharging phase at the current moment based on the current state of charge and the regional power grid's charging and discharging requirements. and ;

[0199] Based on the determination of some resource output in the advance scheduling phase, the real-time scheduling phase collects information such as frequency deviation and tie-line power deviation with a granularity of 4 seconds, calculates the ACE signal, and generates the AGC general command by the controller. The core task of the real-time scheduling phase is to allocate the general AGC command to each energy storage power station according to the technical and economic characteristics and operating status of the energy storage power station, and then each power station executes it. As can be seen from the modeling of the advance scheduling phase, the dynamic cost generated by the energy storage power station participating in AGC frequency regulation is shown in Equation (15). On this basis, in order to reflect the comprehensive impact of the real-time operating status and rated power of energy storage, this section introduces a dynamic task coefficient to constrain its maximum adjustable power, thereby coupling the technical and economic characteristics of energy storage to complete the dynamic optimal allocation.

[0200] Based on the current state of charge and the regional power grid's charging and discharging needs, define the dynamic task coefficient for the charging and discharging phase at the current moment. and When the state of charge (SOC) is relatively high, the dynamic task coefficient equals 1; as the SOC approaches its limit, the dynamic task coefficient is gradually reduced to zero. The functional relationship between the dynamic task coefficient and SOC conforms to an sigmoid function, represented by the following hyperbolic tangent function:

[0201] (33)

[0202] (34)

[0203] (35)

[0204] In the formula, SOC min , SOC max The minimum and maximum values ​​represent the state of charge; in this invention, these values ​​are 0.1 and 0.9. α jc,t , α jd,t These represent the state-of-charge deviation coefficients during the charging and discharging stages, respectively. w This is an adaptive factor, set manually and rounded. w If the value is too large, the curve will converge to 0 or the nominal value too quickly; if the value is too small, the curve will lose continuity due to slow convergence. Therefore, to ensure the rationality of the dynamic task coefficient setting, this invention compares and verifies the curves under different adaptive factors, and finally determines... w The value is set to 3.

[0205] S42. Define the maximum adjustable power during the charging and discharging phase of an energy storage power station. and The maximum adjustable power is adjusted based on the dynamic task coefficient change curve with SOC during the charging and discharging phase to meet the charging and discharging requirements.

[0206] Figure 2 The curves showing the dynamic task coefficient changing with SOC are further presented, divided into charging and discharging curves. As can be seen from the curves, with... SOC ref Based on this, when in a charging state, if the state of charge is less than... SOC ref This indicates a high charging demand at this time; maintain charging at the highest dynamic task coefficient, when it is greater than... SOC ref During charging and discharging, the dynamic task coefficient is gradually reduced to zero at a certain rate, and the process is similar during discharge. Furthermore, the maximum adjustable power during the charging and discharging phases of the energy storage power station is defined. and :

[0207] (36)

[0208] S43. Considering minimizing energy storage costs, construct the objective function for the real-time scheduling stage.

[0209] Objective function of real-time scheduling process F real That is, minimizing the sum of the cost functions of all energy storage power stations participating in real-time scheduling:

[0210] (37)

[0211] S44. A real-time scheduling model is constructed taking into account constraints such as maximum adjustable power, and the total AGC instructions for the real-time stage are allocated to the remaining energy storage units based on the model scheduling results.

[0212] The constraints include satisfying the total command balance and the maximum adjustable power and capacity constraints of the energy storage power station, which can be expressed as:

[0213] (38)

[0214] In the formula, P AGC This indicates the overall command for real-time stage scheduling.

[0215] The purpose of introducing dynamic task coefficients is to adaptively change the adjustable power of the energy storage power station according to the SOC situation, thereby reasonably adjusting the model constraints. Overall, real-time scheduling aims to minimize costs, taking into account the economic and technical characteristics of energy storage in the command allocation process. It can ensure that power stations with good economic efficiency and strong power output capabilities are given priority in output, while avoiding frequent calls to power stations with poor operating conditions that could lead to their premature loss of regulation capabilities.

[0216] S5. Construct a two-stage scheduling frequency response model and verify the effectiveness of the strategy through simulation.

[0217] To verify the effectiveness of the proposed two-stage coordination strategy in the AGC frequency regulation process of an energy storage collaborative participation system, a frequency response model under the proposed strategy was built using MATLAB / Simulink software, as follows: Figure 3 As shown.

[0218] The basic framework of the model adopts a simplified system frequency response (SFR) model, in which the frequency characteristics of each power generation device are represented by the transfer function. The closed-loop simulation of the overall AGC process is used. Existing technologies have demonstrated the feasibility and advantages of using the SFR model to replace the full-device model, which will not be elaborated upon here. Based on this, the model integrates the two-stage scheduling framework of this invention, embedding the two-stage optimization solution code into the aforementioned basic framework of the frequency response model. On one hand, in the advance scheduling stage, based on relevant prediction information and mode determination results, the power of all conventional units and some energy storage power stations is optimized and solved, outputting the output Δ of the advance scheduling portion. P pro On the other hand, the ACE value is calculated in TBC mode and then the AGC command for real-time scheduling is generated by the PI controller to optimize the solution allocated to another part of the energy storage power station.

[0219] To fully demonstrate the superiority of the strategy proposed in this invention, four strategies are compared, all configured with the same resource type and parameters. The specific scheduling strategies employed by each strategy are as follows:

[0220] 1) Strategy 1 only adopts advance scheduling, and uses the advance scheduling modeling method of this invention to perform advance scheduling on all power generation equipment.

[0221] 2) Strategy 2 only adopts real-time scheduling, constructs an SFR model with the same parameters as the present invention, generates AGC instructions based on power deficit, and distributes the instructions to each power generation device according to the principle of minimizing joint cost.

[0222] 3) Strategy 3 adopts a two-stage scheduling method, constructs an SFR model with the same parameters as the present invention, and integrates advance scheduling and real-time scheduling, but resources are fixedly divided, with only conventional units participating in advance scheduling and only energy storage power stations participating in real-time scheduling.

[0223] 4) Strategy 4 adopts the two-stage dynamic coordination and scheduling method based on mode adaptive switching described in this invention.

[0224] To verify the effectiveness of the strategy proposed in this invention, this section analyzes the dynamic response process of various adjustment resources under the scheduling strategy of this invention, using the scheduling cycle as the unit. Figure 5 The simulation shows the unbalanced power disturbance curves over a given period, illustrating the power variations caused by uncertainties in renewable energy sources and loads. Specifically, the unbalanced power curves include minute-level disturbance components with large, slow changes, and second-level disturbance components with small, rapid changes.

[0225] The curve fluctuations show that the unbalanced power fluctuates within a range of approximately ±120MW, with small, second-level disturbances fluctuating around ±10MW. Overall, the power disturbance exhibits a trend of first rising and then falling. During the four scheduling cycles in the simulation period, scheduling cycles 1 and 3 are primarily the rising and falling phases, during which the system's unbalanced power experiences sharp increases and decreases, resulting in relatively high regulation demands. Scheduling cycles 2 and 4 are the flattening phases, during which the unbalanced power fluctuations are weaker, and the system regulation demands are lower, facilitating the recovery of the energy storage power station's State of Charge (SOC).

[0226] Figure 5 , 6 This section showcases the SOC fluctuations of energy storage power stations during various dispatch cycles and the advance dispatch frequency regulation power commands for regulating resources. Combined with... Figure 5 , 6 The dynamic coordination process of various adjustment resources under the strategy of this invention is analyzed as follows:

[0227] 1) Scheduling cycle 1 falls within the unbalanced power rise period. Based on the determined mode 2, only conventional units participate in the advanced stage frequency regulation without energy storage SOC recovery. Figure 6 As shown in (a), the three generating units can effectively fill the power imbalance. There are small regulation deviations at the 5th and 12th minutes. This is because the dispatching model does not demand perfect power balance and allows for frequency deviations within the specified constraints, thereby reducing regulation costs.

[0228] 2) Scheduling cycle 2 falls within a period of stable unbalanced power. Based on the identified mode 1, conventional generating units participate in frequency regulation while simultaneously assisting energy storage in restoring State of Charge (SOC). During scheduling cycle 1, all energy storage devices participate in real-time phased frequency regulation, combined with… Figure 5As can be seen, energy storage units 5 and 6 have reached the advanced SOC recovery threshold and participate in the advanced SOC recovery process. Figure 6(b) shows that the power output direction of the energy storage power station is opposite to the unbalanced power direction. Conventional units provide power higher than the actual demand to achieve positive adjustment of the energy storage SOC. Energy storage is equivalent to a portion of the unbalanced power during the dispatching process. The dashed box shows the unbalanced power value after considering energy storage, which basically coincides with the output of conventional units.

[0229] 3) Scheduling cycle 3 falls within the unbalanced power decline period, with power fluctuations exceeding those of scheduling cycle 1. This corresponds to mode 3, where energy storage devices assist conventional units in completing the advanced stage of frequency regulation. Based on the aforementioned control...

[0230] Based on the principles described in Section 1, energy storage power stations are added sequentially until frequency regulation requirements are met. Finally, it was determined that energy storage 1 and energy storage 2 will participate in the advanced stage regulation, by... Figure 6 As shown in (c), the auxiliary power output of energy storage is mainly concentrated in the period of more intense power fluctuations from the 4th to the 11th minute.

[0231] 4) Scheduling cycle 4 is the second stable period of unbalanced power, also belonging to mode 1, for energy storage SOC recovery. Combined with... Figure 5 It can be seen that energy storage 2 and energy storage 6 reached the deviation threshold of SOC during the participation in advanced and real-time frequency regulation, respectively, and therefore participated in advanced SOC recovery. From Figure 6 As shown in (d), unlike scheduling cycle 2, the direction of energy storage SOC recovery power in this scheduling cycle is consistent with the direction of unbalanced power. Conventional units reduce their output to ensure power balance. It can be seen that the proactive recovery strategy in scheduling cycle 4 not only improves energy storage SOC but also reduces the power output of conventional units, thereby reducing regulation losses.

[0232] Table 1. Determination of Different Scheduling Cycle Modes and Resource Selection

[0233]

[0234] Table 1 summarizes the mode determination results and resource selection for different scheduling cycles. Through the above analysis and summary of the scheduling process, it is verified that the strategy of this invention can dynamically switch operating modes according to actual adjustment needs, and rationally divide and allocate adjustment resources, achieving a balance between system frequency regulation and energy storage state of charge recovery.

[0235] Based on the effective verification of the resource response characteristics of the strategy of this invention, this section compares multiple strategies from the perspectives of frequency deviation, response error rate, and energy storage SOC status to demonstrate the advantages of the dynamic coordination strategy of this invention in improving the frequency regulation capability of the AGC system and the recovery effect of energy storage SOC. Among them, strategies 1 and 2 both use only a single scheduling method to compare and verify the necessity of advance-real-time coordinated scheduling; strategy 3 adopts two-stage scheduling, but the resources are only divided in a fixed manner, to compare and verify the importance and advantages of mode switching and adaptive resource allocation in the strategy of this invention.

[0236] Figure 7 The simulation demonstrates the frequency deviations of various strategies under unbalanced power disturbances during the simulation period. It is evident that the strategy of this invention outperforms the others in terms of frequency deviation, maintaining the amplitude consistently within 0.08Hz, effectively ensuring the safe and stable operation of the interconnected power grid and reducing the risk of frequency exceedances. Strategy 1 exhibits the worst frequency maintenance performance among the four strategies, with the highest frequency deviation exceeding 0.3Hz, easily triggering frequency collapse. The poor performance of Strategy 1 in frequency maintenance is primarily attributed to the insufficient precision of the dispatch commands. All generating equipment participates in the advanced dispatch phase, with command issuance precision only at the minute level, making it difficult to mitigate rapidly changing power disturbances at the second level. Furthermore, ultra-short-term prediction errors further exacerbate the frequency exceedance risk during the Strategy 1 dispatch process.

[0237] Strategy 2 exhibits better frequency deviation performance than Strategy 1 but worse performance than Strategy 3 and 4. Its scheduling process involves real-time scheduling with the participation of all regulation resources. Although the scheme of formulating scheduling instructions based on real-time regulation needs avoids the problem of insufficient instruction precision, conventional units have poor flexibility and slow response speed, making it impossible to accurately track scheduling instructions. Figure 8 The presentation illustrates the response error rates of conventional generating units and energy storage devices under different strategies. It shows that strategies 1, 3, and 4 all maintain response error rates below 2%, indicating that each generating unit can execute commands issued by the upper-level dispatch center with relatively high accuracy. Strategy 2 exhibits a significantly higher response error rate for generating units compared to the other strategies, particularly for conventional generating units, where the response error rate reaches 7.2%, more than 3.5 times higher than the other strategies. This substantial response deviation in Strategy 2 prevents generating units from mitigating power deficits according to dispatch commands, leading to more severe frequency issues.

[0238] Strategy 3 employs a two-stage complementary strategy to compensate for the shortcomings of strategies 1 and 2, and its frequency deviation is not significantly different from that of strategy 4 for most periods. However, since strategy 3 only performs fixed resource allocation, conventional units cannot support large regulation demands during scheduling cycle 3. Figure 9The diagram illustrates the advance scheduling performance of Strategy 3 during this cycle. It shows significant power deviations occurring between minutes 5 and 11, with the maximum deviation reaching 24.36 MW at minute 7. These large power deviations during the advance phase exacerbate the frequency regulation pressure on the generating equipment during the real-time phase, resulting in even larger frequency deviations.

[0239] Table 2 Comparison of System-Level Data for Different Strategies

[0240]

[0241] Based on the above analysis, Table 2 summarizes the relevant data comparison of system-level indicators for each strategy. Among them, in terms of frequency deviation, the strategy of this invention has the best mean and maximum amplitude, improving by 7.19% and 16.90% respectively compared to the fixed-division strategy (Strategy 3). In terms of CPS, the strategy of this invention achieves a mean CPS1 of 146.3%, the highest among the four strategies. Regarding response deviation rate, the deviation rate of Strategy 2 is significantly higher than the other strategies. The deviation rates of the other three strategies are all controlled below 2%, with the strategy of this invention having the lowest deviation rate at only 1.44%. Through multi-indicator comparison, it is verified that the strategy of this invention can effectively improve AGC frequency tuning capability and enhance assessment quality.

[0242] Based on the comparison of system-level indicators Figure 10 The changes in the overall State of Charge (SOC) of the energy storage cluster before and after optimization using different strategies were compared and analyzed. To measure the overall state of charge (SOC) level of the energy storage cluster, the ratio of the remaining capacity to the total capacity of the energy storage cluster was used as its equivalent SOC. This invention uses the difference between the equivalent SOC and the benchmark value SOCref, dSOC, to measure the energy storage SOC level. The smaller the difference, the closer it is to the benchmark value, and the better the overall operating status of the energy storage. In addition, this invention also uses the maximum range δ of SOC and the balance degree Dsoc to further comprehensively measure the overall SOC quality of the energy storage cluster, as shown in equations (39) and (40), respectively.

[0243] (39)

[0244] (40)

[0245] In the formula, {SOC j } max 、{SOC j } min This represents the maximum and minimum SOC values ​​of the energy storage power stations in the energy storage cluster at the end of the optimization process; SOC avg This represents the average SOC of each energy storage power station at the end of the optimization process.

[0246] Depend on Figure 10 As shown in (a)-(c), the strategy of this invention has a significant improvement effect on all SOC indicators. Specifically, the equivalent SOC increases from 0.4653 to 0.5015, further narrowing the deviation from the benchmark SOC value. The maximum range decreases from 0.3 to 0.1708, and the uniformity decreases from 0.0944 to 0.032, indicating that the SOC difference between energy storage power stations is reduced, and the SOC is becoming more uniform. In comparison, strategies 1 and 2 show no improvement or only a slight improvement in various SOC indicators. Strategy 3 shows a more serious deterioration in various SOC indicators. The main reason is that, in response to the frequent power fluctuations in the real-time stage, strategy 3 only provides long-term support for energy storage resources and does not perform advanced SOC recovery for energy storage like strategy 4. Therefore, the overall SOC quality declines during the simulation.

[0247] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0248] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0249] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0250] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

[0251] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A two-stage AGC scheduling method considering adaptive mode switching, characterized in that, Includes the following steps: S1. Construct a two-stage basic scheduling architecture consisting of a proactive scheduling stage and a real-time scheduling stage; S2. Develop the operating mode of the AGC system based on resource regulation capacity and power grid regulation requirements, and perform adaptive switching between modes; S3. Based on the operation mode determination results, construct an advanced scheduling optimization model to pre-determine the output of some resources; S4. Construct a real-time scheduling optimization model, taking into account dynamic task coefficients, and complete the power allocation of remaining resources; S5. Construct a two-stage scheduling frequency response model and verify the effectiveness of the strategy through simulation.

2. The two-stage AGC scheduling method considering adaptive mode switching according to claim 1, characterized in that, In S1: In the advance scheduling phase, with a time granularity of 1 minute, unbalanced power is obtained based on ultra-short-term prediction, and an optimization model for power balance, resource regulation capability, and CPS assessment constraints is constructed. The real-time scheduling phase uses a time granularity of 4 seconds. Based on the advance scheduling of fixed part of the resource output, frequency deviation and tie line deviation are collected to calculate ACE and construct a real-time power allocation model.

3. The two-stage AGC scheduling method considering adaptive mode switching according to claim 1, characterized in that, S2 specifically includes the following steps: S21. Quantitative upper and lower limits of conventional unit regulation capacity P RGU,t , P RGL,t : In the formula, P RGi,t-1 , R rateG,i , P Gi,max , P Gi,min They represent the first i The power value, maximum grade rate, maximum and minimum power of a conventional unit at the previous moment; m Indicates the total number of conventional generating units; The scheduling time interval; S22. Introducing 0-1 decision variables T 1. Determine whether conventional generating units can support grid regulation requirements: when hour T 1=1, otherwise T 1=0, when T When 1=0, priority is given to allocating energy storage to participate in advanced dispatch until... T 1=1; Introducing 0-1 decision variables T 2. Determine the adjustment margin: When hour T 2 = 1; where, To adjust the margin coefficient, To adjust the deviation, Δ P t This refers to regional power imbalance fluctuations. S23, the mode switching command satisfies: In the formula, This is a mode switching command; S24. Determine the operating mode based on the mode switching instruction, including: when Mode 1: Conventional generating units participate in advanced dispatching, while the lowest state of charge energy storage power station is restored to SOC, and the remaining energy storage participates in real-time phased dispatching. when Mode 2: Conventional generating units are scheduled in advance, and energy storage power stations participate in real-time scheduling to complete a two-stage scheduling process; when Mode 3: Some energy storage assists conventional units in advance scheduling, while the remaining energy storage participates in real-time phase scheduling.

4. The two-stage AGC scheduling method considering adaptive mode switching according to claim 3, characterized in that, S3 specifically includes the following steps: S31. Introduce various resource cost functions and define variable objective functions under different operating modes; S32, taking into account CPS assessment to determine advance scheduling constraints.

5. The two-stage AGC scheduling method considering adaptive mode switching according to claim 4, characterized in that, The variable objective function in S31 under different operating modes is expressed as follows: Mode 1: Conventional generating units are sufficient to meet regulation demands and have ample margin. The objective is for conventional generating units to participate in power smoothing during the advanced phase, while simultaneously assisting some energy storage power stations with the lowest state of charge (SOC) to restore their State of Charge (SOC). The objective function is... F Represented as: In the formula, M The penalty factor for energy storage advance recovery is set to a constant, indicating that it has a higher priority; q The number of energy storage units participating in the advanced phase of SOC recovery; T The scheduling period is set to 15 minutes. c f,i Indicates the first i The auxiliary frequency regulation cost coefficient for each thermal power unit For the first j Energy storage t Timing adjustment deviation, P RGi,t Indicates the first i The power value of a conventional unit at the current moment. m Indicates the total number of conventional generating units; Mode 2: When conventional units can meet the frequency regulation requirements during the advanced stage but the margin is insufficient, the objective function is... F Including only the frequency regulation costs of conventional units, it is expressed as: Mode 3: If conventional generating units are insufficient to support the demand during the advanced adjustment phase, some energy storage needs to be allocated to participate in the advanced phase. The determined number of energy storage units participating in the advanced dispatch is... p The optimization aims to minimize the combined cost of all resources participating in the advance stage scheduling, with the objective function being... F Represented as: In the formula, This indicates that the unit is in t Frequency tuning costs at any given time F Rbj,t It equals the sum of dynamic capacity cost, dynamic mileage cost, and lifetime depreciation cost.

6. The two-stage AGC scheduling method considering adaptive mode switching according to claim 5, characterized in that, The constraints in S32 include: Power balance constraints are expressed as: In the formula, K Gi Indicates the first i The primary frequency regulation unit adjustment coefficient for a conventional generating unit, Δ f t Δ P T,t for t Frequency deviation and tie-line power deviation at specific times; Δ P t This is due to regional power imbalance fluctuations caused by uncertainties in load and new energy sources; R RGi,t Indicates the first i A conventional unit t The rate of increase at any given moment; P Rbj,t for t Time of the first j The output power of an energy storage power station; CPS constraints are represented as follows: In the formula, and These are the hard constraint indicators based on the NERC assessment standard, and: In the formula, B This is the frequency deviation coefficient for the control area; ɛ 1min The statistical value of the root mean square of the 1-minute frequency average deviation of the interconnected power grid in the previous year is taken. E ACE,t Indicates the ACE value. ɛ 15min Take the root mean square value of the frequency deviation over 15 minutes in the previous year for the interconnected power grid; B s This refers to the frequency deviation coefficient of the entire interconnected power grid; For the scheduling period, They are respectively The maximum and minimum values; Conventional generating units, considering ramp-up constraints and power constraints, are expressed as follows: In the formula, Let t be the output of the conventional unit. P Gi,max , P Gi,min These are the maximum and minimum ramp power for conventional generating units, respectively. Maximum gradeability; Considering power and capacity constraints, an energy storage power station can be represented as follows: In the formula, Indicates the first j Energy storage power station t Capacity value at any given time These are the maximum and minimum values ​​of the capacity, respectively. Rated power; Power deviation constraint, expressed as: In the formula, This represents the communication power deviation at time t. For the regional power grid frequency deviation, Δ P Tmin Δ P Tmaxx These represent the minimum and maximum power deviations of the tie line, Δ. f min Δ f max These represent the minimum and maximum frequency deviations of the regional power grid, respectively.

7. The two-stage AGC scheduling method considering adaptive mode switching according to claim 1, characterized in that, S4 specifically includes the following steps: S41. Define the dynamic task coefficient for the charging and discharging phase at the current moment based on the current state of charge and the regional power grid's charging and discharging requirements. and The specific calculation method is as follows: In the formula, and The following are the dynamic task coefficients for the charging and discharging phases, in order. SOC min , SOC max Represents the minimum and maximum values ​​of the state of charge. α jc,t , α jd,t These represent the state-of-charge deviation coefficients during the charging and discharging phases, respectively. w As an adaptive factor, SOC j,t Indicates the first j Energy storage t State of charge at time t, SOC ref This is the reference value for the state of charge; S42. Define the maximum adjustable power during the charging and discharging phase of an energy storage power station. and The maximum adjustable power is adjusted based on the dynamic task coefficient change curve with SOC during the charging and discharging phase to meet the charging and discharging requirements. S43. Considering minimizing energy storage costs, construct the objective function for the real-time scheduling phase; S44. A real-time scheduling model is constructed taking into account constraints such as maximum adjustable power, and the total AGC instructions for the real-time stage are allocated to the remaining energy storage units based on the model scheduling results.

8. The two-stage AGC scheduling method considering adaptive mode switching according to claim 7, characterized in that, The maximum adjustable power during the charging and discharging phase of the energy storage power station in S42 and The specific calculation method is as follows: In the formula, and The following are the maximum adjustable power outputs during the charging and discharging phases of the energy storage power station. and The following are the dynamic task coefficients for the charging and discharging phases, in order. This is the rated power.

9. The two-stage AGC scheduling method considering adaptive mode switching according to claim 7, characterized in that, The objective function in S43 is expressed as follows: In the formula, The objective function of the real-time scheduling process is... For the first j A function for the cost of AGC (Automatic Guided Service) auxiliary services for an energy storage power station. This represents the total number of energy storage power stations. The number of energy storage units participating in advanced dispatch.

10. The two-stage AGC scheduling method considering adaptive mode switching according to claim 7, characterized in that, The real-time scheduling constraints in S44 are expressed as follows: In the formula, This is the general instruction for real-time stage scheduling. for t Time of the first j The output power of an energy storage power station This represents the total number of energy storage power stations. The number of energy storage units participating in advanced dispatching. These represent the maximum adjustable power during the charging and discharging phases, for t Time of the first j The capacity value of an energy storage unit. They represent the first j The maximum and minimum capacity of each energy storage power station.

Images