Following-networking type new energy sending end system scheduling method considering dynamic frequency constraint

Abstract

The invention discloses a tracking-networking type new energy sending end system scheduling method considering dynamic frequency constraint, and belongs to the technical field of power system operation and planning. The method comprises the following steps: establishing a multi-resource collaborative frequency response model of a high-proportion new energy sending end system, establishing a maximum frequency deviation, an initial frequency change rate and a steady-state frequency difference dynamic index, and converting the dynamic indexes into constraint conditions in optimal scheduling; and combining an energy storage refinement model, conventional unit operation constraint and fuzzy opportunity constraint to process wind and light load uncertainty, and constructing an optimization framework with the lowest total operation cost as a target function. Through deep fusion of dynamic frequency security constraint and resource collaborative scheduling, new energy output fluctuation is effectively stabilized, the system frequency stability and economy are improved, and the technical defect that a traditional scheduling model neglects a dynamic frequency security index in a high-proportion new energy scene is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system operation and planning, and relates to a dispatching method for a grid-type new energy sending-end system taking dynamic frequency constraints into consideration. Background Art

[0002] The integration of large-scale renewable energy into the grid has become a prominent feature of the global power system transformation. However, the low inertia characteristics and output volatility of renewable energy units such as wind power and photovoltaics have led to severe frequency stability issues in the sending-end system. Traditional power systems rely on the rotating inertia of synchronous units and the response of speed regulators to maintain frequency security. Their dispatch models are relatively simple, considering only constraints such as power balance. However, with the increase in the penetration rate of renewable energy, the system's equivalent inertia has significantly decreased, and system frequency issues have become more prominent. Existing dispatch methods ignore dynamic frequency indicator constraints such as maximum frequency deviation, initial frequency change rate, and steady-state frequency difference, resulting in underestimated safety risks. The frequent frequency over-limit accidents in high-proportion renewable energy systems have exposed this limitation.

[0003] Technically, new energy generators lack the inertial response and primary frequency regulation capabilities of traditional synchronous generators. While inertial support can be simulated through virtual synchronous generator technology or energy storage systems, existing optimization dispatch models rarely quantitatively analyze the dynamic response characteristics of virtual inertia and frequency regulation resources. Furthermore, the high uncertainty of new energy output and the spatiotemporal randomness of fault scenarios further exacerbate the conflict between frequency stability and dispatch economy. Current research often employs multi-scenario stochastic programming or robust optimization methods to address uncertainty. However, these frequency security constraints are still based on static thresholds and fail to couple the dynamic frequency response trajectory with uncertain scenarios. This results in dispatch schemes that are either overly conservative (at the expense of economy) or pose safety risks. In current power systems, new energy stations and energy storage devices are connected to the main grid through either grid-following or grid-forming control. On the timescales of primary and secondary frequency regulation, the differences between the control mechanisms of grid-following and grid-forming no longer dominate the frequency regulation capabilities of these devices; instead, the physical constraints of the energy source dominate.

[0004] In the existing technology, the dispatching strategies for renewable energy power systems mainly focus on power balance and economic optimization. For example, the impact of wind and solar forecast errors is reduced through multi-scenario generation technology and fuzzy opportunity constraints, or conventional frequency regulation reserve constraints are introduced in the unit combination. However, such methods do not start from the dynamic system characteristics and embed indicators such as maximum frequency deviation, initial frequency change rate, and steady-state frequency difference into the optimization framework. For example, some studies have increased system inertia by configuring energy storage, but the coordination of their control strategies and dispatching models is still limited to static reserve power allocation, and they have failed to achieve global optimization of dynamic frequency security dispatching. Therefore, how to construct a dispatching method that takes into account dynamic frequency security and economy in scenarios with a high proportion of renewable energy has become a core issue that needs to be solved in the field of power system operation. Summary of the Invention

[0005] The present invention proposes a dispatching method for a grid-type new energy sending-end system considering dynamic frequency constraints. By constructing a multi-resource collaborative frequency response model of a high-proportion new energy sending-end system, the dynamic frequency regulation characteristics of wind power, photovoltaic, energy storage and thermal power units are quantified, and a dynamic frequency index covering the maximum frequency deviation, initial frequency change rate and steady-state frequency difference is established, and it is converted into linear and nonlinear constraints for optimal scheduling; combined with a refined energy storage model (including charge state, charge and discharge times and cycle energy targets), thermal power unit operation constraints (climbing restrictions, minimum on / off time and spare capacity) and a wind-solar-load uncertainty processing method based on fuzzy chance constraints, an optimization framework with the lowest total system operating cost as the objective function is constructed, realizing the deep integration of dynamic frequency security constraints and economic dispatch, and smoothing the fluctuation of new energy output through multi-resource collaborative dispatching, to ensure the system frequency stability and operation economy in high-proportion new energy scenarios.

[0006] The technical solution adopted by the present invention is as follows: a scheduling method for a grid-type new energy sending-end system considering dynamic frequency constraints, the method comprising the following steps:

[0007] Step S1: Construct an equivalent frequency response model for the coordination of multiple resources in a high-proportion new energy sending-end system, such as wind power, photovoltaic power, energy storage, and conventional units;

[0008] Step S2: Based on the equivalent frequency response model, dynamic frequency indicators are established, including maximum frequency deviation, initial frequency change rate, and steady-state frequency difference;

[0009] Step S3: Based on the constructed dynamic frequency index, considering the system range and dead zone settings of each index, dynamic frequency constraints of maximum frequency deviation, initial frequency change rate, and steady-state frequency difference are established;

[0010] Step S4: Constructing a refined energy storage model, specifically including: energy storage charge state, capacity limit, charge and discharge power limit, energy storage charge and discharge number limit, and setting the energy storage internal energy to the target value after a scheduling cycle.

[0011] Step S5: Construct a conventional unit model. Taking a thermal power unit as an example, the model includes: thermal power unit ramp limit, minimum on / off time limit, output power limit, and standby capacity limit;

[0012] Step S6: Constructing a wind, solar, and load uncertainty processing model. Based on the randomness of wind, solar, and load, fuzzy chance constraints are used to process the uncertainty of wind, solar, and load to a certain extent.

[0013] Step S7: Construct an objective function with the lowest total operating cost of the renewable energy sending-end system, taking into account dynamic frequency constraints, smoothing the fluctuation of renewable energy output and improving system safety and stability.

[0014] Further preferably, in step S1, a conventional thermal power unit frequency response model, an energy storage device frequency response model, a dynamic frequency response model of the system load, a maximum error model of the expected power disturbance step value for the load and new energy prediction are constructed, and a dynamic frequency response model of the system is constructed. An equivalent method is used to obtain a simplified thermal power unit transfer function model, and on this basis, an equivalent frequency response model is finally constructed.

[0015] Further preferably, in step S2, a calculation method for three frequency indices, namely, initial frequency change rate, steady-state frequency difference, and maximum frequency deviation, is derived according to the equivalent frequency response model.

[0016] Further preferably, in step S3, based on the maximum frequency deviation, initial frequency change rate, steady-state frequency difference dynamic frequency index constructed in step S2, dynamic frequency constraint conditions are established taking into account the range allowed by the system.

[0017] Further preferably, the energy storage refined model includes an energy storage state of charge model, energy storage capacity constraint, energy storage charge and discharge power model, energy storage energy management model, and energy storage charge and discharge ramp model.

[0018] Further preferably, the conventional unit model includes thermal power unit ramping constraints, thermal power unit minimum on / off time constraints, thermal power unit output power constraints, and thermal power unit spare capacity constraints.

[0019] Further preferably, in step S6, the uncertainty of the forecast error of wind power, photovoltaic output and load in the high-proportion renewable energy sending-end power system is processed, and the processed model is as follows:

[0020] When using fuzzy chance constraints to clear equivalence classes and perform equivalent calculations, various membership functions need to be used to represent fuzzy parameters. Taking the trapezoidal membership function as an example, the prediction errors of wind power, photovoltaic power, and load are considered using a unified fuzzy parameter P. F To express;

[0021] Fuzzy chance constraints are used to describe system uncertainty and construct a system reserve capacity model;

[0022] The system should reserve some spinning reserve capacity to cope with the uncertainty of renewable energy output and load output; the size of the reserved spinning reserve capacity is determined by the size of the renewable energy output error and the load error; a triangular fuzzy membership function is used to describe the uncertainty of the high-proportion renewable energy sending-end system;

[0023] The system spinning reserve capacity model is transformed into a form with clear equivalence classes that can be solved.

[0024] Further preferably, in step S7, the following model is established with the minimum total operating cost of the new energy sending-end power system as the objective function.

[0025] Further preferably, the large-M method is used to linearize the primary frequency regulation constraint model of the conventional thermal power unit and the primary frequency regulation constraint model of the energy storage.

[0026] The present invention accurately quantifies the system frequency safety boundary through dynamic frequency indicator modeling and constraint embedding, avoiding the problem of underestimation of safety risks in traditional static threshold methods; combines the equivalent frequency regulation characteristic modeling of thermal power and energy storage to achieve coordinated optimization of virtual inertia, droop control and physical inertia, and improves the dynamic response capability of the system; adopts a refined energy storage model to reduce equipment loss, and uses fuzzy opportunity constraints to clearly balance the impact of wind and solar load uncertainty on economy and robustness, thereby reducing the total operating cost of the system while ensuring frequency safety; at the same time, it is compatible with the differentiated parameter settings of grid-following and grid-forming energy storage, providing an optimized scheduling solution that takes into account dynamic safety and economy for high-proportion new energy sending-end systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 Schematic diagram of the frequency response model of the power system at the sending end with a high proportion of renewable energy.

[0028] Figure 2 Schematic diagram of the optimization scheduling process of the grid-type new energy sending-end system considering dynamic frequency constraints. DETAILED DESCRIPTION

[0029] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0030] See also Figure 1 、 Figure 2 The present invention addresses the issues of high-proportion renewable energy sending-end systems with grid-connected energy storage, such as renewable energy fluctuations and poor anti-interference capabilities of high-proportion renewable energy sending-end systems. A scheduling method for grid-connected renewable energy sending-end systems considering dynamic frequency constraints is proposed. The method comprises the following steps:

[0031] Step S1: Construct an equivalent frequency response model for the coordination of multiple resources in a high-proportion new energy sending-end system, such as wind power, photovoltaic power, energy storage, and conventional units;

[0032] Step S2: Based on the equivalent frequency response model, dynamic frequency indicators are established, including maximum frequency deviation, initial frequency change rate, and steady-state frequency difference;

[0033] Step S3: Based on the constructed dynamic frequency index, considering the system range and dead zone settings of each index, dynamic frequency constraints of maximum frequency deviation, initial frequency change rate, and steady-state frequency difference are established;

[0034] Step S4: Constructing a refined energy storage model, specifically including: energy storage charge state, capacity limit, charge and discharge power limit, energy storage charge and discharge number limit, and setting the energy storage internal energy to the target value after a scheduling cycle.

[0035] Step S5: Construct a conventional unit model. Taking a thermal power unit as an example, the model includes: thermal power unit ramp limit, minimum on / off time limit, output power limit, and standby capacity limit;

[0036] Step S6: Constructing a wind, solar, and load uncertainty processing model. Based on the randomness of wind, solar, and load, fuzzy chance constraints are used to process the uncertainty of wind, solar, and load to a certain extent.

[0037] Step S7: Construct an objective function with the lowest total operating cost of the renewable energy sending-end system, taking into account dynamic frequency constraints, smoothing the fluctuation of renewable energy output and improving system safety and stability.

[0038] Step S1: Construct a frequency response model for the coordinated multi-resource system with a high proportion of new energy, such as wind power, photovoltaic power, energy storage, and conventional units;

[0039] In a high-proportion renewable energy transmission system, a frequency response model that coordinates multiple resources is considered, such as the coordinated frequency regulation of thermal power units and energy storage. This model then constructs the frequency response models of thermal power units, energy storage, and the system. The model is as follows:

[0040] (1) The frequency response model of conventional thermal power units is as follows:

[0041]

[0042] Where, They are the proportional coefficient and integral coefficient of the dynamic frequency response of the speed control system of the thermal power unit; The per-unit value that represents the change in the generator's mechanical power over time; Δf * is the per-unit value of the change in system frequency over time; s is the independent variable in the complex frequency domain; μ is the regulation coefficient of the thermal power unit; F HP is the ratio of the mechanical power output of the high-pressure cylinder of the turbine to the total output power of the turbine; T RH is the reheater volume time constant.

[0043] (2) Frequency response model of energy storage device:

[0044] Typical energy storage primary frequency regulation control strategies include virtual inertia control and frequency droop control. Without considering the frequency regulation dead zone, its dynamic frequency response characteristic transfer function is as follows:

[0045]

[0046] Where, The change in the primary frequency modulation power of the energy storage; is the energy storage virtual inertia gain coefficient; is the energy storage frequency droop gain coefficient; by setting and Parameters are used to distinguish between grid-following and grid-building energy storage within the system.

[0047] (3) The dynamic frequency response model of the system load is shown as follows:

[0048]

[0049] Where, is the actual load fluctuation; is the load fluctuation at rated frequency; is the system frequency change under power disturbance; k Ld is the load frequency response coefficient; a j is the load proportion in the system that is proportional to the jth power of the frequency; N L is the maximum power value of the load in the system that is proportional to the frequency, that is, the maximum value of j; Z is an integer set.

[0050] (4) The expected power disturbance step value is the maximum error model of load and new energy prediction as follows:

[0051]

[0052] Where, ΔP Ld is the load disturbance at rated frequency; ΔP W is the wind power disturbance; ΔP PV is the photovoltaic output power disturbance; P Lde is the maximum error of load forecast; N W is the number of wind farms in the system, N V is the number of photovoltaic power stations in the system; P i We is the maximum error of the output prediction of the i-th wind farm; P i PVe is the maximum output prediction error of the i-th photovoltaic power station; P St,L is the sum of the step powers of the power disturbance.

[0053] (5) Equivalent frequency response model of high-proportion new energy sending-end system:

[0054] The dynamic frequency response model of the system is as follows:

[0055]

[0056] Where, is the system frequency variation under power disturbance; G I (s) is the system inertial response transfer function; H G (s), H E (s) are the group transfer functions of the speed regulators of thermal power units and energy storage, respectively; is the system load power at rated frequency. G I (s) is the superposition of the inertial response characteristics of all operational thermal power units in the system, as shown in the following formula:

[0057]

[0058] Where N G is the number of thermal power units in the system; ΔP Ge,i , ΔP Gm,i are the active power variation and mechanical power variation of the i-th thermal power unit respectively; is the start / stop status of the i-th thermal power unit, 1 for start and 0 for stop; is the rated output of the i-th thermal power unit; H i is the inertia time constant of the i-th thermal power unit; D i is the damping coefficient of the i-th thermal power unit.

[0059] H G (s), H E (s) are the superposition of the dynamic frequency response characteristics of all operational thermal power units and energy storage speed regulators in the system, as shown in the following formula:

[0060]

[0061] Where N E The sum of all grid-connected energy storage quantities in the system; is the rated value of the charging and discharging power of the i-th energy storage; μ i is the regulation coefficient of the i-th thermal power unit; is the virtual inertia gain coefficient of the i-th energy storage; is the frequency droop gain coefficient of the i-th energy storage; are respectively the proportional coefficient and integral coefficient of the dynamic frequency response of the speed control system of the i-th thermal power unit.

[0062] Since the parameters of each device in formula (3) are different, it is difficult to calculate the frequency index. Therefore, an equivalent method is used to simplify it. The simplified transfer function model of the thermal power unit is as follows:

[0063]

[0064] Where H equ is the transfer function of the thermal power unit, μ equ They are respectively the proportional coefficient, integral coefficient and differential coefficient of the equivalent speed regulator.

[0065] The equivalent frequency response model of the high-proportion new energy sending-end system can be obtained as follows:

[0066]

[0067] Where, κ L ,λ L , α L , β L are the transformation coefficients of the transfer function of the system dynamic frequency response model under power disturbance; k G is the system equivalent inertia coefficient; are the equivalent virtual inertia coefficient and frequency droop gain coefficient of the energy storage in the system respectively. G 、 The calculation methods are as follows:

[0068]

[0069] In step S2, based on the equivalent frequency response model, dynamic frequency indicators such as maximum frequency deviation, initial frequency change rate, and steady-state frequency difference are established:

[0070] According to formula (10), the calculation methods of the three frequency indicators, namely the initial frequency change rate, the steady-state frequency difference, and the maximum frequency difference, can be derived as follows:

[0071] 1. Steady-state frequency difference:

[0072]

[0073] Where, P St,L is the sum of the step powers of small power disturbances; is the per-unit value of the FM dead zone, is the per-unit value of the steady-state frequency difference, when P St,L >0, Take the negative sign; when P St,L <0, Take the positive sign.

[0074] 2. Initial frequency change rate:

[0075]

[0076] Where, is the per-unit value of the initial frequency change rate under power disturbance.

[0077] 3. Maximum frequency deviation:

[0078] Because β L With α L 2 The size relationship is uncertain, so it needs to be discussed case by case.

[0079] ① When β L >α L 2 season when The maximum frequency deviation is solved as follows:

[0080]

[0081] when When , the maximum frequency deviation is obtained as follows:

[0082]

[0083] Where, is the maximum frequency difference without considering the frequency modulation dead zone under power disturbance; τ os is the moment when the maximum frequency deviation occurs.

[0084] ② When β L <α L 2 season The maximum frequency deviation is solved as follows:

[0085]

[0086] ③ When β L =α L 2 When , the maximum frequency deviation is obtained as follows:

[0087]

[0088] For the above three situations, the maximum frequency deviation after taking into account the frequency modulation dead zone is shown as follows:

[0089]

[0090] When P St,L >0, Take the negative sign; when P St,L <0, Take the positive sign.

[0091] In step S3, based on the maximum frequency deviation, initial frequency change rate, and steady-state frequency difference dynamic frequency index constructed in step S2, and taking into account the range allowed by the system, the dynamic frequency constraint conditions are established as follows:

[0092] 1. Initial frequency change rate constraint:

[0093]

[0094] Where, is the initial frequency change rate limit under power disturbance, which is 0.004 pu. Combining the initial frequency change rate indicator calculation formula of step S2 and the equivalent frequency response model of step S1, equation (19) can be written as follows:

[0095]

[0096] Among them, P Lde,t is the maximum error of load forecast at time t, is the maximum output prediction error of the i-th wind farm at time t, is the maximum output prediction error of the i-th photovoltaic power station at time t, is the start / stop status of the i-th thermal power unit at time t;

[0097] 2. Steady-state frequency difference constraint:

[0098]

[0099] Where, is the steady-state frequency difference limit under small power disturbance, which is 0.01pu. Combining the steady-state frequency difference index calculation formula of step S2 and the equivalent frequency response model of step S1, the above formula can be written as follows:

[0100]

[0101] in, is the system load power at rated frequency at time t;

[0102] 3. Maximum frequency deviation constraint:

[0103]

[0104] Where, is the maximum frequency difference limit under small power disturbance, which is 0.01pu.

[0105] In step S4, a refined energy storage model is constructed, specifically including: energy storage charge state, capacity limit, charge and discharge power limit, energy storage charge and discharge number limit, and setting the energy storage internal energy to the target value after a scheduling cycle.

[0106] To describe the dynamic changes of energy storage during operation, including charging, discharging, and energy conversion processes. To ensure the normal operation and effectiveness of the energy storage system, it is necessary to ensure that the charge state is always non-negative. Therefore, the energy storage charge state model is as follows:

[0107]

[0108] Where, is the initial energy storage level of the i-th energy storage; Respectively represent the charging and discharging efficiency of the i-th energy storage, P i C (t′) is the charging power of the i-th energy storage at time t′, and the corresponding P i D (t′) is the discharge power of the i-th energy storage at time t′.

[0109] To ensure energy storage performance and extend its lifespan, a storage capacity constraint is used. Furthermore, this constraint uses the net charging efficiency of the energy storage to better account for its actual operating conditions, thereby approximately describing its dynamic characteristics. Therefore, the energy storage capacity constraint is as follows:

[0110]

[0111] Where, is the rated capacity of the i-th energy storage; is the net charging efficiency, which is in the range Inside.

[0112] Since the energy storage cannot be charged and discharged simultaneously during operation, and the energy storage charging and discharging power should be less than the maximum limit allowed by the configured converter, the energy storage charging and discharging power model is constructed as shown below:

[0113]

[0114] Where, is the rated value of the charging and discharging power of the i-th energy storage; z i (t) is a binary variable that measures the working state of energy storage i. When it is "1", it means that the energy storage is working in the charging state. Conversely, when it is "0", it means that the i-th energy storage is working in the discharging state. i (t) The principle of single binary variable value avoids the possibility of simultaneous charging and discharging; i C(t) is the charging power of the i-th energy storage at time t, P i D (t) is the discharge power of the i-th energy storage at time t.

[0115] To enable scheduling decision makers to customize the internal energy state of the energy storage after a scheduling cycle, an energy storage energy management model is constructed as follows:

[0116]

[0117] Where, In order to set the maximum and minimum limits for the final energy storage level, the energy storage level after the end of the scheduling cycle can be defined by setting the values of these two parameters to be equal and consistent with the target values desired by the scheduling decision maker; T is the scheduling cycle, and a one-day scheduling plan is formulated, which can be set to 24 hours, and Δt is the time interval of the scheduling cycle.

[0118] In order to carefully describe the charging and discharging process of energy storage, an energy storage charging and discharging ramp model is constructed as shown below:

[0119]

[0120] Where, The lower energy limit set for the i-th energy storage to prevent over-discharge; and They represent the positive / negative spinning reserve capacity of the i-th energy storage at time t, is the energy storage discharge efficiency, is the charging efficiency of energy storage i, is the rated power of the i-th energy storage, Soc i (t) is the state of charge of the i-th energy storage at time t.

[0121] In step S5, a conventional unit model is constructed. Taking a thermal power unit as an example, the model includes: thermal power unit ramping state, minimum on / off time limit, output power limit, and standby capacity limit. Therefore, the model of a conventional thermal power unit is as follows:

[0122] (1) Thermal power unit climbing constraints:

[0123]

[0124] Where, is the output power of the i-th thermal power unit at time t; It is a binary variable that measures whether the i-th thermal power unit is online at time t, where “1” indicates online and “0” indicates offline; is the ramp rate of the ith thermal power unit, in MW / 15min.

[0125] (2) Minimum on / off time constraints for thermal power units:

[0126]

[0127] Where, T on,i is the minimum startup time of the ith thermal power unit, in hours, where the coefficient 4 is multiplied to convert 15 minutes into 1 hour; T off,i is the minimum shutdown time of the i-th thermal power unit, in hours, It is a binary variable that measures whether the i-th thermal power unit is online at time n. T is the scheduling period of 1 day, with 15 minutes as the value.

[0128] (3) Output power constraints of thermal power units:

[0129]

[0130] Where, is the minimum output power of the i-th thermal power unit; is the maximum output power of the i-th thermal power unit.

[0131] (4) Constraints on the reserve capacity of thermal power units:

[0132]

[0133] Where, are the positive reserve and negative reserve capacity of the i-th thermal power unit respectively; in addition, the reserve is 10min spinning reserve, so Δt 10 The value is 10min.

[0134] Step S6: Large-scale wind power and photovoltaic power generation, when connected to the grid, introduces uncertainty into the system, posing a significant challenge to optimal scheduling. To address this challenge, a wind, solar, and load uncertainty processing model is constructed. Based on the random nature of wind, solar, and load, fuzzy chance constraints are employed to address the uncertainty of wind, solar, and load to a certain extent. The established fuzzy chance constraint model is as follows:

[0135] The optimization problem of fuzzy chance constraints with fuzzy parameters based on credibility theory can be expressed as:

[0136]

[0137] Where: f(x,ξ) is the objective function; x is the decision variable; ξ is the fuzzy parameter vector; α is the confidence level of the system; g(x,ξ) is the constraint function; C r {·} indicates the credibility of the fuzzy event.

[0138] The credibility measure is obtained from the possibility measure. For the possibility space (Θ, P(Θ), Pos), the credibility measure of event A is expressed as:

[0139]

[0140] Where: Pos{A} is the possibility measure of event A; A c is the opposite event of A.

[0141] Since fuzzy chance constraints cannot be solved directly, it is necessary to convert them into clear equivalence classes and then use traditional solution methods to calculate them. Therefore, clear equivalence classes of chance constraints with multiple fuzzy parameters are used. When α>0.5, the clear equivalence class of the chance constraint (Equation (49)) is:

[0142]

[0143] Where: r k1 ,r k2 ,r k3 ,r k4 is the four-tuple membership parameter of the k-th fuzzy parameter, corresponding to the boundary value of the prediction error; is the positive / negative deviation function associated with the kth fuzzy variable; h0(x) is a part of the constraint function g(x,ξ).

[0144] The above uncertainty processing method is used to process the uncertainty of wind power, photovoltaic output and load forecast errors in the power system with a high proportion of renewable energy. The processed model is as follows:

[0145] When using fuzzy chance constraints to clear equivalence classes and perform equivalent calculations, various membership functions need to be used to represent fuzzy parameters. Taking the trapezoidal membership function as an example, considering the prediction errors of wind power, photovoltaic power and load, a unified fuzzy parameter P is used. F To express it, the membership function can be expressed as follows:

[0146]

[0147] Where: μ(P F ) is the membership function; P F1 ,P F2 ,P F3 ,P F4 are the 1st, 2nd, 3rd, and 4th membership parameters, and have the following mathematical relationship with the predicted value:

[0148] P Fi =k i P fc (38)

[0149] Among them, PFi is the i-th membership parameter, k i is the relationship coefficient between the i-th membership parameter and the prediction error value.

[0150] The fuzzy parameters of wind power, photovoltaic and load forecast errors can be represented by the following four-tuple:

[0151]

[0152] Where: P fc Used to express the forecast error values of wind power, photovoltaic power, and load output; k1, k2, k3, and k4 are the 1st, 2nd, 3rd, and 4th proportional coefficients, respectively. When k2 = k3 = 1, the fuzzy parameters are triangular parameters.

[0153] The system reserve capacity model that uses fuzzy chance constraints to describe system uncertainty is as follows:

[0154]

[0155] Where, is the actual grid-connected power of wind power and photovoltaic power in the system; Fuzzy representation of wind power, photovoltaic and load forecast errors; N G is the number of thermal power units in the system.

[0156] The system should reserve some spinning reserve capacity to cope with the uncertainty of renewable energy output and load output. The size of the reserved spinning reserve capacity is determined by the size of the renewable energy output error and the load error. A triangular fuzzy membership function is used to describe the uncertainty of the high-proportion renewable energy sending-end system, as shown below:

[0157] The fuzzy parameters of the forecast errors of wind power, photovoltaic power and load using triangular fuzzy membership functions can be represented by the following triples:

[0158] H=(S1,S2,S3) (42)

[0159] Where H is the membership function; S1 is the lower limit of the parameter; S2 is the historical statistical data; and S3 is the upper limit of the parameter.

[0160] Then the fuzzy parameters of the model can be expressed as follows:

[0161]

[0162] Where, are the predicted values of wind power and photovoltaic power, respectively, k 1w 、k 3w are the first and third proportional coefficients of wind power prediction error respectively; k 1v 、k 3vare the first and third proportional coefficients of photovoltaic prediction error respectively; k 11 、k 31 are the first and third proportional coefficients of load forecast error respectively.

[0163] Based on the above analysis, the system with spinning reserve equations (30) and (31) can be transformed into a clear equivalence class that can be solved, as shown below:

[0164]

[0165] in, are the second and third membership parameters of the load forecast error respectively; are the first and second membership parameters of wind power prediction error respectively; are the first and second membership parameters of photovoltaic prediction error, respectively.

[0166] Step S7 constructs an objective function with the lowest total operating cost of the new energy sending-end system, taking into account dynamic frequency constraints, smoothing out new energy output fluctuations and improving system safety and stability.

[0167] To ensure the economy, safety and stability of the new energy sending-end system, we deeply explore the frequency regulation reserve capacity of conventional units and energy storage within the system, ensure that the sending-end power system has a certain amount of inertia support and primary frequency regulation resources during the operation phase, and through the rational formulation of unit start-up and shutdown and output plans, enable the system to maintain frequency indicators within a safe range under the expected power disturbance. With the minimum total operating cost of the new energy sending-end power system as the objective function, the following model is established:

[0168]

[0169] Where C G,r is the operating cost of thermal power units; C G,on is the startup cost of thermal power units; C wind Penalty cost for wind curtailment; C PV C is the penalty cost for abandoning light; ESS is the energy storage operating cost; C C,i (t) is the penalty cost of converting the energy storage from the discharge state to the charge state; C D,i (t) is the penalty cost for converting the energy storage from the charging state to the discharging state. The specific calculation method of each cost is shown in the following formula:

[0170]

[0171] Where T is the total number of unit combination periods, the optimization step is 15 minutes, then T = 96; t is the unit scheduling period: is the power generation cost coefficient of the i-th thermal power unit; is the active power output of the i-th thermal power unit in period t; is the single startup cost of the i-th thermal power unit; c p1 、c p2 Penalty cost coefficient for curtailing wind and solar power; are the predicted output and actual output of the i-th wind farm in period t respectively; are the predicted output and actual output of the i-th photovoltaic power station in period t respectively; is the net charging cost coefficient of the i-th energy storage.

[0172] To extend the service life of energy storage, a penalty constraint for energy storage state transition is added, thereby reducing the number of energy storage state transitions and, to a certain extent, improving the service life of energy storage. The penalty cost for energy storage charge and discharge state transition is defined as:

[0173]

[0174] Where: β is the penalty coefficient; is the rated capacity of the i-th energy storage; Soc i (t-1) is the internal energy of the i-th energy storage at time t-1; Soc i (t) is the internal energy of the i-th energy storage at time t; y i (t) is a binary variable, which is 1 when the energy storage is converted from discharge to charge, and 0 in other states; x i (t) is a binary variable, which is 1 when the energy storage is converted from charging to discharging, and 0 in other states. And the binary variables y of the two energy storage work conversions i (t), x i (t) is also related to the binary variable z that measures the working status of energy storage i (t) There is also the following mathematical relationship:

[0175]

[0176] Among them, z i (t) is a binary variable that measures the working status of energy storage at time t.

[0177] The primary frequency regulation constraint model of conventional thermal power units is as follows:

[0178]

[0179] Where, is the steady-state power change of the primary frequency regulation of the i-th thermal power unit under the power disturbance in period t, as shown below:

[0180]

[0181] From Equations (50) and (51), we can see that the model contains nonlinear terms of variable multiplication. Considering the use of the large M method to linearize the model, the constraint model after processing is as follows:

[0182]

[0183]

[0184] Where, X t,i,l and Y t,i,l is an intermediate variable used to assist linearization, where

[0185] The primary frequency regulation constraint model of energy storage is as follows:

[0186]

[0187] Where, is the steady-state power change of the primary frequency modulation of the i-th energy storage under the power disturbance in period t, as shown below:

[0188]

[0189] Among them, μ l is the regulation coefficient of thermal power unit l; It is a binary variable indicating whether the thermal power unit l is online, which is "1" if online and "0" if offline; are the maximum prediction errors of wind farm l and photovoltaic power station l, respectively; is the maximum charge and discharge power of energy storage l.

[0190] From Equations (54) and (55), we can see that the model contains nonlinear terms of variable multiplication. Considering the use of the large M method to linearize the model, the constraint model after processing is as follows:

[0191]

[0192] Where M t,i,l is an intermediate variable used to assist linearization, where

[0193] The system operation constraint model is as follows:

[0194] The new energy output constraint model is as follows:

[0195]

[0196] The power balance constraint model is as follows:

[0197]

[0198] The remaining system operation constraints are the maximum frequency difference constraint, the initial frequency change rate constraint, the steady-state frequency difference constraint in step S3, and the clear equivalence class of the system spare capacity in step S6.

[0199] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A dispatching method for a grid-connected renewable energy sending-end system considering dynamic frequency constraints. The method is characterized in that it comprises the following steps: Step S1: Construct an equivalent frequency response model for multi-resource coordination of a high-proportion new energy sending-end system; Step S2: Based on the equivalent frequency response model, dynamic frequency indicators are established, including maximum frequency deviation, initial frequency change rate, and steady-state frequency difference; Step S3: Based on the constructed dynamic frequency index, considering the system range and dead zone settings of each index, dynamic frequency constraints of maximum frequency deviation, initial frequency change rate, and steady-state frequency difference are established; Step S4: constructing a refined energy storage model; Step S5: constructing a conventional unit model; Step S6: Constructing a wind, solar, and load uncertainty processing model. Based on the randomness of wind, solar, and load, fuzzy chance constraints are used to process the uncertainty of wind and solar load to a certain extent. Step S7: Construct an objective function with the lowest total operating cost of the renewable energy sending-end system, taking into account dynamic frequency constraints, smoothing the fluctuation of renewable energy output and improving system safety and stability.

2. The method for dispatching a new energy transmission system of a grid-connected type according to claim 1, characterized in that: In step S1, a conventional thermal power unit frequency response model, an energy storage device frequency response model, a dynamic frequency response model of the system load, a maximum error model of the expected power disturbance step value as the load and new energy prediction are constructed, and a dynamic frequency response model of the system is constructed. An equivalent method is used to obtain a simplified thermal power unit transfer function model, and on this basis, an equivalent frequency response model is finally constructed; the simplified thermal power unit transfer function model is as follows: ； Where, is the transfer function of the thermal power unit, s is the independent variable in the complex frequency domain; 、 、 They are respectively the proportional coefficient, integral coefficient and differential coefficient of the equivalent speed regulator.

3. The method for dispatching a new energy transmission system of a grid-connected type according to claim 2, wherein the equivalent frequency response model is as follows: ； Where, 、 、 、 are the transformation coefficients of the transfer function of the system dynamic frequency response model under power disturbance; is the equivalent inertia coefficient of the system; 、 are the equivalent virtual inertia coefficient and frequency droop gain coefficient of the energy storage in the system respectively; is the system frequency variation under power disturbance; is the load frequency response coefficient; is the sum of the step powers of the power disturbance; 、 、 The calculation methods are as follows: ； in, is the number of thermal power units in the system; is the start / stop status of the i-th thermal power unit, 1 for start and 0 for stop; is the rated output of the i-th thermal power unit; is the inertia time constant of the i-th thermal power unit; The sum of all grid-connected energy storage quantities in the system; is the rated value of the charging and discharging power of the i-th energy storage; is the virtual inertia gain coefficient of the i-th energy storage; is the frequency droop gain coefficient of the i-th energy storage.

4. The method for dispatching a grid-connected new energy sending-end system according to claim 3 is characterized in that, in step S2, the calculation method of the three frequency indicators, namely the initial frequency change rate, the steady-state frequency difference, and the maximum frequency deviation, is derived based on the equivalent frequency response model.

5. The method for dispatching a new energy transmission system of a grid-connected type according to claim 3, wherein: In step S3, based on the maximum frequency deviation, initial frequency change rate, and steady-state frequency difference dynamic frequency index constructed in step S2, dynamic frequency constraints are established taking into account the range allowed by the system.

6. The method for dispatching a grid-type new energy sending-end system according to claim 1 is characterized in that the energy storage refinement model includes an energy storage state of charge model, energy storage capacity constraints, an energy storage charging and discharging power model, an energy storage energy management model, and an energy storage charging and discharging ramping model.

7. The method for dispatching a new energy transmission system of a grid-connected type according to claim 1, wherein: The conventional unit model includes thermal power unit ramping constraints, thermal power unit minimum on / off time constraints, thermal power unit output power constraints, and thermal power unit spare capacity constraints.

8. The method for dispatching a new energy transmission system of a grid-connected type according to claim 1, wherein: In step S6, the uncertainty of the forecast error of wind power, photovoltaic output and load in the high-proportion renewable energy sending-end power system is processed, and the processed model is as follows: When using fuzzy chance constraints to clear equivalence classes and perform equivalent calculations, various membership functions are required to represent fuzzy parameters. Taking the trapezoidal membership function as an example, the prediction errors of wind power, photovoltaic power, and load are considered to use a unified fuzzy parameter. To express; Fuzzy chance constraints are used to describe system uncertainty and construct a system reserve capacity model. The system should reserve some spinning reserve capacity to cope with the uncertainty of renewable energy output and load output; the size of the reserved spinning reserve capacity is determined by the size of the renewable energy output error and the load error; a triangular fuzzy membership function is used to describe the uncertainty of the high-proportion renewable energy sending-end system; The system spinning reserve capacity model is transformed into a form with clear equivalence classes that can be solved.

9. The method for dispatching a new energy transmission system of a grid-connected type according to claim 1, wherein: In step S7, the following model is established with the minimum total operating cost of the renewable energy sending-end power system as the objective function: ； Where, The operating cost of thermal power units; The startup cost of thermal power units; Penalty costs for wind curtailment; Penalty cost for abandoned light; The operating cost of energy storage; The penalty cost for converting energy storage from a discharge state to a charge state; It is the penalty cost for converting energy storage from charging state to discharging state.

10. The method for dispatching a new energy sending-end system of a grid-connected network according to claim 9, wherein the penalty cost for energy storage charge and discharge state conversion is expressed by the following formula: ； Where: is the penalty coefficient; is the rated capacity of the i-th energy storage; for The internal energy of the i-th energy storage at time; for The internal energy of the i-th energy storage at time; It is a binary variable, which is 1 when the energy storage is converted from discharge to charge, and 0 in other states; It is a binary variable, which is 1 when the energy storage is converted from charging to discharging, and 0 in other states; and the binary variables of the two energy storage working conversions 、 and a binary variable that measures the working status of energy storage There is also the following mathematical relationship: ； in, It is a binary variable that measures the working status of energy storage at time t.

Images