A dispatching method for virtual power plants to participate in inertia and reserve markets

Abstract

The present invention discloses a scheduling method for a virtual power plant to participate in an inertia and reserve market, comprising the following steps: (1) establishing a power system model and a virtual power plant model; (2) establishing an uncertainty power output model based on the power system model and the virtual power plant model; (3) establishing a frequency response model of the power system based on the power system model and the virtual power plant model; (4) establishing an optimal scheduling model for a synchronous generator and a virtual power plant based on the uncertainty power output model and the frequency response model; (5) solving the optimal scheduling model to obtain a scheduling instruction for the virtual power plant.

Description

Technical Field

[0001] The invention relates to a scheduling method for a virtual power plant to participate in an inertia and standby market, and belongs to the technical field of virtual power plants. Background Art

[0002] Renewable energy and electrochemical energy storage connected to the grid through power electronic inverters are flexible and fragile, and cannot spontaneously provide inertia and droop response like traditional synchronous generators. Therefore, the large-scale grid connection of power electronic inverter-based power supplies affects the frequency security of the power system, and the large-scale distributed source and load resources increase the burden of power system operation and management. In addition, the volatility of renewable energy generation brings uncertainty to the dispatch of the power system, affects the effective absorption of new energy, and affects the reliability of the dispatch and operation of the power system. The above problems have brought challenges to the safe and reliable operation of power systems with a high proportion of renewable energy.

[0003] To address the issue of safety, academia and industry have proposed a variety of inverter power control methods, such as droop control, virtual inertia control, virtual synchronous generator control, and virtual oscillator control, so that inverter-based power supplies can spontaneously perform virtual inertia response and frequency response. Wind power and photovoltaic power can both participate in system frequency response through the above inverter power control methods, and battery energy storage can provide backup and participate in frequency response at the same time. To address the issue of uncertainty in renewable energy, existing methods use the method of reserving unit backup to compensate for the uncertain new energy power deviation in real-time operation, and solve the uncertainty optimization scheduling problem of power systems through uncertainty optimization methods such as robust optimization, stochastic optimization, and chance-constrained optimization. Based on the above technologies, power system operators can guide controllable generators, renewable energy or energy storage to provide power backup and virtual inertia response by setting up backup markets and inertia markets, thereby improving the reliability and safety of power system operation.

[0004] In existing technical research, the dispatch of units participating in the power system reserve and inertia market is usually based on the day-ahead or intraday power system security constraint dispatch method. The system operator solves the dispatch instruction based on the bids of controllable synchronous generators and inverter power sources to determine the reserve capacity and inertia response capacity of each unit and power source. Existing technologies include market participation dispatch methods for synchronous generators and various power sources, including the application of wind power, photovoltaics, electric vehicles, and batteries in the reserve and inertia markets. However, ubiquitous source and load resources with small capacity and different characteristics are difficult to meet the market participation conditions, and the regulation potential of flexible resources cannot be fully utilized. Virtual power plant (VPP) technology, as an effective technical approach for the operation and management of large-scale, small-capacity distributed resources, can coordinate distributed source and load resources to create economic value and support the efficient operation of the power system, but there is currently a lack of optimized dispatch methods for virtual power plants to participate in the reserve and inertia markets.

[0005] At present, the dispatching method for virtual power plants to participate in the reserve and inertia markets is mainly based on the traditional power system construction based on synchronous generators. With the increase in the penetration rate of new energy and the withdrawal of traditional units, it is necessary to design a feasible optimization dispatching method for participating in the reserve and inertia markets based on the operating characteristics of distributed resources within the virtual power plant. Summary of the invention

[0006] In view of this, the present invention proposes a scheduling method for virtual power plants to participate in the inertia and reserve markets based on the operating characteristics of distributed resources within virtual power plants, so as to solve the problem that the current virtual power plant scheduling model lacks feasibility in participating in the new power system reserve and inertia markets.

[0007] In order to solve the above problems, the present invention proposes the following technical solutions:

[0008] A scheduling method for a virtual power plant to participate in an inertia and reserve market comprises the following steps: (1) establishing a power system model and a virtual power plant model; (2) establishing an uncertain power output model based on the power system model and the virtual power plant model; (3) establishing a frequency response model of the power system based on the power system model and the virtual power plant model; (4) establishing an optimal scheduling model for synchronous generators and virtual power plants based on the uncertain power output model and the frequency response model; and (5) solving the optimal scheduling model to obtain a scheduling instruction for the virtual power plant.

[0009] Furthermore, the step (1) of establishing the power system model includes: using a graph model of interconnected nodes and lines to describe the power system, and using a node admittance matrix, a generator node position matrix, a wind farm node position matrix, a virtual power plant node position matrix, and a load node position matrix to respectively describe the line parameters, the generator, the wind farm, the virtual power plant, and the location information of the load, so as to construct the power system model; wherein the time period set of the scheduling problem of the transmission network is defined as T , by the time t index.

[0010] Furthermore, the virtual power plant model in step (1) is described by the active power output, power regulation reserve and virtual inertia response capacity of the virtual power plant, and establishing the virtual power plant model includes:

[0011] Based on the power prediction results of distributed photovoltaic and adjustable load, the charging and discharging power and electric energy level of energy storage are considered, and the virtual power plant is solved in an aggregated manner. t The upper and lower limits of active power output at the moment;

[0012] Based on the demand response range of adjustable load, the reserve capacity of energy storage and the electric energy level of energy storage, the virtual power plant is solved in an aggregated manner. t The maximum power increase and power decrease values ​​set at the time;

[0013] Based on the virtual inertia control coefficient and rated capacity of distributed energy storage, the virtual power plant is solved in an aggregated manner. t The upper limit of the virtual inertia response capacity at the moment.

[0014] Furthermore, step (2) of establishing an uncertain power output model includes: describing the random error between the predicted power and the actual power based on a vector composed of uncertain variables; and representing the uncertain power regulation amount based on an affine proportional control strategy of a synchronous generator and a virtual power plant.

[0015] Furthermore, step (3) of establishing the frequency response model includes: establishing a primary frequency modulation output model of the synchronous generator; solving the total inertia capacity of the power system; using a swing equation constructed based on the primary frequency modulation output model and the total inertia capacity to establish the frequency response model and derive a frequency safety index.

[0016] Furthermore, step (4) of establishing the optimization scheduling model includes: (4.1) establishing the objective function of the synchronous generator and the virtual power plant participating in the optimization scheduling; (4.2) establishing the power regulation constraints of the synchronous generator based on the uncertainty power output model; (4.3) establishing the power regulation constraints of the virtual power plant based on the virtual power plant model and the uncertainty power output model; (4.4) establishing the line flow constraints of the power system; (4.5) establishing the frequency security constraints of the power system based on the frequency security index.

[0017] Furthermore, the objective function in step (4.1) includes the operating cost of the synchronous generator and the operating cost of the virtual power plant; wherein the operating cost of the synchronous generator includes the power generation fuel cost at all scheduling times, the primary frequency regulation reserve cost, and the upward and downward power regulation reserve cost, and the operating cost of the virtual power plant includes the power generation operating cost at all scheduling times, the inertia response cost, and the upward and downward power regulation reserve cost.

[0018] Furthermore, the power regulation constraints of the synchronous generators in step (4.2) include: constraints on the participation factors of the generation control of the synchronous generators; system balance constraints of the generation control, which are used to limit the sum of the generation control participation factors of all synchronous generators and all virtual power plants to 1, so as to balance the global uncertainty power deviation; upper and lower limit constraints on the output power of the synchronous generators; upper and lower limit constraints on the power increase and decrease reserve of the generation control of the synchronous generators, as well as the primary frequency regulation reserve; reliability distribution robustness opportunity constraints on the power increase and decrease reserve of the generation control of the synchronous generators;

[0019] The power regulation constraints of the virtual power plant described in step (4.3) include: constraints on the power generation control participation factor of the virtual power plant; constraints on the virtual inertia response capacity of the virtual power plant; upper and lower limit constraints on the output power of the virtual power plant; upper and lower limit constraints on the power increase and decrease of the virtual power plant; and reliability distribution constraint on the power increase and decrease of the virtual power plant.

[0020] Furthermore, step (4.4) of establishing the line flow constraint of the power system includes: first, establishing a line flow model of the power system; then, establishing the line flow constraint based on the line flow model to ensure that the probability of the line flow being between the upper flow limit and the lower flow limit is within a set range;

[0021] Establishing the frequency safety constraint of the power system in step (4.5) includes: obtaining the frequency minimum point safety threshold of the power system from the frequency safety index, and limiting the frequency of the power system to not be lower than the frequency minimum point safety threshold under power loss disturbance; limiting the total inertia of the power system to not be lower than the inertia level corresponding to the frequency change rate safety threshold; constraining the quasi-steady-state frequency deviation of the power system.

[0022] Furthermore, step (5) specifically includes: (5.1) based on the historical data of random variables existing in the constraints of the optimization scheduling model, solving the moment information of the uncertainty probability distribution, and constructing a fuzzy set of the probability distribution of random variables; (5.2) converting the distributionally robust chance constraint containing the uncertainty vector described by the fuzzy set into a standard form to obtain a non-convex distributionally robust chance constraint; (5.3) based on inequality approximation, reorganizing the non-convex distributionally robust chance constraint obtained in step (5.2) into a convex linear constraint or a second-order cone constraint; (5.4) using a solver to solve the optimization scheduling model containing the convex linear constraint or the second-order cone constraint to obtain the scheduling instructions of the virtual power plant.

[0023] The beneficial effects of the technical solution of the present invention are embodied in:

[0024] 1) The optimization dispatch model of the present invention can enable virtual power plants to participate in energy, inertia and reserve markets, aggregate and utilize the flexibility of distributed source and load resources;

[0025] 2) The power system frequency response model of the present invention takes into account load damping, frequency modulation response dead zone and other conditions to establish frequency safety constraints, and has higher accuracy;

[0026] 3) The present invention is based on the distributed robust chance constrained optimization method which does not rely on the probability information of uncertainty. It has better economy than the robust optimization method and higher computational efficiency than the scenario-based uncertainty optimization method. It can optimize the scheduling so that the synchronous generators and virtual power plants can compensate for the uncertainty of the output power of renewable energy and improve the reliability of system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a scheduling method for a virtual power plant participating in an inertia and reserve market according to an embodiment of the present invention.

[0028] Figure 2 It is a framework diagram of the virtual power plant participating in the reserve and inertia markets according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation methods and examples. The purpose of providing the examples is only to illustrate, and not to impose any limitation.

[0030] The present invention proposes a method for scheduling a virtual power plant to participate in the inertia and reserve market. Figure 1 The scheduling method includes the following steps (1) to (5):

[0031] Step (1), establish the power system model and the virtual power plant model, the specific process is as follows:

[0032] (1.1) Power system model:

[0033] Consider a transmission network where I , E , G , W , V , L They represent the node set, transmission line set, synchronous generator set, wind farm set, virtual power plant set, and load set in the power system respectively; correspondingly, i , j Represents the index of the node. ij , g , w , v , l They represent the index of transmission lines, the index of synchronous generators, the index of wind farms, the index of virtual power plants, and the index of loads respectively. i or j There will be one of the power sources or loads such as synchronous generators, wind farms, and virtual power plants on the line. ij Indicates that the nodes in the power system are connected i and nodes j Based on this, the above-mentioned graph model of interconnected nodes and lines can be used ( I , E ) describes the power system, and uses node admittance matrix, generator node location matrix, wind farm node location matrix, virtual power plant node location matrix, and load node location matrix to describe the line parameters, generator, wind farm, virtual power plant, and load location information, respectively, to build a power system model. Among them, the time period set of the transmission network scheduling problem is defined as T ,Depend on t index.

[0034] (1.2) Virtual power plant model:

[0035] Distributed resources in the power system can be divided into distributed photovoltaics, distributed energy storage, and adjustable loads, which are distributed in the distribution network of a single transmission grid load node. The distributed resources on a single load node are aggregated and managed by a commercial virtual power plant (Commercial VPP) to participate in power system dispatching and market transactions.

[0036] Distributed source and load resources are characterized by being numerous, wide, and having different characteristics. When modeling a virtual power plant, it is necessary to first aggregate the distributed source and load resource models to which they belong to obtain the overall output model of the virtual power plant. At present, the aggregation of distributed source and load resources in virtual power plants usually uses geometric methods to describe the feasible domain after aggregation. Aggregation of the output power feasible domain of distributed source and load resources, aggregation of the power reserve feasible domain, and aggregation of the virtual inertia feasible domain can be performed in the self-dispatching link of the virtual power plant. The specific technical means for the aggregation of distributed source and load resources can adopt existing methods, and the present invention does not limit this. The virtual power plant model of an embodiment of the present invention is described by its active power output, power regulation reserve, and virtual inertia response capacity. Therefore, establishing a virtual power plant model in this step specifically includes:

[0037] (1.2.1) Based on the power prediction results of distributed photovoltaic and adjustable load, considering the charging and discharging power and electric energy level of energy storage, the existing geometric aggregation method is used to solve the virtual power plant t Active power output upper limit at the moment With lower limit ;

[0038] (1.2.2) Based on the demand response range of adjustable loads, the reserve capacity of energy storage and the energy level of energy storage, the existing geometric aggregation method is used to solve the virtual power plant in t The maximum power increase value set at the moment With power down to maximum value ;

[0039] (1.2.3) Based on the virtual inertia control coefficient and rated capacity of distributed energy storage, the existing geometric aggregation method is used to solve the virtual power plant in t The upper limit of the virtual inertia response capacity at the moment .

[0040] In the embodiment of the present invention, it is assumed that the uncertainty fluctuation of the distributed photovoltaic output power in the virtual power plant can be compensated and adjusted by the energy storage and load inside the virtual power plant, which is carried out in the self-dispatching link of the virtual power plant.

[0041] Step (2): Establish an uncertain power output model based on the power system model and the virtual power plant model. This step describes the random error between the predicted power and the actual output power of the wind farm based on the vector composed of uncertain variables, and then represents the uncertain power regulation amount based on the affine proportional control strategy of the synchronous generator and the virtual power plant to achieve the establishment of the uncertain power output model.

[0042] First, the power system operator will dispatch synchronous generators and controllable sources based on the load and wind power forecast values. However, the actual output power of the wind farm usually deviates from the power forecast value with random errors.w At the scheduling time t The uncertainty of the power output error is ξ w,t , the uncertain power output of the wind farm can be modeled as:

[0043]

[0044] in, Represents wind farm w The output power, Represents the deterministic wind power forecast. t Global uncertainty power deviation of power system at every moment The uncertainty of power output error of all wind farms The sum is obtained, expressed as:

[0045] .

[0046] In order to ensure the reliability of power supply in the power system, synchronous generators and virtual power plants will reserve active power for backup, and compensate for the uncertainty of wind power output errors in real time with proportional control based on the automatic power generation control strategy. Considering the uncertainty output of power generation control (or "uncertain power adjustment"), based on the affine proportional control strategy of synchronous generators and virtual power plants, the uncertainty of the power output of synchronous generators and virtual power plants can be modeled as follows:

[0047]

[0048] in, represents the uncertainty of the synchronous generator power output, represents the deterministic synchronous generator planned output power, represents the uncertain power regulation of the synchronous generator, where is the power generation control participation factor of the synchronous generator; represents the uncertain power output of the virtual power plant, represents the deterministic planned output power of the virtual power plant, Indicates the composition of the uncertain power regulation of the virtual power plant, is the power generation control participation factor of the virtual power plant.

[0049] Step (3): Based on the aforementioned power system model and virtual power plant model, a frequency response model of the power system is established. This step mainly includes: first, establishing a primary frequency modulation output model of the synchronous generator; then, solving the total inertia capacity of the power system; and finally, using the swing equation constructed based on the primary frequency modulation output model and the total inertia capacity to establish the frequency response model of the power system and derive the frequency security index.

[0050] When the power system frequency fluctuates due to fault power disturbance, the synchronous generator can provide a primary frequency response through the droop response of its speed regulator. To simplify the calculation and obtain a solvable result, we model the frequency response of the synchronous generator as a linear response model, assuming that the ramp time constant of the synchronous generator frequency response power output is , considering the dead zone of the frequency response (usually 0.03~0.05Hz) corresponding to the FM response dead zone time The ramp time constant means the generator self From the moment, the regulated power increases linearly from 0 to the maximum regulated power. The maximum power regulation of the frequency response is determined by the frequency regulation reserve. Limited by the primary frequency regulation output model of synchronous generator It is expressed as follows:

[0051]

[0052] in, is the time variable of the frequency response.

[0053] The synchronous stepper can spontaneously provide rotational inertia when it is turned on. The inertia is recorded as ; The virtual power plant can provide virtual inertia response through internal distributed energy storage, which is expressed as ;Total inertia capacity of the power system The inertia of all synchronous generators and the virtual power plant inertia are added together to obtain:

[0054]

[0055] In the present invention, the power loss caused by the worst-case grid failure is modeled as a step power loss disturbance , whose amplitude is , frequency response model of power system It can be expressed by the following swing equation:

[0056]

[0057] in, D is the load damping coefficient of the power system, L t is the total load demand of the power system.

[0058] The frequency safety of power systems usually focuses on three indicators: the maximum rate of change of frequency (RoCoF), the lowest frequency point (Frequency nadir), and the quasi steady state frequency deviation, which correspond to the safety performance of the power system frequency at the time of fault occurrence, transient process, and quasi steady state process. Based on the frequency response model, we can solve these three frequency safety indicators.

[0059] Maximum frequency change rate At the moment of disturbance, it is related to the magnitude of fault power disturbance and the inertia level of the power system, expressed as:

[0060]

[0061] Substituting the primary frequency modulation output model of the synchronous generator into the power system frequency response model, a piecewise nonlinear expression can be obtained. At a certain moment in the period, the frequency regulation output power of all generators in the power system reaches the power disturbance value, and the system frequency reaches the lowest point. Take the synchronous generator primary frequency regulation output model for this period of time Substitute into the swing equation of the power system frequency response model mentioned above and solve the differential equation corresponding to the frequency deviation change rate equal to 0 The time variable expression of the lowest frequency point can be obtained , and then Substitution Frequency response model for time segment The expression can get the expression of the lowest frequency point For the nonlinear frequency minimum point expression, it can be solved by piecewise linearization method to obtain The linear approximation expression of .

[0062] Quasi-steady-state frequency deviation is the frequency deviation of the power system after the frequency response of all generators is fully output. The result is determined by the sum of the frequency regulation reserves (frequency regulation reserves) of all generators. The sum of the frequency regulation reserves of all synchronous generators is , then the quasi-steady-state frequency deviation of the power system is The calculation expression is:

[0063]

[0064] The frequency response model and frequency security index of the power system established in this step will be used in the power system frequency security constraint conditions in the subsequent step (4) to characterize the frequency security requirements of the power system.

[0065] Step (4): Based on the uncertain power output model and frequency response model, an optimal dispatch model of synchronous generators and virtual power plants is established. This step specifically includes the following aspects:

[0066] (4.1) Establish the objective function for the optimal dispatch of synchronous generators and virtual power plants;

[0067] (4.2) Based on the uncertain power output model, establish the power regulation constraints of the synchronous generator;

[0068] (4.3) Based on the virtual power plant model and the uncertain power output model, establish the power regulation constraints of the virtual power plant;

[0069] (4.4) Establish line flow constraints for the power system;

[0070] (4.5) Based on the frequency security index obtained in step (3), establish the frequency security constraint of the power system.

[0071] The optimization dispatch model of this step is the core part of the method proposed in this invention, which is mainly composed of the objective function of (4.1) and the constraints in (4.2)~(4.5). The role of this step is to consider the operating costs, operating constraints and safety constraints of the power system, synchronous generators and virtual power plants, establish a solvable optimization dispatch model, and obtain the operating instructions corresponding to the virtual power plant participating in the inertia and reserve markets through solving. The following is a detailed description of steps (4.1)~(4.5).

[0072] (4.1) Establish the objective function of synchronous generators and virtual power plants participating in the optimal dispatch:

[0073] The objective function of synchronous generators and virtual power plants participating in optimal scheduling includes the operating costs of synchronous generators and virtual power plants; among them, the operating costs of synchronous generators include the power generation fuel costs at all scheduling times, the primary frequency regulation reserve costs, and the upward and downward power regulation reserve costs; the operating costs of virtual power plants include the power generation operating costs at all scheduling times, the inertia response costs, and the upward and downward power regulation reserve costs.

[0074] The goal of optimizing scheduling in the present invention is to minimize the objective function, which is expressed as follows:

[0075]

[0076] in, , , , Synchronous generator The power generation cost coefficient, primary frequency regulation reserve cost coefficient, power generation control upward reserve cost coefficient and downward reserve cost coefficient; and It is used to adjust the power up and down for the synchronous generator power generation control. , , , Virtual Power Plant v exist t The moment-aggregated power generation cost coefficient, the aggregated virtual inertia control cost coefficient, the aggregated power generation control upward reserve cost coefficient and the downward reserve cost coefficient; and Provide standby for increasing and decreasing the power of virtual power plants.

[0077] (4.2) Based on the uncertain power output model, the power regulation constraints of the synchronous generator are established:

[0078] The constraint of the synchronous generator generation control participation factor is expressed as:

[0079]

[0080] The system balance constraint of power generation control limits the sum of the power generation control participation factors of all synchronous generators and all virtual power plants to 1, which balances the global uncertainty power deviation. .

[0081]

[0082] The upper and lower limits of the synchronous generator output power are expressed as:

[0083]

[0084] in, and It is the upper and lower limits of synchronous generator power output.

[0085] The power up and down reserve of synchronous generator generation control, as well as the upper and lower limit constraints of primary frequency regulation reserve are expressed as follows:

[0086]

[0087] in, and They are respectively the maximum value of the power-up reserve and the maximum value of the power-down reserve for the synchronous generator power generation control.

[0088] The reliability distribution constraint of the power up and down reserve of synchronous generator generation control is expressed as:

[0089]

[0090] The meaning of the reliability distribution robust chance constraint is explained as follows: for the uncertainty vector In the fuzzy set All probability distributions within , ensuring that the probability of the constraints in the above brackets being true is greater than a certain lower bound probability , that is, the uncertainty output of synchronous generator power generation control The probability that the reserve is less than the upward reserve and greater than the downward reserve is at least is the pre-set feasibility probability threshold of the backup reliability constraint.

[0091] (4.3) Based on the virtual power plant model and the uncertain power output model, the power regulation constraints of the virtual power plant are established:

[0092] The constraint of the generation control participation factor of the virtual power plant is expressed as:

[0093]

[0094] The constraint on the virtual inertia response capacity of the virtual power plant is expressed as:

[0095]

[0096] The upper and lower limits of the virtual power plant output power are expressed as:

[0097]

[0098] in, and The virtual power plant t The upper and lower limits of the predicted power output at the time.

[0099] The upper and lower limit constraints of the virtual power plant power increase and decrease are expressed as:

[0100]

[0101] in, and The virtual power plant at time t The maximum value of power increase and power decrease that are set.

[0102] The reliability distribution constraint of the virtual power plant power up and down reserve is:

[0103]

[0104] For the uncertainty vector In the fuzzy set The above two constraints ensure the uncertainty output of virtual power plant power regulation. The probability that the reserve is less than the upward reserve and greater than the downward reserve is at least is the pre-set feasibility probability threshold of the backup reliability constraint.

[0105] (4.4) Establish line flow constraints for the power system:

[0106] First, establish the line flow of the power system The model is represented as follows:

[0107]

[0108] in, represents the shift factor matrix of the power system, with the subscript Represents the matrix OK; , , They are the power injection matrices of synchronous generators, virtual power plants, and wind farms; They are t The generator output power vector, virtual power plant output power vector, wind farm output power vector and load demand vector are synchronized at all times.

[0109] Based on the above line power flow model, the power flow constraint is established as follows:

[0110]

[0111] For the uncertainty vector In the fuzzy set All probability distribution conditions within the line are guaranteed At the upper limit and lower limit The probability between is the pre-set feasibility probability threshold of the power flow constraint.

[0112] (4.5) Based on the frequency security index obtained in step (3), establish the frequency security constraint of the power system.

[0113] The power system frequency minimum point constraint is solved by the method in step (3) of the present invention. It should be noted that the original frequency minimum point expression is a nonlinear function. In order to make the optimization scheduling model solvable, we use the piecewise linearization method to obtain an approximate linear expression set of the constraint, which is expressed as follows:

[0114]

[0115] Among them, the definition SThe multidimensional space feasible domain of the nonlinear frequency minimum point function is the piecewise linearization method. S Divide it into multiple sub-feasible domains, and use linear expressions to approximate the original function on the current sub-feasible domain. , The decision variables are , The linear coefficient of The above three parameters can be solved by the least square method of linear approximation. The frequency minimum point constraint limits the system frequency to not less than the system frequency minimum point safety threshold under power loss disturbance. .

[0116] The power system frequency change rate constraint is actually a limit on the system inertia, so that the total system inertia is not lower than the frequency change rate safety threshold. The corresponding inertia level is expressed as:

[0117]

[0118] The constraint of the quasi-steady-state frequency deviation of the power system is actually a constraint on the primary frequency regulation reserve of the generator, which is expressed as follows:

[0119]

[0120] in, is the safety threshold of quasi-steady-state frequency deviation.

[0121] Step (5) solves the optimal dispatch model of the synchronous generator and the virtual power plant to obtain the dispatch instruction of the virtual power plant. It mainly includes the following steps:

[0122] (5.1) Based on the historical data of random variables in the constraints of the optimization scheduling model, the moment information of the uncertainty probability distribution is solved and the fuzzy set of the random variable probability distribution is constructed;

[0123] (5.2) The distributed robustness chance constraint containing uncertainty vectors described by fuzzy sets is transformed into a standard form, and a non-convex distributed robustness chance constraint is obtained;

[0124] (5.3) Based on the inequality approximation, the non-convex distributed robustness constraint obtained in step (5.2) is reorganized into a convex linear constraint or a second-order cone constraint;

[0125] (5.4) The solver is used to solve the optimization scheduling model containing the convex linear constraints or second-order cone constraints to obtain the scheduling instructions of the virtual power plant.

[0126] The random power deviation vector (uncertainty vector) of all wind farms in the power system is recorded as (Note: Here T means transpose, |W | is the number of wind farms). Based on the moment information of the random power deviation vector, the fuzzy set of random power deviation probability distribution can be established: , defined as follows:

[0127]

[0128] in, is the random power deviation vector The first-order moment information of The mean of historical sample data ; is a random vector The second-order moment information of The covariance of the historical sample data .

[0129] For the fuzzy set The uncertainty vector described The distributional robustness chance constraint of can be modeled as the following standard distributional robustness chance constraint:

[0130]

[0131] in, a ( x )and b ( x ) is the decision variable x The mapping function of a ( x )The T in the upper right corner indicates transpose. is the permissible probability of exceeding the limit of the pre-set constraint.

[0132] The reliability distributed robustness opportunity constraints of the synchronous generator power regulation reserve, the reliability distributed robustness opportunity constraints of the virtual power plant power regulation reserve, and the line flow constraints involved in step (4) of the present invention can all be expressed as the distributed robustness opportunity constraints of the above standards. Based on the inequality approximation, the non-convex distributed robustness opportunity constraints can be reorganized into convex linear constraints or second-order cone constraints, which are expressed as follows:

[0133]

[0134] Using this formula, the distributed robustness chance constraints that cannot be directly solved in steps (4.2) to (4.3) can be reformulated into convex linear constraints or second-order cone constraints.

[0135] Furthermore, the optimization scheduling model containing linear constraints or second-order cone constraints is a second-order cone programming problem, which can be solved using commercial solvers such as CPLEX, GUROBI, MOSEK and other software.

[0136] The solution of the optimization scheduling model includes the following: Any synchronous generator The active power output, synchronous generator power generation control participation factor, synchronous generator primary frequency regulation standby, synchronous generator power generation control power up and down standby are expressed as ; The solution results related to the virtual power plant include each dispatch time Install a virtual power plant The active power output, virtual power plant generation control participation factor, virtual power plant inertia response capacity, virtual power plant generation control power up and down reserve are expressed as After solving the optimal dispatch model, the dispatching agency of the power system can use the above solution results to dispatch at different times. t The dispatching instructions are issued to the synchronous generators and virtual power plants in the system, thereby realizing the dispatching of synchronous generators and virtual power plants. The virtual power plant outputs the corresponding virtual inertia and reserved reserve based on the dispatching results, and can participate in the reserve and inertia market in the power market.

[0137] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art of the present invention, several equivalent substitutions or obvious variations can be made without departing from the concept of the present invention, and the performance or use is the same, which should be regarded as belonging to the protection scope of the present invention.

Claims

1. A scheduling method for a virtual power plant to participate in the inertia and reserve market, characterized in that: The following steps are involved: (1) Establish power system model and virtual power plant model; The virtual power plant model is described by the active power output, power regulation reserve and virtual inertia response capacity of the virtual power plant. Establishing the virtual power plant model includes: Based on the power prediction results of distributed photovoltaic and adjustable load, the charging and discharging power and electric energy level of energy storage are considered, and the virtual power plant is solved in an aggregated manner. t The upper and lower limits of active power output at the moment; Based on the demand response range of adjustable load, the reserve capacity of energy storage and the electric energy level of energy storage, the virtual power plant is solved in an aggregated manner. t The maximum power increase and power decrease values ​​set at the time; Based on the virtual inertia control coefficient and rated capacity of distributed energy storage, the virtual power plant is solved in an aggregated manner. t The upper limit of the virtual inertia response capacity at the moment; (2) establishing an uncertain power output model based on the power system model and the virtual power plant model; (3) establishing a frequency response model of the power system based on the power system model and the virtual power plant model; (4) establishing an optimal scheduling model of synchronous generators and virtual power plants based on the uncertain power output model and the frequency response model; (5) Solve the optimization scheduling model to obtain the scheduling instructions of the virtual power plant.

2. The scheduling method according to claim 1, characterized in that: The establishment of the power system model in step (1) includes: The power system is described by using a graph model of interconnected nodes and lines, and the node admittance matrix, generator node position matrix, wind farm node position matrix, virtual power plant node position matrix, and load node position matrix are used to describe the line parameters, generators, wind farms, virtual power plants, and load location information, respectively, so as to construct the power system model; wherein the time period set of the transmission network scheduling problem is defined as T , by the time t index.

3. The scheduling method according to claim 1, characterized in that: Step (2) of establishing an uncertain power output model includes: The random error between the predicted power and the actual power is described based on the vector composed of uncertain variables; Based on synchronous generators, the affine proportional control strategy of virtual power plants represents the uncertain power regulation quantity.

4. The scheduling method according to claim 1, characterized in that: Step (3) of establishing the frequency response model comprises: Establish the primary frequency modulation output model of synchronous generator; Solve for the total inertia capacity of the power system; The frequency response model is established by using the swing equation constructed based on the primary frequency modulation output model and the total inertia capacity, and the frequency safety index is derived.

5. The scheduling method according to claim 4, characterized in that: Step (4) of establishing the optimization scheduling model includes: (4.1) Establish the objective function for the optimal dispatch of synchronous generators and virtual power plants; (4.2) establishing a power regulation constraint of a synchronous generator based on the uncertain power output model; (4.3) establishing power regulation constraints of the virtual power plant based on the virtual power plant model and the uncertain power output model; (4.4) Establish line flow constraints for the power system; (4.5) Based on the frequency security index, establish frequency security constraints for the power system.

6. The scheduling method according to claim 5, characterized in that: The objective function in step (4.1) includes the operating cost of the synchronous generator and the operating cost of the virtual power plant; wherein the operating cost of the synchronous generator includes the power generation fuel cost at all scheduling times, the primary frequency regulation reserve cost, and the upward and downward power regulation reserve cost; the operating cost of the virtual power plant includes the power generation operating cost at all scheduling times, the inertia response cost, and the upward and downward power regulation reserve cost.

7. The scheduling method according to claim 5, characterized in that: The power regulation constraints of the synchronous generator in step (4.2) include: Constraints on the participation factor of synchronous generator generation control; System balance constraints for power generation control, which are used to limit the sum of the power generation control participation factors of all synchronous generators and all virtual power plants to 1, so as to balance the global uncertain power deviation; Upper and lower limits of synchronous generator output power; The power up and down reserve of synchronous generator power generation control, as well as the upper and lower limit constraints of primary frequency regulation reserve; The reliability of the synchronous generator power generation control power up and down standby is based on the opportunity constraints of the blue stick; The power regulation constraints of the virtual power plant in step (4.3) include: Constraints on the participation factors of generation control in virtual power plants; The virtual inertia response capacity constraints of the virtual power plant; Upper and lower limits of the output power of the virtual power plant; The upper and lower limits of the virtual power plant's power adjustment and reserve adjustment; The reliability of the virtual power plant's power adjustment up and down is subject to the opportunistic constraints.

8. The scheduling method according to claim 5, characterized in that: Step (4.4) of establishing the line flow constraint of the power system includes: first, establishing a line flow model of the power system; then, establishing the line flow constraint based on the line flow model to ensure that the probability of the line flow being between the upper flow limit and the lower flow limit is within a set range; Establishing the frequency safety constraint of the power system in step (4.5) includes: obtaining the frequency minimum point safety threshold of the power system from the frequency safety index, and limiting the frequency of the power system to not be lower than the frequency minimum point safety threshold under power loss disturbance; limiting the total inertia of the power system to not be lower than the inertia level corresponding to the frequency change rate safety threshold; constraining the quasi-steady-state frequency deviation of the power system.

9. The scheduling method according to claim 1, characterized in that: Step (5) specifically includes: (5.1) Based on the historical data of random variables in the constraints of the optimization scheduling model, solving the moment information of the uncertainty probability distribution and constructing the fuzzy set of the random variable probability distribution; (5.2) The distributed robust chance constraint containing uncertainty vectors described by fuzzy sets is transformed into a standard form, and a non-convex distributed robust chance constraint is obtained; (5.3) Based on the inequality approximation, the non-convex distributed robustness constraint obtained in step (5.2) is reorganized into a convex linear constraint or a second-order cone constraint; (5.4) The solver is used to solve the optimization scheduling model containing the convex linear constraints or second-order cone constraints to obtain the scheduling instructions of the virtual power plant.

Images