Quantitative evaluation method and device for adaptive frequency modulation active regulation power of wind storage system

Abstract

The present application relates to the technical field of wind storage system participating in power grid frequency modulation, and specifically provides an active regulation power quantitative evaluation method and device for adaptive frequency modulation of a wind storage system, comprising: constructing a frequency response model of the energy storage system to calculate the active power of the energy storage participating in frequency modulation; constructing a wind speed fluctuation model of the wind turbine to simulate the wind speed fluctuation in the same wind farm according to the wind speed probability distribution to generate a random wind speed disturbance; constructing a frequency response model of the wind turbine to calculate the active power participating in frequency modulation; constructing an adaptive allocation model of the wind storage system based on a deep deterministic policy gradient algorithm to realize optimal allocation of virtual droop control coefficients and virtual inertia control coefficients of the wind storage system; and calculating the active regulation power after adaptive frequency modulation of the wind storage system. The adjustable active power of the wind turbine and the energy storage device of each participating frequency modulation strategy is calculated to achieve the best frequency modulation effect and equipment life balance.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wind storage system participating in grid frequency modulation, and particularly relates to a method and device for quantitatively evaluating active regulation power of adaptive frequency modulation of a wind storage system. BACKGROUND

[0002] Due to the natural characteristics of wind power, such as the instability of wind power, the power generation of wind power has the characteristics of discontinuity and unpredictability. When a large amount of wind power is connected to the power grid, the inertia of the power grid will be reduced, thereby reducing the stability of the power grid in the face of frequency fluctuations. In order to enhance the stability of the power grid, the wind farm must be able to provide inertia response and primary frequency modulation function to help maintain the frequency stability of the power grid.

[0003] After the wind farm integrates the energy storage system, the energy storage needs to cooperate with the wind power to participate in the frequency modulation to improve the frequency modulation effect. While performing peak shaving and absorbing or emitting active power, a certain amount of active power should also be left to meet the frequency modulation demand. In order to make full use of the support power of the wind storage system as much as possible and avoid resource waste, it is necessary to quantitatively evaluate the frequency modulation capacity of the wind storage system. However, under the fluctuation of wind speed, the strategies and states of wind turbine units participating in frequency modulation are not the same, and there is no unified quantitative standard for the power used for frequency modulation. When the state of charge (SOC) of the energy storage is close to the upper and lower limit values, the problem of insufficient energy storage power may occur. If the energy storage is controlled to charge and discharge at the maximum, and the active power is insufficient, it may lead to failure of primary frequency modulation, and ultimately cause the system frequency deviation to increase. Or the energy storage is frequently operated beyond the maximum power due to frequency modulation, which may cause damage to the energy storage equipment. In addition, due to the capacity limitation of the energy storage, a certain amount of active power needs to be left for charging and discharging at low state of charge and high state of charge to meet the frequency modulation demand and prevent overcharging or overdischarging.

[0004] During the integration of the energy storage system in the wind farm, the wind storage system as a part of the overall power system is not considered; the frequency modulation strategy is relatively single, and the adaptive frequency modulation control potential of the wind storage system cannot be fully utilized, so that the adjustable active power of the wind storage system cannot be accurately quantified. SUMMARY

[0005] In view of the above deficiencies of the prior art, the present application provides a method for quantitatively evaluating the active regulation power of adaptive frequency modulation of a wind storage system, to solve the above technical problems.

[0006] In a first aspect, the present application provides a method for quantitatively evaluating the active regulation power of adaptive frequency modulation of a wind storage system, comprising the following steps:

[0007] obtaining the charging and discharging state and the state of charge of the energy storage system, and adjusting the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage participating in frequency modulation according to the charging and discharging state and the state of charge;

[0008] According to the power output characteristics of the wind turbine, a second virtual droop control coefficient and a second virtual inertia control coefficient of the wind turbine participating in frequency modulation are obtained;

[0009] An adaptive allocation model of the wind storage system is constructed, and optimal weight configurations are respectively performed on the first virtual droop control coefficient and the first virtual inertia control coefficient based on the adaptive allocation model of the wind storage system, so as to calculate corresponding first virtual droop control frequency modulation power and first virtual inertia control frequency modulation power;

[0010] The second virtual droop control coefficient is updated based on the first virtual droop control frequency modulation power after the weight configuration, and the second virtual inertia control coefficient is updated based on the first virtual inertia control frequency modulation power after the weight configuration, so as to calculate the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power;

[0011] Based on the first virtual droop control frequency modulation power, the first virtual inertia control frequency modulation power, the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power, an active power adjustment power of the wind storage system after adaptive frequency modulation is calculated.

[0012] In an optional embodiment, the adaptive allocation model of the wind storage system is constructed based on a deep deterministic policy gradient algorithm, a wind speed fluctuation model of the wind turbine is constructed based on a Weibull distribution, and an intelligent agent is trained based on the deep deterministic policy gradient algorithm under wind speed fluctuation of the wind turbine, so as to achieve optimal allocation of the first virtual droop control coefficient and the first virtual inertia control coefficient, specifically including:

[0013] Initial parameters are set for a policy network, a target policy network, a value network and a target value network;

[0014] Simulated wind speeds are generated according to the Weibull distribution of the wind speed, and continuous disturbance data is generated through the wind turbine;

[0015] The intelligent agent determines an action based on the current system frequency deviation and the frequency deviation change rate, and the energy storage system performs weight configuration of the corresponding first virtual droop control coefficient and virtual inertia control coefficient;

[0016] After the intelligent agent performs the action, a current reward and a system state at the next time are fed back, experience data of the current system state, the action taken, the reward obtained and the next system state are stored in a replay buffer, and samples are extracted from the replay buffer to update network parameters;

[0017] The above steps are repeated until the maximum number of steps in a single round is reached;

[0018] The round iteration is repeatedly performed until a preset maximum training round number is reached, and optimal weight configurations of the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system are obtained.

[0019] In an optional implementation, the obtained reward includes a reward and a reward :

[0020]

[0021] wherein, and are the proportion coefficients of the two-part reward, is the frequency difference of the system, is the power frequency modulation control change of the wind storage system, the reward measures the pros and cons of the effect of the action of the intelligent agent on the frequency modulation of the energy storage, measures the situation that the long-term power participation in frequency modulation control of the wind storage system affects its service life.

[0022] In an optional implementation, after obtaining the optimal weight configuration, the corresponding first virtual inertia control frequency modulation power and first virtual droop control frequency modulation power are calculated, including:

[0023]

[0024]

[0025]

[0026] wherein, is the first virtual inertia control frequency modulation power after the optimal weight configuration, is the first virtual droop control frequency modulation power after the optimal weight configuration, is the active power of the energy storage participating in frequency modulation after the optimal weight configuration, , is the optimal first virtual inertia weight coefficient, is the optimal first virtual droop weight coefficient, denotes the first virtual inertia control coefficient of the energy storage system, denotes the first virtual droop control coefficient of the energy storage system, is the frequency change rate, is the frequency deviation.

[0027] In an optional implementation, the updated second virtual droop control coefficient and the updated second virtual inertia control coefficient specifically include:

[0028]

[0029]

[0030] wherein, and is the adjustable power constraint, is the MPPT point output power, is the initial load shedding output power, is the rotor kinetic energy inertia response release power, is the updated second virtual droop control coefficient, is the updated second virtual inertia control coefficient.

[0031] In an optional embodiment, the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power are specifically calculated as:

[0032] The active power change value of the wind turbine through virtual inertia control is calculated as:

[0033]

[0034]

[0035]

[0036] wherein, is the rated power frequency of the power grid, is the rated active power of the wind turbine, is the second virtual inertia control coefficient of the wind turbine, is the active power change value of the wind turbine, is the inertia response power change limiting lower limit value, is the inertia response power change limiting upper limit value, is the active power given reference value of the wind turbine during the inertia response, is the active power value at the time when the inertia response frequency modulation frequency starts to change;

[0037] The active power change value of the wind turbine through virtual droop control is calculated as:

[0038]

[0039]

[0040]

[0041] wherein, is the second virtual droop control coefficient of the primary frequency modulation at the time of frequency drop disturbance or frequency rise disturbance, is the primary frequency modulation power change limiting lower limit value, is the primary frequency modulation power change limiting upper limit value, is the wind turbine operating change frequency,​​ a deviation value for the primary frequency regulation frequency; a reference value for the wind turbine active power setpoint during the primary frequency regulation, an active power value at the time when the primary frequency regulation frequency starts to change;

[0042] calculating the active power of the wind turbine participating in the frequency regulation:

[0043]

[0044] wherein, the active regulation power value of the wind turbine participating in the frequency regulation.

[0045] In an optional embodiment, the calculating the active regulation power of the wind storage system after adaptive frequency regulation specifically comprises:

[0046] calculating the active regulation power of the wind turbine after adaptive frequency regulation :

[0047]

[0048]

[0049] wherein, N represents the number of wind turbines in the wind storage system; is the active regulation power of the virtual inertia and virtual droop control of the Nth wind turbine participating in the frequency regulation; calculating the active regulation power of the wind storage system after adaptive frequency regulation

[0050] :

[0051]

[0052]

[0053] wherein, N represents the number of battery energy storage devices in the wind storage system; is the active regulation power of the virtual inertia and virtual droop control of the Nth battery energy storage device participating in the frequency regulation; the active regulation power of the wind storage system after adaptive frequency regulation

[0054] is calculated as:

[0055] .

[0056] In an optional embodiment, the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system are configured according to the comprehensive optimal weight, and a transfer function of a frequency response model of the energy storage system is calculated.​​

[0057]

[0058] wherein, is the energy storage system response time constant, is the transformation parameter in the Laplace transform.

[0059] In an optional embodiment, the virtual inertia control of the wind turbine is a rotor kinetic energy control, the virtual droop control of the wind turbine is mainly a pitch angle control, and the frequency response model transfer function of the wind turbine is calculated by integrating the updated second virtual droop control coefficient and the updated second virtual inertia control coefficient:

[0060]

[0061] wherein, is the rotor inertia response time constant, is the pitch response time constant.

[0062] In a second aspect, an active regulation power quantification evaluation device for adaptive frequency regulation of a wind storage system is provided, comprising:

[0063] An energy storage coefficient acquisition module acquires the charge and discharge state and the state of charge of the energy storage system, and adjusts the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage participating in frequency regulation according to the charge and discharge state and the state of charge;

[0064] A wind turbine coefficient acquisition module acquires the second virtual droop control coefficient and the second virtual inertia control coefficient of the wind turbine participating in frequency regulation according to the power output characteristics of the wind turbine;

[0065] An energy storage coefficient weight configuration module constructs an adaptive allocation model of the wind storage system, and performs optimal weight configuration for the first virtual droop control coefficient and the first virtual inertia control coefficient based on the adaptive allocation model of the wind storage system, and calculates the corresponding first virtual droop control frequency regulation power and the first virtual inertia control frequency regulation power;

[0066] A wind turbine coefficient update module updates the second virtual droop control coefficient based on the first virtual droop control frequency regulation power after the weight configuration and calculates the second virtual droop control frequency regulation power, and updates the second virtual inertia control coefficient based on the first virtual inertia control frequency regulation power after the weight configuration and calculates the second virtual inertia control frequency regulation power;

[0067] An active regulation power calculation module calculates the active regulation power of the adaptive frequency regulation of the wind storage system based on the first virtual droop control frequency regulation power, the first virtual inertia control frequency regulation power, the second virtual droop control frequency regulation power, and the second virtual inertia control frequency regulation power.

[0068] In an optional embodiment, the wind storage system adaptive allocation model is constructed based on a deep deterministic policy gradient algorithm, a wind turbine wind speed fluctuation model is constructed based on a Weibull distribution, and an agent is trained based on the deep deterministic policy gradient algorithm under wind turbine wind speed fluctuation to achieve optimal allocation of the first virtual droop control coefficient and the first virtual inertia control coefficient, and specifically includes:

[0069] Setting initial parameters for the policy network, the target policy network, the value network, and the target value network;

[0070] Generating simulated wind speeds according to the Weibull distribution of wind speeds, and generating continuous disturbance data through wind turbines;

[0071] The agent determines the action based on the current system frequency deviation and the frequency deviation change rate, and the energy storage system executes the corresponding weight configuration of the first virtual droop control coefficient and the virtual inertia control coefficient;

[0072] After the agent executes the action, the current reward and the next system state are fed back, and the experience data of the current system state, the action taken, the reward obtained, and the next system state are stored in the replay buffer, and samples are extracted from the replay buffer to update the network parameters;

[0073] Repeat the above steps until the maximum number of steps per round is reached;

[0074] Continuously repeat the round iteration until the maximum preset training round number is reached, and obtain the optimal weight configuration of the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system.

[0075] In an optional embodiment, the reward obtained includes reward and reward :

[0076]

[0077] wherein, and are the proportion coefficients of the two parts of the reward, is the frequency deviation of the system, is the power frequency control change of the wind storage system, reward measures the pros and cons of the action of the agent on the effect of the energy storage on frequency regulation, and reward measures the situation that the long-term power participation in frequency regulation control of the wind storage system affects its service life.

[0078] In an optional embodiment, after obtaining the optimal weight configuration, the corresponding first virtual inertia control frequency regulation power and first virtual droop control frequency regulation power are calculated, and the frequency regulation power includes:

[0079]

[0080]

[0081]

[0082] wherein, is the first virtual inertia control frequency modulation power after optimal weight configuration, is the first virtual droop control frequency modulation power after optimal weight configuration, is the active power of energy storage participating in frequency modulation after optimal weight configuration, , is the optimal first virtual inertia weight coefficient, is the optimal first virtual droop weight coefficient, represents the first virtual inertia control coefficient of the energy storage system, represents the first virtual droop control coefficient of the energy storage system, is the frequency change rate, is the frequency deviation.

[0083] In an optional implementation, the updated second virtual droop control coefficient and the updated second virtual inertia control coefficient specifically include:

[0084]

[0085]

[0086] wherein, and is the adjustable power constraint, is the MPPT point output power, is the initial load shedding output power, is the rotor kinetic energy inertia response release power, is the updated second virtual droop control coefficient, is the updated second virtual inertia control coefficient.

[0087] In an optional implementation, the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power are specifically calculated as:

[0088] the active power change value of the wind turbine through virtual inertia control is calculated as:

[0089]

[0090]

[0091] ​

[0092] wherein, is the rated power frequency of the power grid, is the rated active power of the wind turbine, is the second virtual inertia control coefficient of the wind turbine, is the active power change value of the wind turbine, is the lower limit value of the inertia response power change limit, is the upper limit value of the inertia response power change limit, is the given reference value of the active power of the wind turbine during the inertia response, is the active power value at the time when the frequency adjustment frequency starts to change;

[0093] calculating the active power change value of the wind turbine through virtual droop control :

[0094]

[0095]

[0096]

[0097] wherein, is the second virtual droop control coefficient of the primary frequency regulation when the frequency drops or rises, is the lower limit value of the primary frequency regulation power change limit, is the upper limit value of the primary frequency regulation power change limit, is the working frequency change of the wind turbine, is the deviation value set for the primary frequency regulation frequency; is the given reference value of the active power of the wind turbine during the primary frequency regulation, is the active power value at the time when the primary frequency regulation frequency starts to change;

[0098] calculating the active power of the wind turbine participating in the frequency regulation:

[0099]

[0100] wherein, is the active regulation power value of the wind turbine participating in the frequency regulation.

[0101] In an optional embodiment, the calculation of the active regulation power of the wind storage system after adaptive frequency regulation specifically includes:

[0102] calculating the active regulation power of the wind turbine after adaptive frequency regulation :

[0103]

[0104]

[0105] wherein, denotes the number of wind turbines in the wind storage system; is the active regulation power of the virtual inertia and virtual droop control of the th wind turbine participating in frequency regulation;

[0106] the active regulation power of the wind storage system after adaptive frequency regulation is calculated as:

[0107]

[0108]

[0109] wherein, denotes the number of battery energy storage devices in the wind storage system; is the active regulation power of the virtual inertia and virtual droop control of the th battery energy storage device participating in frequency regulation;

[0110] the active regulation power of the wind storage system after adaptive frequency regulation is calculated as:

[0111] .

[0112] In an optional embodiment, the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system are calculated by integrating the optimal weight configuration, and the transfer function of the frequency response model of the energy storage system is calculated as:

[0113]

[0114] wherein, is the response time constant of the energy storage system, is the transformation parameter in Laplace transformation.

[0115] In an optional embodiment, the virtual inertia control of the wind turbine is rotor kinetic energy control, and the virtual droop control of the wind turbine is mainly pitch angle control. The transfer function of the frequency response model of the wind turbine is calculated by integrating the updated second virtual droop control coefficient and the updated second virtual inertia control coefficient:

[0116]

[0117] wherein, is the response time constant of the rotor inertia, is the response time constant of the pitch control.

[0118] The beneficial effects of the application are that, in view of the natural fluctuation of wind speed in an actual wind farm and the performance of the wind storage system in different operating states, an adaptive frequency modulation control strategy is adopted to ensure that the wind storage system optimally modulates frequency under the best conditions of system life, quantitatively analyzes the frequency modulation active power of the wind power and energy storage combined system, calculates the adjustable active power of the wind turbine and energy storage equipment participating in the frequency modulation strategy, evaluates the performance of the system, and further optimizes the frequency modulation strategy according to the evaluation results, so as to achieve the best frequency modulation effect and equipment life balance.

[0119] In addition, the design principle of the application is reliable, the structure is simple, and it has very wide application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0120] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, for those skilled in the art, other drawings can also be obtained based on these drawings without creative labor.

[0121] Figure 1 is a schematic flow chart of an active regulation power quantitative evaluation method of a wind storage system adaptive frequency modulation according to an embodiment of the application.

[0122] Figure 2 is a detailed flow chart of an active regulation power quantitative evaluation method of a wind storage system adaptive frequency modulation according to an embodiment of the application.

[0123] Figure 3 is a curve adaptive coefficient provided for the embodiment of the application and droop control coefficient variation.

[0124] Figure 4 is a maximum power tracking curve provided for the embodiment of the application.

[0125] Figure 5 is a load shedding operation principle provided for the embodiment of the application.

[0126] Figure 6 is a wind turbine frequency modulation overall strategy diagram provided for the embodiment of the application.

[0127] Figure 7 is an adaptive controller of energy storage participating in frequency modulation provided for the embodiment of the application.

[0128] Figure 8 is the overall framework of a virtual inertia and virtual droop control adaptive allocation model based on a DDPG algorithm provided for the embodiment of the application.

[0129] Figure 9A schematic structural block diagram of an active regulation power quantitative evaluation device of a wind storage system adaptive frequency modulation is an embodiment of the present application. DETAILED DESCRIPTION

[0130] In order for those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0131] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terminology used in the description of the present application herein is only for the purpose of describing the specific embodiments and is not intended to limit the present application.

[0132] Figure 1 A schematic flow chart of a method is an embodiment of the present application. Wherein, Figure 1 The execution subject can be an active regulation power quantitative evaluation device of a wind storage system adaptive frequency modulation. According to different needs, the order of steps in the flow chart can be changed, and some can be omitted.

[0133] As shown in Figure 1 and Figure 2 , the method comprises:

[0134] Step S1, obtaining the charge and discharge state and state of charge of the energy storage system, adjusting the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage participating in frequency modulation according to the charge and discharge state and the state of charge;

[0135] According to the power output characteristics of the wind turbine, the second virtual droop control coefficient and the second virtual inertia control coefficient of the wind turbine participating in frequency modulation are obtained;

[0136] Step S2, constructing a wind storage system adaptive allocation model, and based on the wind storage system adaptive allocation model, the first virtual droop control coefficient and the first virtual inertia control coefficient are respectively configured with optimal weights, and the corresponding first virtual droop control frequency modulation power and the first virtual inertia control frequency modulation power are calculated;

[0137] Step S3, updating the second virtual droop control coefficient based on the first virtual droop control frequency modulation power after the weight configuration and calculating the second virtual droop control frequency modulation power, updating the second virtual inertia control coefficient based on the first virtual inertia control frequency modulation power after the weight configuration and calculating the second virtual inertia control frequency modulation power;

[0138] Step S4: Calculate the active power regulation of the wind-storage system after adaptive frequency regulation based on the first virtual droop control frequency regulation power, the first virtual inertial control frequency regulation power, the second virtual droop control frequency regulation power, and the second virtual inertial control frequency regulation power.

[0139] Optionally, as an embodiment of the present invention, in step S1, the two basic control strategies for battery energy storage to participate in frequency regulation are virtual inertial control and virtual droop control. The expression for the energy storage frequency regulation control command obtained based on the system frequency change rate and frequency difference and through the above two control strategies is as shown in the following formula:

[0140]

[0141]

[0142]

[0143] in, This indicates the active power of energy storage participating in frequency regulation. This indicates that energy storage achieves active power frequency regulation control through virtual inertial control. This indicates that energy storage achieves active power frequency regulation control through virtual droop control. This represents the first virtual inertial control coefficient of the energy storage system. This represents the first virtual droop control coefficient of the energy storage system. The rate of change of frequency, This refers to frequency deviation. Under step disturbances in new energy power, since virtual inertial control is proportional to the rate of change of frequency deviation, the system frequency deviation change rate is usually large in the initial stage of the disturbance. Energy storage can quickly control this through virtual inertial control, hindering the change in the rate of change of frequency deviation. However, it has no effect on the steady-state frequency deviation of the system, and it will suppress frequency recovery when the direction of the rate of change of frequency deviation is inconsistent with the direction of frequency deviation. Virtual droop control is proportional to the system frequency deviation and has a certain delay in frequency adjustment, mainly adjusting the steady-state deviation of the system frequency. Considering that the two control strategies for current battery energy storage frequency regulation play different roles at different stages of frequency regulation, if the two can be appropriately combined during the frequency regulation process to make their advantages complementary, it can better suppress the frequency disturbances caused by wind turbine output fluctuations, further leverage the advantages of rapid frequency regulation of energy storage, and reduce the pressure on grid-side frequency regulation.

[0144] Energy storage control is closely related to its state of charge (SBC). Ignoring changes in SBC during system operation can lead to overcharging and over-discharging, affecting not only the storage's lifespan but also causing premature shutdown due to inherent limitations, resulting in greater harm to the system. To effectively utilize the charging and discharging characteristics of energy storage systems, the droop control coefficient, which participates in frequency regulation, is adjusted based on the SBC to maintain the energy storage's SBC within a healthy operating range. This prevents the energy storage system from operating at excessively high or low charge levels for extended periods, thus achieving efficient frequency regulation. The definition of the energy storage droop coefficient is as follows:

[0145]

[0146] in, This is the droop control coefficient during energy storage discharge; The droop control coefficient during energy storage charging. To ensure good frequency regulation performance of the energy storage, when the SOC of the energy storage is too high, Set to the maximum value. It decreases as SOC increases; when the SOC of the energy storage is relatively small, Set to the maximum value. The droop coefficient decreases as SOC decreases, therefore the S-function (sigmoid function) is chosen to simulate the change in the energy storage droop coefficient, and the specific relationship is shown below:

[0147]

[0148]

[0149] In the formula, This represents the maximum value of the droop control coefficient. The adaptive coefficient of the curve determines the trend of the curve's change. This is the maximum limit of the state of charge during charging; This is the minimum limit for the state of charge during discharge; This is the state of charge value. When The S-function curves change with different values. The value of has a significant impact on system performance, specifically as follows: Figure 3 As shown. The choice of value needs to balance the energy storage frequency regulation effect and Adaptability to changes in energy storage SOC. When When the value is large, The range of variation is small, and the adaptability is weak; while When the value is too low, it may affect the frequency regulation performance of energy storage. Therefore, selecting an appropriate value is important. The value is key. In practical applications, Choosing 15 ensures both the effectiveness of energy storage frequency regulation and takes into account... Adaptability with SOC change.

[0150] Optionally, as an embodiment of the present application, in step S2, the wind storage system adaptive allocation model is constructed based on a deep deterministic policy gradient algorithm, and an agent is trained under wind turbine output control disturbance to achieve optimal allocation of the first virtual droop control coefficient and the first virtual inertia control coefficient. According to the established regional frequency response model, the reinforcement learning agent needs to continuously learn to update the agent parameters under the wind farm output disturbance according to the historical experience, and the control of the wind farm is mainly obtained by the wind turbine model input by the wind speed. The distribution model commonly used to fit the wind speed at present is Weibull distribution, and its probability distribution is:

[0151]

[0152] wherein, is a size parameter, is a shape parameter, is the wind speed at the location of the wind turbine. In order to train the agent with historical control characteristics, it is necessary to first collect a large amount of historical wind speed data of the wind farm to calculate the wind speed probability statistical distribution parameters and obtain the statistical distribution law of the wind speed. Then a random sequence is generated according to the wind speed probability distribution, and finally the control fluctuation data can be obtained according to the indicated wind turbine model.

[0153] Optionally, as an embodiment of the present application, in step S3, the mechanical power captured by the wind turbine can be expressed as:

[0154]

[0155] wherein, is the tip speed ratio. Wherein, is the air density; is the wind energy utilization coefficient; is the pitch angle; is the rotor speed of the wind turbine; is the radius of the wind wheel; is the wind speed at the location of the wind turbine; is a calculation intermediate variable. The captured power of the wind turbine is mainly related to the wind speed, the speed and the size of the pitch angle. The operating point at which the wind turbine power output is maximum is called the maximum power point tracking (MPPT) point. With the change of wind speed, the speed corresponding to the maximum power output point changes, and the two form an MPPT working curve. The power extremum is obtained at the vertex of the quadratic curve :

[0156]

[0157] wherein, , , are linear expressions of the wind energy utilization coefficient with respect to the pitch angle, and are updated immediately after the pitch angle changes, is an MPPT curve fitting coefficient.

[0158] In the normal operation of the wind turbine, the maximum power tracking can be achieved through the converter control and the pitch angle control. The maximum power tracking curve is shown in FIG. 1, and the wind turbine output power reference instruction is shown in formula (1). Figure 4

[0159]

[0160] In the formula, is a wind turbine maximum power tracking curve proportion coefficient; , , are respectively the minimum electrical angular velocity of the wind turbine, the constant speed region electrical angular velocity, and the maximum electrical angular velocity; is the maximum output power of the wind turbine.

[0161] In the normal operation of the wind turbine, different operation regions are divided according to the wind speed. When the wind speed is low, the rotor speed of the wind turbine is also low, and the wind turbine is in an idling state and does not emit active power. With the gradual increase of the wind speed, the output power increases with the increase of the rotor speed, and the wind turbine operates in the maximum power tracking mode. In the constant speed region, the rotor speed reaches the maximum value, at this time, the wind speed change has little effect on the rotor speed of the wind turbine, but the power still continues to increase. With the further increase of the wind speed, the wind turbine operates in the constant power region, at this time, the pitch angle control needs to be introduced to limit the output power of the wind turbine. According to the operation characteristics of the wind turbine, when the frequency oscillation occurs in the power grid, the wind turbine still outputs power to the power grid according to the current speed, and cannot provide necessary power support for the power grid, so the kinetic energy reserved in the rotor needs to be provided to participate in the system frequency modulation.

[0162] Optionally, as an embodiment of the present application, the current main frequency modulation strategy of the wind turbine is divided into two categories: one is the rotor kinetic energy control for participating in the system frequency modulation by using its own inertia, and the other is the active reserve control for participating in the frequency modulation by reserving active power. The rotor kinetic energy control of the wind turbine is mainly the virtual inertia control, the inertia response function responds to the frequency change rate of the power grid, and the kinetic energy stored in the rotor can be adjusted to support the power for a short time, and the response speed is fast. The active reserve control is mainly the pitch angle control, which responds to the frequency deviation of the power grid through the virtual droop control, adjusts the pitch angle to support the power for a long time, and the response speed is relatively slow. As the most common frequency support strategy of the wind farm, the two methods have certain representative significance, and the frequency modulation capacity of the wind farm will be significantly improved under the joint action of the two methods.​​

[0163] Rotor kinetic energy control (inertia response control) adjusts the input / output of kinetic energy stored in the impeller to change the output power of the fan and suppress the rate of change of system frequency. Inertia reflects the characteristics of the unit using its rotor kinetic energy to suppress the rate of change of system frequency, and the inertia time constant The action characteristics are represented as follows:

[0164]

[0165] In the formula, is the inertia time constant of the fan; is the rotor kinetic energy of the wind turbine; is the rated power of the wind turbine; is the moment of inertia of the wind turbine; is the rated speed of the fan.

[0166] For a synchronous generator set, the unit operates at the rated speed in the steady state, and when the system power (torque) imbalance causes the frequency to drop, the unit will release the stored rotor kinetic energy to suppress the frequency fluctuation. In this process, the inertia response expression is as follows:

[0167]

[0168] In the formula, and are the mechanical torque and the electromagnetic torque; and are the mechanical power and the electromagnetic power.

[0169] By analogy with the synchronous generator set, if the doubly-fed wind turbine is to have inertia support capability, its output electromagnetic power can respond to the change in system frequency. When the rotor kinetic energy of the doubly-fed wind turbine is output in the form of electromagnetic power to respond to the frequency fluctuation, the inertia support power can be obtained as shown in the following formula:

[0170]

[0171]

[0172] In the formula, is the initial speed corresponds to the rotor kinetic energy; is the rotor speed of the unit after participating in frequency modulation corresponds to the rotor kinetic energy.

[0173] After simplification, the inertia response power output by the doubly-fed wind turbine when providing inertia support can be obtained as shown in the following formula:

[0174]

[0175] Active reserve control (pitch angle control) is to increase the pitch angle of wind turbine at a certain wind speed , to realize the power reserve of load reduction control. When the frequency drops due to the sudden increase of grid load, the wind turbine can reduce the pitch angle and increase the output power of the wind turbine. When the grid load suddenly decreases, the wind turbine can limit the power output by increasing the pitch angle to participate in frequency modulation and improve the stability of the system. The pitch control can realize the power reserve of load reduction at various wind speeds. With the development of pitch technology, the pitch angle adjustment delay is further reduced, so the pitch angle control is used to reserve power. The principle of pitch load reduction operation is shown in Figure 5 , wherein, , is the pitch angle of different sizes.

[0176] When the wind turbine adopts the active reserve strategy, the pitch angle of the wind turbine is increased to realize the power reserve of load reduction control, and the working operating point will be fixed on the load reduction curve set in advance. At this time, the relationship between the initial load reduction output power and the MPPT point output power is:

[0177]

[0178] In the formula, is the load reduction coefficient of the wind turbine, which is generally set in percentage form.

[0179] The frequency modulation strategy combining rotor kinetic energy control and active reserve control mainly considers the frequency support process under the inertia scale and the primary scale, and analyzes and evaluates the frequency modulation capacity of the wind farm. The rotor kinetic energy power component mainly acts on the inertia time scale in a short time after the disturbance occurs, and plays a short-time support role; and the reserve power component mainly acts on the primary time scale, and plays a long-time support role.

[0180] Optionally, as an embodiment of the present application, when an active power disturbance occurs in the power system, the wind farm will adjust its active power to support the frequency. The active power response of the wind power system can be divided into two components according to the source of energy: the rotor kinetic energy power component stored in the rotor rotational kinetic energy and the reserve power component stored in the wind energy.

[0181] When the inertia response function is triggered, the wind turbine system will adjust the active power change value of the unit according to the following formula:

[0182]

[0183]

[0184] wherein, is the rated frequency of the grid. is the rated active power of the wind turbine; is the second virtual inertia control coefficient of the wind turbine.

[0185] The active power increase during the wind turbine inertia response mainly comes from the kinetic energy stored in the wind turbine blades. Considering the load limit of the wind turbine and the safe operation of the wind turbine, the active power change value is limited to obtain The new expression of

[0186]

[0187] wherein, is the lower limit value of the inertia response power change limit, generally -10%; is the upper limit value of the inertia response power change limit, generally 10%. Therefore, the active power given value of the wind turbine during the inertia response is as follows:

[0188]

[0189] wherein, is the active power given reference value of the wind turbine during the inertia response, is the active power value at the time when the frequency of the primary frequency modulation starts to change.

[0190] The calculation of the active power given value of the primary frequency modulation is as follows: when is as follows:

[0191]

[0192] when is as follows:

[0193]

[0194] wherein, is the lower limit deviation value of the frequency of the primary frequency modulation, is the upper limit deviation value of the frequency of the primary frequency modulation.

[0195] During the primary frequency modulation process, the pitch system needs to be coordinated to open and close the pitch. The wind turbine gradually releases about 10% of the power, and at most limits 20% of the power. The active power change value of the primary frequency modulation needs to be limited, and the new expression is as follows:

[0196]

[0197] wherein, , are the second virtual droop control coefficient of the primary frequency modulation when the frequency is falling and rising, respectively; is the lower limit of the primary frequency modulation power change limit, generally 10%; is the upper limit of the primary frequency modulation power change limit, generally 20%; the active power given value of the wind turbine during the primary frequency modulation is shown as follows:

[0198]

[0199] wherein, is the reference value of the active power given value of the wind turbine during the primary frequency modulation, is the active power value at the time when the primary frequency modulation starts to change.

[0200]

[0201] wherein, is the active power value of the wind turbine inertia response and the primary frequency modulation participating frequency support.

[0202] The specific situation of the wind turbine participating in the frequency modulation is shown as follows: Figure 6

[0203] Optionally, as an embodiment of the present application, the total control of the energy storage when participating in the frequency modulation is composed of virtual inertia control and virtual droop control, and the weight of the two in different frequency modulation periods is adjusted by and and The specific values of and are obtained by the intelligent agent trained, Figure 7 are the scaling coefficients of the frequency difference and the frequency difference rate of change. When the battery energy storage participates in the frequency modulation, the system frequency difference enters the energy storage controller after the dead zone, at this time the frequency difference and the frequency difference rate of change pass through the proportional link to obtain the virtual inertia distribution coefficient through the intelligent agent, and the virtual droop distribution coefficient is calculated at the same time; then the virtual inertia and the virtual droop control are calculated according to the distribution coefficient and the corresponding frequency difference and the frequency difference rate of change; finally, the control of the two control strategies is added to obtain the final frequency modulation control instruction of the energy storage frequency modulation, and the specific principle is shown as follows: is the differential meaning.

[0204] ​​Optionally, as one embodiment of the present application, the deep deterministic policy gradient (DDPG) algorithm is a reinforcement learning algorithm suitable for solving continuous control problems, and its main framework is a policy-value network. The DDPG algorithm is used to solve the adaptive allocation model of virtual inertia and virtual droop control when the wind storage system participates in frequency modulation. Based on the output fluctuation data of the wind turbine, the agent interacts and learns in the frequency response model of the wind storage system, and the agent is continuously trained according to the obtained experience, so that the weight of the two control modes can be adaptively allocated when the energy storage participates in frequency modulation, and the frequency modulation effect and control optimization can be met.

[0205] wherein the state variable is defined as the system frequency difference and the frequency difference change rate after the battery energy storage dead zone link; the action of the agent is defined as the virtual inertia allocation coefficient when the battery energy storage participates in frequency modulation, satisfying The specific value of the virtual droop allocation coefficient can be calculated according to the relationship between the action of the agent and the virtual inertia allocation coefficient and the virtual droop allocation coefficient, and then the coefficient of the wind turbine participating in frequency modulation control is adaptively updated. The reward function mainly includes and two parts, wherein and are the proportion coefficients of the two parts of the reward, is the frequency difference of the system, is the power frequency modulation control change amount of the wind storage system, the reward measures the pros and cons of the action of the agent on the effect of the energy storage participating in frequency modulation, and the reward avoids the situation that the wind storage system long-term power participates in frequency modulation control and affects its service life.

[0206]

[0207] wherein, and are the proportion coefficients of the two parts of the reward, is the frequency difference of the system, is the power frequency modulation control change amount of the wind storage system, the reward measures the pros and cons of the action of the agent on the effect of the energy storage participating in frequency modulation, and the reward measures the situation that the wind storage system long-term power participates in frequency modulation control and affects its service life

[0208] The overall framework of the virtual inertia and virtual droop control adaptive allocation model based on the DDPG algorithm is shown in Figure 7 , wherein, and the transfer function of the frequency response of the wind turbine and the battery energy storage system, respectively, the system inertia of the wind storage system.

[0209] Optionally, as an embodiment of the present application, in step S4, the battery energy storage handles the frequency fluctuation by adopting virtual inertia response and virtual droop control, in combination with the analysis of the energy storage frequency modulation control strategy, the transfer function of the frequency model of the energy storage system is:

[0210]

[0211] wherein, is the response time constant of the energy storage system, is the transformation parameter in Laplace transform.

[0212] To achieve the optimal allocation of virtual inertia control and virtual droop control when the battery energy storage participates in frequency modulation, the principle of the proposed adaptive allocation method is shown in the formula:

[0213]

[0214]

[0215]

[0216] wherein, and and are the first virtual inertia control and the first virtual droop control frequency modulation power after optimal weight configuration, is the active power of the energy storage participating in frequency modulation after optimal weight configuration; wherein, . and are the virtual inertia allocation coefficient and the virtual droop allocation coefficient obtained by the proposed optimal allocation method, respectively.

[0217] In the process of joint frequency modulation of the wind storage system, the wind turbine can adjust the second virtual inertia control coefficient and the second virtual droop control coefficient when the wind turbine participates in frequency modulation through its power margin adjustment control, according to the adaptive allocation frequency modulation coefficient model of the energy storage, to assist the energy storage in joint frequency modulation of the system, and the adjustment strategy principle is shown in the following formula:

[0218]

[0219]

[0220] wherein, is the second virtual inertia control coefficient of the wind turbine; is the second droop control coefficient of the wind turbine; , This is an adjustable power constraint.

[0221] Optionally, as an embodiment of the present invention, in step S4,

[0222] When regulating the frequency of a wind turbine, the turbine speed can be adjusted, and the blade angle can be modified, thereby increasing control over rotor inertia and pitch angle. Therefore, wind turbine systems mainly include two frequency regulation methods: rotor inertia control and variable pitch control. Rotor inertia control technology is suitable for simulating the inertial response of a traditional generator, and its frequency model transfer function is:

[0223]

[0224] In the formula, The inertial response coefficient of the wind turbine unit; The rotor inertial response time constant; These are the transformation parameters in the Laplace transform.

[0225] The variable pitch control technology of wind turbines is suitable for simulating the primary frequency regulation of traditional generators, and its frequency model transfer function is:

[0226]

[0227] In the formula, This is the second virtual droop control coefficient for the primary frequency regulation of the wind turbine. is the variable pitch response time constant.

[0228] By combining inertial control and pitch control of the doubly-fed induction generator (DFIG) wind turbine, the wind farm acquires inertial response and primary frequency regulation capabilities similar to those of traditional generator sets. Its frequency model transfer function is:

[0229] .

[0230] Optionally, as an embodiment of the present invention, the DDPG algorithm as a whole comprises four neural networks: such as Figure 8 As shown, the strategy network and its target strategy network, and the value network and its target value network. The evaluation function represents the value network. Represents the policy function of the target policy network; Represents the network parameters of the target network; Network parameters representing the strategy or value network; Update the parameters for the target network. The training environment for the agent is the frequency response model of the wind-storage system built in Part 1. Before the start of each training round, a random wind speed disturbance is first generated based on the regional wind speed probability distribution, and then a continuous power disturbance is generated through the wind turbine model to simulate the frequency modulation scenario during training in the simulation environment. Assume that in the first round... The state of the system at the moment The action at this moment can be obtained through the policy network The virtual inertia distribution coefficient of the energy storage is:

[0231]

[0232] Wherein, is the output of the policy network; is the output noise of the policy network, which can be used to increase the exploration degree of the agent. The energy storage performs the action in the simulation environment The reward at the current moment And the state at the next moment The DDPG algorithm stores the Sequence generated by the interaction between the agent and the environment in the experience replay pool through the experience replay technology during training, and randomly extracts Group of historical data from it through batch sampling method to update the parameters of the policy network and the value network, and the parameters of the target network are updated through soft updating method in each round.

[0233] Optionally, as an embodiment of the present application, in step S5, the operating characteristics of the wind turbine and the energy storage system, and the dynamic interaction between them are analyzed. Based on the historical wind speed data of the wind farm, a wind speed fluctuation model is constructed to simulate the change of wind speed, and on the basis of dynamically adjusting the active power output of the system, an adaptive frequency modulation strategy is realized to realize the optimization of frequency modulation effect and wind storage system life, obtain the second virtual inertia control coefficient and the second virtual droop control coefficient of the wind turbine and the energy storage participating in frequency modulation. Finally, the efficiency of the wind storage combined system during frequency modulation is evaluated, and the active regulation power of each part of the wind storage system participating in frequency modulation is quantified.

[0234] When the load surge disturbance occurs, the frequency fluctuates. After the wind turbine is distributed by the adaptive controller, the frequency modulation active power is as follows:

[0235]

[0236]

[0237] Wherein, Indicates the number of wind turbines in the wind storage system; Is the active regulation power of the virtual inertia and virtual droop control of the Wind turbine participating in frequency modulation.

[0238] After the battery energy storage device is distributed by the adaptive controller, the frequency modulation active power is as follows.

[0239]

[0240]

[0241] wherein, represents the number of energy storage devices in the wind storage system; is the first virtual inertia control coefficient of the th battery energy storage device participating in frequency modulation.

[0242] Therefore, the active regulation power of the wind storage system adaptive frequency modulation is as follows:

[0243] .

[0244] In some embodiments, the active regulation power quantitative evaluation device of the wind storage system adaptive frequency modulation can include a plurality of functional modules composed of computer program segments. The computer programs of each program segment in the active regulation power quantitative evaluation device of the wind storage system adaptive frequency modulation can be stored in the memory of the computer device and executed by at least one processor to perform the functions of the active regulation power quantitative evaluation of the wind storage system adaptive frequency modulation (see Figure 1 Description) for details.

[0245] In this embodiment, the active regulation power quantitative evaluation device of the wind storage system adaptive frequency modulation can be divided into a plurality of functional modules according to the functions it performs, as shown in Figure 9 The functional modules of the device can include: an energy storage coefficient adjustment module, a wind turbine coefficient acquisition module, a coefficient optimal allocation module, and an active regulation power calculation module. The module referred to by the present application refers to a series of computer program segments that can be executed by at least one processor and can complete a fixed function, which is stored in the memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0246] The device includes:

[0247] The energy storage coefficient acquisition module acquires the charge and discharge state and state of charge of the energy storage system, and adjusts the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage participating in frequency modulation according to the charge and discharge state and the state of charge;

[0248] The wind turbine coefficient acquisition module acquires the second virtual droop control coefficient and the second virtual inertia control coefficient of the wind turbine participating in frequency modulation according to the power output characteristics of the wind turbine;

[0249] The energy storage coefficient weight configuration module constructs a wind storage system adaptive distribution model, and performs optimal weight configuration for the first virtual droop control coefficient and the first virtual inertia control coefficient based on the wind storage system adaptive distribution model, to calculate corresponding first virtual droop control frequency modulation power and first virtual inertia control frequency modulation power.

[0250] The fan coefficient updating module updates the second virtual droop control coefficient based on the first virtual droop control frequency modulation power after the weight configuration and calculates the second virtual droop control frequency modulation power, and updates the second virtual inertia control coefficient based on the first virtual inertia control frequency modulation power after the weight configuration and calculates the second virtual inertia control frequency modulation power.

[0251] The active regulation power calculation module calculates the active regulation power after adaptive frequency modulation of the wind storage system based on the first virtual droop control frequency modulation power, the first virtual inertia control frequency modulation power, the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power.

[0252] Through the cooperative operation of the modules, the corresponding coefficients can be obtained, adjusted and optimally distributed according to the relevant states of the energy storage system and the power output characteristics of the wind turbine, and the active regulation power after adaptive frequency modulation of the wind storage system can be accurately calculated, which helps to improve the accuracy and adaptability of the frequency modulation of the wind storage system.

[0253] Optionally, as an embodiment of the present application, the wind storage system adaptive distribution model is constructed based on a deep deterministic policy gradient algorithm, and an intelligent agent is trained under wind turbine output control disturbance to realize optimal distribution of the first virtual droop control coefficient and the first virtual inertia control coefficient, specifically including:

[0254] Initial parameters are set for the policy network, the target policy network, the value network and the target value network;

[0255] Simulated wind speed is generated according to the probability distribution of wind speed, and continuous disturbance data is generated through the wind turbine;

[0256] The intelligent agent determines the action based on the current system frequency deviation and the frequency deviation change rate, and the energy storage system executes the weight configuration of the corresponding first virtual droop control coefficient and virtual inertia control coefficient;

[0257] After the intelligent agent executes the action, the current reward and the next system state are fed back, and the experience data of the current system state, the action taken, the reward obtained and the next system state are stored in the replay buffer, and samples are extracted from the replay buffer to update the network parameters;

[0258] The above steps are repeated until the maximum number of steps in a single round is reached;

[0259] The iteration of the round is repeatedly performed until a preset maximum number of training rounds is reached, so as to obtain the optimal first virtual droop control coefficient and first virtual inertia control coefficient weight configuration of the energy storage system.

[0260] Optionally, as one embodiment of the present application, the obtained reward includes a reward and a reward :

[0261]

[0262] wherein, and is a proportionality coefficient of the two-part reward, is a frequency difference of the system, is a power frequency control change of the wind storage system, the reward measures the pros and cons of the action of the intelligent agent on the effect of the energy storage participating in frequency regulation, the reward measures the situation that the long-term power participation in frequency regulation control of the wind storage system affects its service life.

[0263] Optionally, as one embodiment of the present application, after obtaining the optimal weight configuration, the corresponding first virtual inertia control frequency regulation power and first virtual droop control frequency regulation power are calculated, which include:

[0264]

[0265]

[0266]

[0267] wherein, is the first virtual inertia control frequency regulation power after the optimal weight configuration, is the first virtual droop control frequency regulation power after the optimal weight configuration, is the active power of the energy storage participating in frequency regulation after the optimal weight configuration, , is the optimal first virtual inertia weight coefficient, is the optimal first virtual droop weight coefficient, represents the first virtual inertia control coefficient of the energy storage system, represents the first virtual droop control coefficient of the energy storage system, is a frequency change rate, is a frequency deviation.

[0268] Optionally, as one embodiment of the present application, the updated second virtual droop control coefficient and the updated second virtual inertia control coefficient specifically include:

[0269]

[0270]

[0271] wherein, and is an adjustable power constraint, is an MPPT point output power, is an initial load shedding output power, is a rotor kinetic energy inertia response release power, is an updated second virtual droop control coefficient, is an updated second virtual inertia control coefficient.

[0272] Optionally, as one embodiment of the present application, the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power are specifically calculated as:

[0273] calculating an active power change value of the wind turbine through virtual inertia control :

[0274]

[0275]

[0276]

[0277] wherein, is a rated power frequency of the power grid, is a rated active power of the wind turbine, is a second virtual inertia control coefficient of the wind turbine, is an active power change value of the wind turbine, is an inertia response power change limiting lower limit value, is an inertia response power change limiting upper limit value, is an active power given reference value of the wind turbine during inertia response, is an active power value at the time when the inertia response frequency modulation frequency starts to change;

[0278] calculating an active power change value of the wind turbine through virtual droop control :

[0279]

[0280]

[0281]

[0282] wherein, is a second virtual droop control coefficient of the primary frequency modulation when the frequency drops or when the frequency rises, is a lower limit value of the limit amplitude of the primary frequency regulation power variation, is an upper limit value of the limit amplitude of the primary frequency regulation power variation, is a working variation frequency of the wind turbine, is a deviation value set for the primary frequency, is a reference value of the active power given value of the wind turbine during the primary frequency regulation, is an active power value at the time when the primary frequency starts to vary,

[0283] calculating the active power of the wind turbine participating in the frequency regulation:

[0284]

[0285] wherein, is the active regulation power value of the wind turbine participating in the frequency regulation.

[0286] Optionally, as an embodiment of the present application, the calculating the active regulation power of the wind storage system after adaptive frequency regulation specifically comprises:

[0287] calculating the active regulation power of the wind turbine after adaptive frequency regulation :

[0288]

[0289]

[0290] wherein, represents the number of wind power generators in the wind storage system; is the active regulation power of the virtual inertia and virtual droop control of the th wind power generator participating in the frequency regulation;

[0291] calculating the active regulation power of the wind storage system after adaptive frequency regulation :

[0292]

[0293]

[0294] wherein, represents the number of energy storage devices in the wind storage system; is the active regulation power of the virtual inertia and virtual droop control of the th battery energy storage device participating in the frequency regulation;

[0295] calculating the active regulation power of the wind storage system after adaptive frequency regulation :

[0296] .

[0297] Optionally, as one embodiment of the present application, the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system configured by the integrated optimal weight are used to calculate the transfer function of the frequency response model of the energy storage system:

[0298]

[0299] wherein, is the response time constant of the energy storage system, is the transformation parameter in the Laplace transform.

[0300] Optionally, as one embodiment of the present application, the virtual inertia control of the wind turbine is the rotor kinetic energy control, and the virtual droop control of the wind turbine is mainly the pitch angle control. The updated second virtual droop control coefficient and the updated second virtual inertia control coefficient are used to calculate the transfer function of the frequency response model of the wind turbine:

[0301]

[0302] wherein, is the response time constant of the rotor inertia, is the response time constant of the pitch control

[0303] Those skilled in the art can clearly understand that the technology in the embodiments of the present application can be realized by means of software and necessary general hardware platform. Based on such understanding, the technical solutions in the embodiments of the present application can be embodied in the form of a software product, which is stored in a storage medium such as a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. Various storage media that can store program codes include a plurality of instructions for causing a computer terminal (which can be a personal computer, a server, or a second terminal, a network terminal, etc.) to execute all or part of the steps of the method described in the embodiments of the present application.

[0304] In the present specification, the same or similar parts among various embodiments can be referred to each other. Especially, for the terminal embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0305] In several embodiments provided by the present application, it should be understood that the disclosed system and method can be implemented in other manners. For example, the system embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There can be another division manner for the actual implementation. For example, a plurality of modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different modules can be indirect couplings or communication connections through some interfaces, and electrical, mechanical or other forms.

[0306] The modules illustrated as separated components can or can not be physically separated, and the components illustrated as modules can or can not be physical modules, i.e., can be located in one place, or can be distributed to a plurality of network modules. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment.

[0307] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0308] Although the present application has been described in detail by referring to the preferred embodiments thereof, it is to be understood that the present application is not limited to the embodiments described above. Rather, it should be appreciated that those skilled in the art, upon consideration of the disclosure, can make modifications and / or improvements to the embodiments of the application without deviating from the spirit and scope of the application. Any modifications and / or improvements made to the embodiments of the application by those skilled in the art are to be considered within the scope of the application.

Claims

1. A method for quantitatively evaluating the active regulation power of adaptive frequency modulation of a wind storage system, characterized in that, The method comprises the following steps: obtaining the charge-discharge state and state of charge of the energy storage system, and adjusting the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system participating in frequency modulation according to the charge-discharge state and the state of charge; obtaining the second virtual droop control coefficient and the second virtual inertia control coefficient of the wind turbine participating in frequency modulation according to the power output characteristics of the wind turbine; constructing a wind storage system adaptive allocation model, and configuring optimal weights for the first virtual droop control coefficient and the first virtual inertia control coefficient based on the wind storage system adaptive allocation model to calculate the corresponding first virtual droop control frequency modulation power and the first virtual inertia control frequency modulation power; updating the second virtual droop control coefficient based on the first virtual droop control frequency modulation power after the weight configuration and calculating the second virtual droop control frequency modulation power, and updating the second virtual inertia control coefficient based on the first virtual inertia control frequency modulation power after the weight configuration and calculating the second virtual inertia control frequency modulation power; calculating the active power adjustment power of the wind storage system after adaptive frequency modulation based on the first virtual droop control frequency modulation power, the first virtual inertia control frequency modulation power, the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power.

2. The method of claim 1, wherein the method further comprises: The wind storage system adaptive allocation model is constructed based on a deep deterministic policy gradient algorithm, a wind turbine wind speed fluctuation model is constructed based on a Weibull distribution, and an intelligent agent is trained under wind turbine wind speed fluctuation based on the deep deterministic policy gradient algorithm to achieve optimal allocation of the first virtual droop control coefficient and the first virtual inertia control coefficient, specifically including: setting initial parameters for a policy network, a target policy network, a value network and a target value network; generating simulated wind speed according to the Weibull distribution of wind speed, and generating continuous disturbance data through a wind turbine; the intelligent agent determines an action based on the current system frequency deviation and frequency deviation change rate, and the energy storage system executes the corresponding weight configuration of the first virtual droop control coefficient and the virtual inertia control coefficient; after the intelligent agent executes the action, the current reward and the system state at the next time are fed back, the experience data of the current system state, the action taken, the reward obtained and the next system state are stored in a replay buffer, and samples are extracted from the replay buffer to update network parameters; repeat the above steps until the maximum number of steps per round is reached; iteratively repeat the round until the maximum preset number of training rounds is reached to obtain the optimal first virtual droop control coefficient and first virtual inertia control coefficient weight configuration of the energy storage system.

3. The method of claim 2, wherein the method further comprises: Rewards obtained include and rewards : wherein, and is a proportionality coefficient of the two-part reward, is a frequency difference of the system, is a power frequency modulation control change of the wind storage system, and the reward measures the pros and cons of the effect of the agent's action on the frequency modulation of the energy storage, and the reward measures the situation that the long-term power participation in frequency modulation control of the wind storage system affects its service life.

4. The method of claim 2, wherein the method further comprises: After obtaining the optimal weight configuration, the corresponding first virtual inertia control frequency modulation power and first virtual droop control frequency modulation power are calculated, including: In the formula, is the first virtual inertia control frequency modulation power after the optimal weight configuration, is the first virtual droop control frequency modulation power after the optimal weight configuration, is the active power of energy storage participating in frequency modulation after the optimal weight configuration, , is the optimal first virtual inertia weight coefficient, is the optimal first virtual droop weight coefficient, represents the first virtual inertia control coefficient of the energy storage system, represents the first virtual droop control coefficient of the energy storage system, is the frequency change rate, is the frequency deviation.

5. The method of claim 4, wherein the method further comprises: The updated second virtual droop control coefficient and the updated second virtual inertia control coefficient specifically include: wherein, and is an adjustable power constraint, is an MPPT point output power, is an initial load shedding output power, is a rotor kinetic energy inertia response release power, is an updated second virtual droop control coefficient, is an updated second virtual inertia control coefficient.

6. The method of claim 5, wherein the method further comprises: The second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power are specifically calculated as: Computing an active power change value for a wind turbine through virtual inertia control : wherein, is the nominal power grid frequency, is the nominal active power of the wind turbine, is a second virtual inertia control coefficient of the wind turbine, is the active power change value of the wind turbine, is the lower limit value of the inertia response power change limiter, is the upper limit value of the inertia response power change limiter, is the given reference value of the active power of the wind turbine during the inertia response, is the active power value at the time when the frequency of the frequency modulation starts to change. Computing an active power variation value obtained by a virtual droop control of a wind turbine : wherein, is a second virtual droop control coefficient for primary frequency regulation at the time of frequency drop disturbance or at the time of frequency rise disturbance, is a lower limit value of the primary frequency regulation power change limit, is an upper limit value of the primary frequency regulation power change limit, is a frequency of operation change of the wind turbine, is a deviation value set for the primary frequency regulation; is a reference value of the wind turbine active power setpoint during the primary frequency regulation, is an active power value at the time when the primary frequency regulation starts to change. The active power of the wind turbine participating in frequency modulation is calculated as: wherein, is the active regulation power value for the wind turbine to participate in frequency regulation.

7. The method of claim 6, wherein the method further comprises: The active power adjustment power of the wind storage system after adaptive frequency modulation specifically includes: Computing the active regulation power of a wind turbine after adaptive frequency regulation : wherein, represents the number of wind turbines within the wind storage system; is the active regulation power of the th wind turbine participating in frequency regulation when the virtual inertia and virtual droop control is applied. Adaptive frequency modulation of a power storage system : wherein, represents the number of energy storage devices within the wind storage system; is the first active regulation power of the virtual inertia and virtual droop control of the nth battery energy storage device participating in frequency modulation; Active regulation power of the wind storage system after adaptive frequency modulation The calculation is: 。 8.The method of claim 6, wherein, The first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system are comprehensively configured with optimal weights, and a frequency response model transfer function of the energy storage system is calculated as: wherein is the response time constant of the energy storage system, is a transformation parameter in the Laplace transform. 9.The method of claim 8, wherein, The virtual inertia control of the wind turbine is rotor kinetic energy control, and the virtual droop control of the wind turbine is mainly pitch angle control. The updated second virtual droop control coefficient and the updated second virtual inertia control coefficient are combined to calculate the transfer function of the frequency response model of the wind turbine: wherein, is a rotor inertia response time constant, is a pitch response time constant.

10. An active regulation power quantification evaluation device for adaptive frequency modulation of a wind storage system, characterized in that, The method comprises the following steps: An energy storage coefficient acquisition module acquires the charge and discharge state and state of charge of the energy storage system, and adjusts the first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system participating in frequency modulation according to the charge and discharge state and the state of charge; A wind turbine coefficient acquisition module acquires the second virtual droop control coefficient and the second virtual inertia control coefficient of the wind turbine participating in frequency modulation according to the power output characteristics of the wind turbine; An energy storage coefficient weight configuration module constructs an adaptive allocation model of the wind storage system, and performs optimal weight configuration for the first virtual droop control coefficient and the first virtual inertia control coefficient based on the adaptive allocation model of the wind storage system, and calculates the corresponding first virtual droop control frequency modulation power and the first virtual inertia control frequency modulation power; A wind turbine coefficient update module updates the second virtual droop control coefficient based on the first virtual droop control frequency modulation power after the weight configuration, and calculates the second virtual droop control frequency modulation power, and updates the second virtual inertia control coefficient based on the first virtual inertia control frequency modulation power after the weight configuration, and calculates the second virtual inertia control frequency modulation power; An active regulation power calculation module calculates the active regulation power of the adaptive frequency modulation of the wind storage system based on the first virtual droop control frequency modulation power, the first virtual inertia control frequency modulation power, the second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power. 11.The wind storage system adaptive frequency modulation active regulation power quantification evaluation device according to claim 10, characterized in that, The adaptive allocation model of the wind storage system is constructed based on a deep deterministic policy gradient algorithm, a wind turbine wind speed fluctuation model is constructed based on a Weibull distribution, and an intelligent agent is trained based on the deep deterministic policy gradient algorithm under wind turbine wind speed fluctuation, so as to realize optimal allocation of the first virtual droop control coefficient and the first virtual inertia control coefficient, and specifically comprises: Initial parameters are set for a policy network, a target policy network, a value network and a target value network; Simulated wind speed is generated according to the Weibull distribution of wind speed, and continuous disturbance data is generated through a wind turbine; The intelligent agent decides an action based on the current system frequency difference and frequency difference change rate, and the energy storage system executes the weight configuration of the corresponding first virtual droop control coefficient and virtual inertia control coefficient; After the intelligent agent executes the action, the current reward and the system state at the next time are fed back, experience data of the current system state, the action taken, the reward obtained and the next system state are stored in a replay buffer, and samples are extracted from the replay buffer to update network parameters; The above steps are repeated until the maximum number of steps in a single round is reached; The round iteration is repeatedly performed until the maximum preset training round number is reached, and the optimal first virtual droop control coefficient and first virtual inertia control coefficient weight configuration of the energy storage system are obtained. 12.The wind storage system adaptive frequency modulation active regulation power quantification evaluation device according to claim 11, characterized in that, Rewards obtained include and rewards : wherein, and is a proportionality coefficient of the two-part reward, is a frequency difference of the system, is a power frequency modulation control change of the wind storage system, and the reward measures the pros and cons of the effect of the agent's action on the storage participating in frequency modulation, and the reward measures the situation that the long-term power participating in frequency modulation control of the wind storage system affects its service life. 13.The wind storage system adaptive frequency modulation active regulation power quantification evaluation device of claim 11, wherein, After the optimal weight configuration is obtained, the corresponding first virtual inertia control frequency modulation power and first virtual droop control frequency modulation power are calculated, which comprises: In the formula, is the first virtual inertia control frequency modulation power after the optimal weight configuration, is the first virtual droop control frequency modulation power after the optimal weight configuration, is the active power of energy storage participating in frequency modulation after the optimal weight configuration, , is the optimal first virtual inertia weight coefficient, is the optimal first virtual droop weight coefficient, represents the first virtual inertia control coefficient of the energy storage system, represents the first virtual droop control coefficient of the energy storage system, is the frequency change rate, is the frequency deviation.

14. The wind storage system adaptive frequency regulation power quantification evaluation device of claim 13, wherein, The updated second virtual droop control coefficient and the updated second virtual inertia control coefficient specifically comprise: wherein, and is an adjustable power constraint, is an MPPT point output power, is an initial load shedding output power, is a rotor kinetic energy inertia response release power, is an updated second virtual droop control coefficient, is an updated second virtual inertia control coefficient.

15. The wind storage system adaptive frequency regulation power quantification evaluation device of claim 14, wherein, The second virtual droop control frequency modulation power and the second virtual inertia control frequency modulation power are specifically calculated as: Computing an active power change value for a wind turbine through virtual inertia control : wherein, is the nominal power grid frequency, is the nominal active power of the wind turbine, is a second virtual inertia control coefficient of the wind turbine, is the active power change value of the wind turbine, is the lower limit value of the inertia response power change limiter, is the upper limit value of the inertia response power change limiter, is the given reference value of the active power of the wind turbine during the inertia response, is the active power value at the time when the frequency of the frequency modulation starts to change. Computing an active power change value for a wind turbine by virtual droop control : wherein, is a second virtual droop control coefficient for primary frequency regulation at the time of frequency drop disturbance or at the time of frequency rise disturbance, is a lower limit value of the primary frequency regulation power variation limit, is an upper limit value of the primary frequency regulation power variation limit, is a wind turbine operating change frequency, is a deviation value set for the primary frequency regulation; is a wind turbine active power setpoint reference value during the primary frequency regulation, is an active power value at the time when the primary frequency regulation starts to change. The active power of the wind turbine participating in frequency modulation is calculated as: wherein, is the active regulation power value for the wind turbine to participate in frequency regulation. 16.The wind storage system adaptive frequency modulation active regulation power quantification evaluation device of claim 15, wherein, The active power of the wind storage system after adaptive frequency modulation specifically includes: Computing the active regulation power of a wind turbine after adaptive frequency regulation : wherein, represents the number of wind turbines within the wind storage system; is the active regulation power of the th wind turbine participating in frequency regulation when virtual inertia and virtual droop control are applied. Adaptive frequency modulation of a power storage system : wherein, represents the number of energy storage devices within the wind storage system; is the first active regulation power of the virtual inertia and virtual droop control of the nth battery energy storage device participating in frequency modulation; Active regulation power of wind storage system after adaptive frequency modulation The calculation is: 。 17. The wind storage system adaptive frequency regulation power quantification evaluation device of claim 16, wherein, The first virtual droop control coefficient and the first virtual inertia control coefficient of the energy storage system configured with the optimal weight are integrated to calculate a transfer function of a frequency response model of the energy storage system: wherein is the response time constant of the energy storage system, is a transformation parameter in the Laplace transform. 18.The wind storage system adaptive frequency modulation active regulation power quantification evaluation device of claim 17, wherein, The virtual inertia control of the wind turbine is rotor kinetic energy control, and the virtual droop control of the wind turbine is mainly pitch angle control. The transfer function of the frequency response model of the wind turbine is calculated by integrating the updated second virtual droop control coefficient and the updated second virtual inertia control coefficient: wherein, is a rotor inertia response time constant, is a pitch response time constant.

Images