A low-frequency oscillation suppression method for a wind power / pumped storage combined grid-connected power system

Abstract

The application discloses a low-frequency oscillation suppression method of a wind power / pumped storage combined grid-connected power system, which is composed of four grid-connected operation modes considering different working conditions of a wind power plant and a pumped storage power station, and comprises the following steps: 1. wind power and pumped storage two kinds of units are added with oscillation suppression links to suppress low-frequency oscillation together; 2. based on a wind speed probability distribution model, steady-state operation points and linearization models representing different grid-connected operation modes are determined; 3. eigenvalue analysis is adopted to define a dangerous mode by using a damping ratio, and a dangerous degree evaluation index is established; 4. eigenvalue sensitivity is calculated to determine key parameters of oscillation suppression and the adjustment ability of the key parameters to the dangerous mode; and 5. an optimization model of the damping ratio of the dangerous mode is established, and an artificial bee colony algorithm is extended with sensitivity information and inequality constraints, a fitness function is improved, and control parameters are coordinated and determined. The application can define the dangerous degree of an oscillation mode and the suppression ability of control parameters to low-frequency oscillation, so that low-frequency oscillation under different grid-connected operation modes can be effectively suppressed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of low-frequency oscillation suppression of wind power / pumped storage combined grid-connected power system, and particularly relates to a low-frequency oscillation suppression method of wind power / pumped storage combined grid-connected power system. BACKGROUND

[0002] The high penetration of wind power in the power grid exacerbates the peak load crisis of the power system, while the pumped storage system can effectively smooth the volatility of wind power and meet the system load peak shaving demand. Based on the complementary characteristics of the two, a wind power / pumped storage combined power generation system is formed. However, the large-scale grid connection of wind turbines and pumped storage units changes the original power flow distribution and energy interaction of the power system, affects the small signal stability of the system, and exacerbates the low-frequency oscillation.

[0003] At present, the potential of pumped storage system is not fully tapped, and the coordinated operation of the two units in the wind power / pumped storage combined power generation system has not yet formed an industry consensus. In addition, the existing low-frequency oscillation suppression methods for wind power or pumped storage units are independent of each other and do not consider the coordinated suppression of low-frequency oscillation. At the same time, since wind turbines and pumped storage units are connected to the grid in combination, the danger level of the oscillation mode and the suppression ability of the control parameters to the oscillation have not been defined. If wind power or pumped storage units are used to suppress low-frequency oscillation alone, the best suppression effect cannot be achieved, and the operation state of the other unit may be adversely affected. SUMMARY

[0004] To solve the problems in the background art, the present application provides a low-frequency oscillation suppression method of wind power / pumped storage combined grid-connected power system, so as to define the danger level of each low-frequency oscillation mode and the suppression ability of the control parameters of the two units to the oscillation, and to realize the coordinated operation of wind turbines and pumped storage units in the system, so as to achieve the best suppression effect on low-frequency oscillation.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0006] The low-frequency oscillation suppression method of wind power / pumped storage combined grid-connected power system provided by the present application has the characteristics that it comprises the following steps:

[0007] Step 1: According to the different working conditions of the wind power / pumped storage combined grid-connected power system, four kinds of grid-connected operation modes are divided, including: wind turbine MPPT working condition and pumped storage unit water diversion power generation working condition, wind turbine MPPT working condition and pumped storage unit pumped storage working condition, wind turbine active standby working condition and pumped storage unit water diversion power generation working condition, and wind turbine active standby working condition and pumped storage unit pumped storage working condition, and are respectively marked as Wm&G, Wm&P, Wd&G and Wd&P.

[0008] The wind power and pumped storage units in the wind power / pumped storage combined grid-connected power system are both provided with oscillation suppression links, wherein a power oscillation damper (POD) is added to the wind power unit, and a power system stabilizer (PSS) is added to the pumped storage unit.

[0009] Step 2, a wind power / pumped storage combined grid-connected power system model under each grid-connected operation mode is established, and based on the operation condition characteristics of unit combination and the probability distribution model of different wind speed intervals, the steady state operating points, linearized models and system state matrices representing different grid-connected operation modes are determined; wherein the system state matrices under the four grid-connected operation modes of Wm&G, Wm&P, Wd&G and Wd&P are respectively denoted as A WmG , A WmP , A WdG and A WdP .

[0010] Step 3, the eigenvalue analysis is performed on the system state matrices under the four grid-connected operation modes respectively, the eigenvalues, damping ratios and oscillation modes under each grid-connected operation mode are obtained, and all dangerous modes of each grid-connected operation mode are screened according to the damping ratios, the danger degree of each dangerous mode under each grid-connected operation mode is evaluated by using the constructed danger coefficient, so as to determine the priority of adjustment of each dangerous mode;

[0011] Step 4, the eigenvalue sensitivity of the control parameters in the POD and the PSS is calculated, the adjustment ability of each control parameter to the eigenvalues and damping ratios corresponding to the dangerous modes is quantified, and the importance index of the control parameter to the dangerous mode is defined by the sensitivity of the relative damping ratio;

[0012] Step 5, an optimization model for adjusting the damping ratio of the dangerous mode is established, a penalty function is constructed by using inequality constraints and the importance index, and the fitness function in the artificial bee colony algorithm is improved by using the penalty function, so as to solve the optimization model, thereby obtaining the optimal control parameters, and the output of the wind power and pumped storage units is coordinated to suppress the oscillation, the system damping ratio is increased, and the best suppression of the system low-frequency oscillation is realized.

[0013] The low-frequency oscillation suppression method of the wind power / pumped storage combined grid-connected power system has the characteristics that the steady state operating points representing different grid-connected operation modes in step 2 are determined as follows:

[0014] Step 2.1: the wind speed probability distribution V MPPT of the wind turbine unit corresponding to the wind speed interval [V M_min , V M_max ] under the MPPT operating condition is obtained by using formula (1):

[0015]

[0016] In equation (1), V w Indicates wind speed, f p (·) represents the probability density function of the distribution model, V M_min V is the minimum permissible wind speed under the MPPT condition of the wind turbine. M_max F represents the maximum permissible wind speed under MPPT conditions for wind turbines. p (·) represents the cumulative distribution function of the distribution model;

[0017] Step 2.2: Use equation (2) to obtain the wind speed range [V] corresponding to the active standby condition of the wind turbine unit. d_min V d_max Wind speed probability distribution V del :

[0018]

[0019] In equation (2), V d_min V is the minimum permissible wind speed under active standby conditions for wind turbines. d_max This refers to the maximum permissible wind speed under active standby conditions for wind turbine units.

[0020] Steps 2-3: Based on the probability distribution model for different wind speed ranges, the system determines the corresponding wind speed value V under the Wm&G and Wm&P operating modes. MPPT The system takes wind speed values ​​V under Wd&G and Wd&P operating modes. del , respectively, represent all possible wind speed conditions under each grid-connected operation mode, and the steady-state operation point corresponding to different wind speed values ​​is determined through power flow calculation.

[0021] The degree of danger of each hazard mode in step 3 is assessed according to the following steps:

[0022] Step 3-1, Define ξ cr The critical damping ratio is defined, and the damping ratios less than ξ under each grid-connected operation mode are selected. cr The oscillation mode corresponding to the damping ratio is defined as a dangerous mode, and let any i-th dangerous mode be denoted as ξ. di ;

[0023] Step 3-2, calculate the i-th hazard mode ξ using equation (3). di Risk factor χ i :

[0024]

[0025] Step 3-3, for any i-th dangerous mode, if the dangerous coefficient corresponding to the i-1-th dangerous mode is less than the dangerous coefficient of the i-th dangerous mode, it indicates that the i-th dangerous mode has higher adjustment priority than the i-1-th dangerous mode.

[0026] The importance index in the step 4 is calculated as follows:

[0027] Step 4-1, the control parameter to be adjusted is denoted as a parameter vector α = [K POD T d1 T d2 K PSS T p1 T p2 ] T , and any j-th parameter in the parameter vector α is denoted as α j ; wherein K POD , T d1 , T d2 respectively represent the gain multiple, the first constant of lead-lag, and the second constant of lead-lag of the additional POD of the wind turbine generator set, K PSS , T p1 , T p2 respectively represent the gain multiple, the first constant of lead-lag, and the second constant of lead-lag of the additional PSS of the pumped storage unit;

[0028] Step 4-2, the j-th control parameter α j is calculated by using formula (4):

[0029]

[0030] In formula (4), α j0 is the initial value of the j-th control parameter; is the j-th relative control parameter, and

[0031] The step 5 comprises:

[0032] Step 5-1, a target function of the optimization model of the adjustment dangerous mode damping ratio is constructed by using formula (5):

[0033]

[0034] In formula (5), l is the number of dangerous modes existing in any grid-connected operation mode, is the corresponding weight of the i-th dangerous mode;

[0035] An inequality constraint of the optimization model is constructed by using formula (6):

[0036]

[0037] In equation (6), m represents the number of parameters to be determined;

[0038] Step 5-2, use equation (7) to obtain the j-th control parameter α j Corresponding to the i-th danger mode ξ di Penalty item e i (α j ):

[0039] e i (α j )=max{0,-z maxj (α j )}+max{0,-z minj (α j )},1≤i≤l (7)

[0040] The penalty term e is obtained using equation (8). i (α j The corresponding penalty coefficient R ij :

[0041]

[0042] Step 5-3: Add the penalty term to the objective function to obtain the augmented penalty function of the optimized model. The fitness function fit(α) of the improved artificial bee colony algorithm is used; let α be the feasible solution vector of any v-th group. v Using equation (9), the fitness function fit(α) corresponding to the feasible solution vector of the vth group is established. v ) and penalty function

[0043]

[0044] In equation (9), α vj Let α represent the feasible solution vector of the vth group. v The j-th feasible solution and its corresponding j-th control parameter; σ represents the penalty factor in the penalty function; e i (α vj ) represents the j-th feasible solution α vj For the i-th dangerous mode ξ di Penalties;

[0045] Step 5-4: The feasible solution vector α composed of the control parameters... v As a nectar source in the artificial bee colony algorithm, it utilizes the fitness function fit(α) v Artificial bee colony algorithm for penalty function The optimization model is solved to obtain the optimal control parameters.

[0046] The electronic device of the present application comprises a memory and a processor, and the memory is used to store a program supporting the processor to execute the low-frequency oscillation suppression method, and the processor is configured to execute the program stored in the memory.

[0047] The computer readable storage medium of the present application stores a computer program, and when the computer program is run by a processor, the steps of the low-frequency oscillation suppression method are executed.

[0048] Compared with the prior art, the beneficial effects of the present application are reflected in:

[0049] 1. The method of the present application is aimed at the low-frequency oscillation suppression problem of a wind power / pumped storage combined grid-connected power system, and compared with the prior art, four kinds of grid-connected operation modes composed of wind farms and pumped storage power stations under different working conditions are considered at the same time, and based on the probability distribution model of different wind speed intervals, the steady-state operating point and the linearized model representing different grid-connected operation modes are determined, which more accurately describes the small signal stability of the combined system under multiple grid-connected operation modes.

[0050] 2. In the present application, a power oscillation damper POD is designed in the wind turbine generator set, and a power system stabilizer PSS is added in the pumped storage unit, and according to the damping ratio and the eigenvalue sensitivity, the danger degree of each low-frequency oscillation mode is defined, thereby quantifying the suppression ability of the control parameters of the two kinds of units on the system oscillation; based on the suppression ability, the values of the control parameters are comprehensively changed, which can realize the coordinated adjustment of the control parameters, and ensure that the output of the two kinds of generating units can be coordinated in the application of the present application, so as to effectively suppress the low-frequency oscillation of the grid-connected system.

[0051] 3. According to the divided danger degree and the suppression ability of the wind power and pumped storage units on different oscillation modes, a penalty function extended ABC (Artificial Bee Colony) algorithm is established, the fitness function in the algorithm is improved, and the optimal solution is obtained as the optimal control parameters of the POD of the wind turbine generator set and the PSS of the pumped storage unit, thereby ensuring that the low-frequency oscillation of the grid-connected system has the best suppression effect under the coordinated operation of the two kinds of units. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 It is a working principle schematic diagram of the wind power / pumped storage combined system with an oscillation suppression link in the present application;

[0053] Figure 2 It is a conventional pumped storage unit internal structure and control model;

[0054] Figure 3This invention provides a hazard coefficient curve for evaluating the degree of hazard corresponding to various hazard modes;

[0055] Figure 4 This is a flowchart of the algorithm for suppressing low-frequency oscillations in the wind power / pumped storage combined unit in this invention;

[0056] Figure 5 A flowchart for extending the artificial bee colony algorithm to find the global optimal value of control parameters. Detailed Implementation

[0057] In this embodiment, the implementation process of a method for suppressing low-frequency oscillations in a wind power / pumped storage combined grid-connected power system is as follows: Figure 4 As shown in the diagram. In this embodiment, based on the structural characteristics of the wind power / pumped storage combined system, power oscillation suppression components POD and PSS are added to both the wind turbine and the pumped storage unit. A schematic diagram of this structural characteristic is shown below. Figure 1 As shown, the specific modeling of the pumped-storage unit is as follows: Figure 2 As shown. An optimization model is designed based on the system damping ratio and the severity of the low-frequency oscillations. The eigenvalue sensitivity is used to quantify the adjustment capability of wind power and pumped-storage unit control parameters for different hazard modes. A penalty function-extended ABC algorithm is constructed, and the fitness function in the algorithm is improved. The optimal solution is obtained, and the POD and PSS control parameters are tuned. The coordinated operation of wind power and pumped-storage units effectively suppresses low-frequency oscillations in the grid-connected system. The optimization process of the extended ABC algorithm is as follows: Figure 5 As shown.

[0058] Please see Figure 4 Specifically, the method for suppressing low-frequency oscillations in the wind power / pumped storage combined grid-connected power system includes the following steps:

[0059] Step 1: Based on the different operating conditions of the wind power / pumped storage combined grid-connected power system, four grid-connected operation modes are divided, including: wind turbine MPPT operation mode and pumped storage unit water diversion power generation operation mode, wind turbine MPPT operation mode and pumped storage unit pumped storage operation mode, wind turbine active standby operation mode and pumped storage unit water diversion power generation operation mode, and wind turbine active standby operation mode and pumped storage unit pumped storage operation mode, which are respectively denoted as Wm&G, Wm&P, Wd&G, and Wd&P;

[0060] In wind power / pumped storage combined grid-connected power systems, both wind power and pumped storage units are equipped with oscillation suppression devices. Specifically, wind turbine units are equipped with power oscillation suppressors (PODs), and pumped storage units are equipped with power system stabilizers (PSSs).

[0061] In step 1, based on different combinations of operating conditions under grid-connected operation of wind turbines and pumped storage units, four grid-connected operation modes are identified, and oscillation suppression mechanisms are added to both types of units, specifically including:

[0062] Step 1-1, the overall operation structure of the combined system is as shown in Figure 1 The wind turbine and the pumped storage unit are connected to the point of common coupling (PCC) through a transformer. Considering different operating conditions of the wind turbine and the pumped storage unit, four kinds of grid-connected operation modes are provided: wind turbine MPPT condition and pumped storage unit water diversion power generation condition, wind turbine MPTT condition and pumped storage unit water pumping and storage condition, wind turbine active standby condition and pumped storage unit water diversion power generation condition, and wind turbine active standby condition and pumped storage unit water pumping and storage condition, which are denoted as Wm&G, Wm&P, Wd&G and Wd&P respectively.

[0063] Step 1-2, the additional oscillation suppression links are connected to the wind turbine and the pumped storage unit respectively, and the structures of the two are basically the same. Due to the difference in application objects, they are denoted as POD and PSS respectively. Both controllers are based on the active power P pcc As input, the corresponding model transfer functions are formula (1) and formula (2):

[0064]

[0065]

[0066] In formula (1) and formula (2), K POD , T d1 , T d2 represent the gain multiple, the first constant of lead-lag, and the second constant of lead-lag of the additional POD of the wind turbine respectively, and K PSS , T p1 , T p2 represent the gain multiple, the first constant of lead-lag, and the first constant of lead-lag of the additional PSS of the pumped storage unit respectively.

[0067] Step 2, the wind power / pumped storage combined grid-connected power system model under each grid-connected operation mode is established, and based on the operating condition characteristics of the unit combination and the probability distribution model of different wind speed intervals, the steady-state operating point, the linearized model and the system state matrix representing different grid-connected operation modes are determined. The system state matrices under the four kinds of grid-connected operation modes are denoted as A WmG , A WmP , A WdG , A WdP .

[0068] Step 2-1, as shown in Figure 2 , this embodiment shows the internal specific modeling of a conventional pumped storage unit, which includes motor / generator unit, excitation system, speed regulator, water pump / turbine unit and water diversion system. The model equations of the speed regulator and the water pump / turbine unit under different operating conditions have great differences.

[0069] The mathematical model of the power system after the two units are connected to the grid is expressed by state equations and algebraic equations as formula (3) including the structures of formula (1) and formula (2):

[0070]

[0071] In formula (3), p represents a differential operator; x and y represent vector forms of state variables and algebraic variables respectively; f and g represent function forms of state equations and algebraic equations respectively; the subscripts SG, WT, PSU and sys represent a synchronous generator, a wind turbine generator, a pumped storage unit and a system network respectively. The specific equations corresponding to different grid-connected operation modes of the wind power / pumped storage combined power station are different, and the embodiment of the present application is applicable to all modes without discussing the model equations of a specific grid-connected operation mode.

[0072] Step 2-2: The steady-state operating point of the grid-connected system is related to the wind speed value, and the wind speed values representing the wind turbine in the MPPT working condition and the active standby working condition are determined based on the wind speed probability distribution model, and the specific steps include:

[0073] Step 2-2-1, the wind speed probability distribution V M_min of the wind turbine unit corresponding to the wind speed interval [V M_max , V MPPT ] in the MPPT working condition is obtained by using formula (4):

[0074]

[0075] In formula (4), V w represents the wind speed, f p (·) is the probability density function of the distribution model, V M_min is the minimum allowable wind speed in the MPPT working condition of the wind turbine, V M_max is the maximum allowable wind speed in the MPTT working condition of the wind turbine, and F p (·) is the cumulative distribution function of the distribution model;

[0076] Step 2-2-2: the wind speed probability distribution V d_min of the wind turbine unit corresponding to the wind speed interval [V d_max , V del ] in the active standby working condition is obtained by using formula (5):

[0077]

[0078] In formula (5), V d_min is the minimum allowable wind speed in the active standby working condition of the wind turbine, and V d_max is the maximum allowable wind speed in the active standby working condition of the wind turbine.

[0079] Step 2-2-3: The system wind speed value V under the Wm&G, Wm&P operation mode of step 1 MPPT , the wind speed value V under the Wd&G, Wd&P operation mode del , respectively represent all possible wind speed conditions under the corresponding grid-connected operation mode.

[0080] Step 2-3, based on the wind speed value obtained in step 2-2, the power flow distribution of the system is obtained as the steady-state operating point for small disturbance stability analysis, and formula (3) is linearized, and formula (6) is used to obtain the linearized model of the wind power / pumped storage combined power generation system:

[0081]

[0082] In formula (6), A, B, C, and D are coefficient matrices; A sys is the system state matrix; Δ represents the increment. A sys in formula (6) corresponds to the system state matrix of four different grid-connected operation modes Wm&G, Wm&P, Wd&G, and Wd&P, which are respectively denoted as A WmG , A WmP , A WdG , and A WdP .

[0083] Step 3, the eigenvalue analysis is performed on the system state matrix under four grid-connected operation modes respectively, the eigenvalue, damping ratio, and oscillation mode under each grid-connected operation mode are obtained, and according to the damping ratio, all dangerous modes of each grid-connected operation mode are screened out, the dangerous degree of each dangerous mode under each grid-connected operation mode is evaluated by using the constructed dangerous coefficient, so as to determine the priority of the regulation of each dangerous mode, including:

[0084] Step 3-1, according to the system state matrix A sys , the system eigenvalue, eigenvector, oscillation mode, damping ratio, and oscillation frequency information are solved by using formula (7):

[0085]

[0086] In formula (7), λ is the eigenvalue; σ and ω are the real part and imaginary part of the eigenvalue, respectively; ξ is the damping ratio, and f sys corresponds to the oscillation frequency of various oscillation modes of the system;

[0087] Step 3-2, according to the damping ratio and the upper and lower limits of the low-frequency oscillation mode frequency, the dangerous mode is screened out, the larger the damping ratio, the faster the oscillation decay, and for the mode with small damping ratio, the decay speed is slow, which is easy to cause sustained oscillation of the system, endangering the stability of the system, and the low-frequency oscillation mode with a damping ratio less than the critical value is a dangerous mode, that is:

[0088]

[0089] f min and f max are the upper and lower limits of the oscillation mode frequency; ξ di is the damping ratio of the i-th selected dangerous mode; ξ cr is the critical value of the damping ratio.

[0090] Step 3-3, using the dangerous coefficient analytical expression of formula (9) as the evaluation index of the dangerous degree corresponding to various dangerous modes, the dangerous coefficient curve formed is as shown in Figure 3 , and the dangerous coefficient χ di of the i-th selected dangerous mode ξ i is:

[0091]

[0092] Step 3-4, according to the dangerous degree evaluation index and the dangerous coefficient curve, for the i-th selected dangerous mode, if the dangerous coefficient corresponding to the i-1-th dangerous mode ξ di-1 is less than the dangerous coefficient of the i-th dangerous mode ξ di , then the dangerous degree of the i-th dangerous mode ξ di is higher than that of the i-1-th dangerous mode, and the adjustment priority is higher.

[0093] Step 4, calculate the eigenvalue sensitivity of the control parameters in the POD and PSS, quantify the adjustment ability of each control parameter to the eigenvalue and damping ratio corresponding to the dangerous mode, and define the importance index of the control parameter to the dangerous mode as the relative damping ratio sensitivity, including:

[0094] Step 4-1, the control parameters to be adjusted are denoted as a parameter vector α = [K POD T d1 T d2 K PSS T p1 T p2 ] T , and the j-th parameter in the parameter vector α is denoted as α j ; wherein K POD , T d1 , T d2 represent the gain multiple, the first constant of lead-lag, and the second constant of lead-lag of the additional POD of the wind turbine, respectively, and K PSS , T p1 , T p2 represent the gain multiple, the first constant of lead-lag, and the second constant of lead-lag of the additional PSS of the pumped storage unit, respectively.

[0095] Step 4-2, the eigenvalue sensitivity of the to-be-tuned control parameter in step 4-1 is calculated to quantify the adjustment capability of each control parameter on the damping ratio of each dangerous mode, and the calculation of the eigenvalue sensitivity and the damping ratio sensitivity is as follows:

[0096]

[0097] In formula (10), is the sensitivity of the a th eigenvalue to the j th control parameter, is the sensitivity of the i th dangerous mode damping ratio to the j th control parameter;

[0098] Step 4-3, the adjustment capability of all control parameters is normalized, and the importance index of the j th to-be-solved parameter to the i th dangerous mode is calculated by using formula (11):

[0099]

[0100] In formula (11), α j0 is the initial value of the control parameter, is the relative control parameter, and

[0101] Step 5, an optimization model for adjusting the damping ratio of the dangerous mode is established, a penalty function is constructed through inequality constraints and the importance index, and the fitness function in the artificial bee colony algorithm is improved by using the penalty function, and then the optimization model is solved, so as to obtain the optimal control parameter, and the output of the oscillation suppression of the wind power and the pumped storage unit is coordinated to increase the system damping ratio, so as to realize the best suppression effect on the system low-frequency oscillation.

[0102] Step 5-1, the control parameter is tuned by using an optimization algorithm, and the purpose of the parameter optimization is to improve the system damping ratio of the dangerous mode, so the optimization target shown in formula (12) is obtained:

[0103]

[0104] In formula (12), there are l optimization targets, and l is the number of dangerous modes existing in the system under the grid-connected operation mode; the dangerous degree of different dangerous modes is different, and the mode with higher dangerous degree corresponds to higher damping ratio adjustment priority;

[0105] Step 5-2, in combination with the dangerous coefficient in formula (9), the multi-objective optimization is converted into single-objective optimization through linear weighting, and the objective function of the optimization model for adjusting the damping ratio of the dangerous mode is constructed by using formula (13):

[0106]

[0107] In formula (13), l is the number of dangerous modes existing in any grid-connected operation mode, corresponding weight of the i th dangerous mode;

[0108] Step 5-3, taking the value range of the parameter to be set as a constraint condition:

[0109]

[0110] The inequality constraint in formula (14) can be constructed into the standard form of inequality constraint of the optimization model by using formula (15):

[0111]

[0112] In formula (15), m represents the number of parameters to be solved;

[0113] Step 5-4, solving the optimization model by extending the artificial bee colony algorithm to realize the setting of the control parameter, combining Figure 5 The main process of the extended ABC algorithm is as follows:

[0114] 1) Initialize the state variable, and the scout bee randomly generates NS groups of feasible solution vectors (i.e. control parameter vectors α), wherein the j th feasible solution in the v th vector is α vj The generation mode is shown in formula (16):

[0115] α vj = α vjmin + rand(0, 1) × (α vjmax - α vjmin ) (16)

[0116] In formula (16), rand(0, 1) represents a random number taken from [0, 1], α vjmax and α vjmin respectively represent the upper and lower limits of the j th feasible solution in the solution vector.

[0117] 2) Extend the artificial bee colony algorithm, establish a penalty function, and improve the fitness function;

[0118] Construct the penalty term by using the inequality constraint of formula (15) and the importance index of the control parameter to different dangerous modes, and get the penalty term e i (α j ) of the j th control parameter α i corresponding to the i th dangerous mode ξ j by using formula (17):

[0119] e maxj (α j ) = max{0, -z minj (α j )} + max{0, -z ij (α i )} (17)}, 1≤i≤l (17)

[0120] By each penalty term and its corresponding penalty coefficient R ij The adjustment ability of the reaction control parameters to the objective function, the importance index normalization processing of formula (11), and the penalty term e i (α j ) corresponding to the penalty coefficient R ij :

[0121]

[0122] The penalty term is added to the objective function to obtain the augmented penalty function of the optimization model The fitness function fit(α) of the improved artificial bee colony algorithm is obtained; let any vth feasible solution vector be α v , and the fitness function fit(α v ) corresponding to the vth feasible solution vector is established by formula (19)

[0123]

[0124] In formula (19), α vj represents the jth feasible solution in the vth feasible solution vector α v , and corresponds to the jth control parameter; σ represents the penalty factor in the penalty function; e i (α vj ) represents the feasible solution α vj The penalty term for the ith risk mode ξ di ;

[0125] 3) Random search in the known honey source field to generate new honey sources, and the search formula is shown in formula (20):

[0126] β vj = α vj + rand(-1, 1) × (α vj - α cj ) (20)

[0127] In formula (20), c≠v; rand(-1, 1) represents a random number taken from [-1, 1]; each new solution β vj searched is in the solution space, and finally a new feasible solution vector β v is composed.

[0128] The greedy criterion T s (·) is used to update a new feasible solution as shown in formula (21):

[0129]

[0130] If the original honey source is the best honey source of the current generation α v , then the new feasible solution vector obtained by T s is the better honey source, and is taken as the best honey source of the new generation α v ′.

[0131] 4) Observe the bees to select a group of honey sources (feasible solutions) from the NS group of honey sources as the final selection, according to the value of the fitness function of the honey source, and use the roulette wheel selection method to select the honey sources, and the selection probability of each group of honey sources is shown in formula (22):

[0132]

[0133] 5) The information of the selected new generation of honey sources is recorded; when the search times of a honey source around reach a certain upper limit but the honey source is still not updated, then the scout bee randomly generates an initial honey source according to formula (16) and performs search and update again.

[0134] 6) Determine whether the stopping criterion is met, if the stopping criterion is met, stop the calculation and output the information of the optimal solution, otherwise return to step 2) and continue to solve.

[0135] Step 5-5, after completing the solution of the optimal value of the optimization model by the extended ABC algorithm, the control parameters are updated, and the eigenvalues and damping ratio distribution of the system after parameter updating are judged again. If there is still a dangerous mode, return to step 2 to solve again with the new control parameters as the initial value; if there is no dangerous mode, it is considered that under the control parameters, the wind power and pumped storage units can effectively realize the best effect of collaborative suppression of low-frequency oscillation.

[0136] In this embodiment, an electronic device includes a memory for storing a program supporting a processor to execute the above low-frequency oscillation suppression method, and a processor configured to execute the program stored in the memory.

[0137] In this embodiment, a computer readable storage medium has a computer program stored thereon, and the computer program is executed by a processor to perform the steps of the above low-frequency oscillation suppression method.

Claims

1. A method for low frequency oscillation suppression of a wind power / pumped storage combined grid-connected power system, characterized in that, The method comprises the following steps: Step 1, four grid-connected operation modes are divided according to different working condition combinations of the wind power / pumped storage combined grid-connected power system, including: wind turbine MPPT working condition and pumped storage unit water diversion power generation working condition, wind turbine MPPT working condition and pumped storage unit pumped storage working condition, wind turbine active standby working condition and pumped storage unit water diversion power generation working condition, and wind turbine active standby working condition and pumped storage unit pumped storage working condition, and are respectively denoted as Wm&G, Wm&P, Wd&G and Wd&P; In the wind power / pumped storage combined grid-connected power system, an oscillation suppression link is additionally added to the wind power and pumped storage units, wherein a power oscillation damper (POD) is added to the wind turbine, and a power system stabilizer (PSS) is added to the pumped storage unit; Step 2, establish the wind power / pumped storage combined grid-connected power system model under each grid-connected operation mode, and determine the steady-state operating point, linear model and system state matrix representing different grid-connected operation modes based on the operating condition characteristics of unit commitment and the probability distribution model of different wind speed intervals; wherein the system state matrices under the four grid-connected operation modes of Wm&G, Wm&P, Wd&G and Wd&P are respectively denoted as A WmG , A WmP , A WdG , A WdP ; Step 3, eigenvalue analysis is performed on the system state matrix under the four grid-connected operation modes, to obtain the eigenvalue, damping ratio and oscillation mode under each grid-connected operation mode, and all dangerous modes of each grid-connected operation mode are screened according to the damping ratio, the dangerous degree of each dangerous mode under each grid-connected operation mode is evaluated by using the constructed dangerous coefficient, and thus the priority of adjustment of each dangerous mode is determined; Step 3-1, definition is the critical damping ratio, and the oscillation mode corresponding to the damping ratio less than in each damping ratio under each grid-connected operation mode is defined as the dangerous mode, and let the arbitrary ith dangerous mode be ; Step 3-2, calculate the risk factor of the i-th risk mode using formula (3) :​ (3) Step 3-3, for any ith dangerous mode, if the dangerous coefficient corresponding to the (i-1)th dangerous mode is less than the dangerous coefficient of the ith dangerous mode, it is indicated that the adjustment priority of the ith dangerous mode is higher than that of the (i-1)th dangerous mode; Step 4, the eigenvalue sensitivity of the control parameters in the POD and PSS is calculated, the adjustment capability of each control parameter on the eigenvalue and damping ratio corresponding to the dangerous mode is quantified, and the importance index of the control parameter on the dangerous mode is defined by the sensitivity of the relative damping ratio; Step 5, an optimization model for adjusting the damping ratio of the dangerous mode is established, a penalty function is constructed by using inequality constraints and the importance index, the fitness function in the artificial bee colony algorithm is improved by using the penalty function, and then the optimization model is solved, to obtain the optimal control parameters, and the output of the oscillation suppression of the wind power and pumped storage units is coordinated, to increase the system damping ratio, and thus the best suppression of the system low-frequency oscillation is realized.

2. The low-frequency oscillation suppression method for a wind power / pumped storage combined grid-connected power system according to claim 1, characterized in that, The steady-state operating points representing different grid-connected operation modes in step 2 are determined in the following manner: Step 2.1: Obtain the wind speed probability distribution V MPPT of the wind turbine unit corresponding to the wind speed interval [V M_max ] at the MPPT operating condition by using formula (1) V MPPT = V M_max + (V M_max - V M_min ) * (1 - e - (V M_max - V M_min ) / (V M_max - V (1) In formula (1), V w represents the wind speed, f p (·) is the probability density function of the distribution model, is the minimum allowable wind speed under the MPPT working condition of the wind turbine generator, is the maximum allowable wind speed under the MPPT working condition of the wind turbine generator, F p (·) is the cumulative distribution function of the distribution model; Step 2.2: Obtain the wind speed probability distribution V del of the wind turbine unit corresponding to the wind speed interval [V d_min , V d_max ] in the active standby working condition by using formula (2) del : (2) In formula (2), V d_min is the minimum allowable wind speed under the active standby operating mode of the wind turbine generator, V d_max is the maximum allowable wind speed under the active standby operating mode of the wind turbine generator; Step 2-3, based on the probability distribution model of different wind speed intervals, the system takes the value V of the corresponding wind speed under the Wm&G and Wm&P operation modes MPPT , the system takes the value V of the corresponding wind speed under the Wd&G and Wd&P operation modes del , respectively representing all possible wind speed conditions under each grid-connected operation mode, and determining the steady-state operating point of the operation mode corresponding to different wind speed values through power flow calculation.

3. The method according to claim 2, wherein the method is characterized by, The importance index in step 4 is calculated in the following steps: Step 4-1, the control parameters to be adjusted are denoted as a parameter vector , and any jth parameter in the parameter vector is denoted as ; wherein, K POD , T d1 , T d2 respectively represent a gain multiple of a POD added to a wind turbine generator, a first constant of lead-lag, and a second constant of lead-lag, K PSS , T p1 , T p2 respectively represent a gain multiple of a PSS added to a pumped storage unit, a first constant of lead-lag, and a first constant of lead-lag; Step 4-2, calculating the jthcontrol parameter using formula (4) importance indicator for the ithhazard mode : (4) In formula (4), is an initial value of the jth control parameter; is the jth relative control parameter, and .

4. The low-frequency oscillation suppression method for the wind power / pumped storage combined grid-connected power system according to claim 1, characterized in that: Step 5 comprises: Step 5-1, the objective function of the optimization model for adjusting the damping ratio of the dangerous mode is constructed by using formula (5): (5) In formula (5), l is the number of dangerous modes in any grid-connected operation mode, is the corresponding weight of the i-th dangerous mode; The inequality constraint of the optimization model is constructed by using formula (6): (6) In formula (6), m represents the number of to-be-solved parameters; Step 5-2, the jthcontrol parameter is obtained using formula (7) corresponding to the ithdanger mode penalty term : (7) The penalty term is obtained using equation (8) The corresponding penalty coefficient : (8) Step 5-3, a penalty term is added to the objective function to obtain the augmented penalty function of the optimization model to improve the fitness function of the artificial bee colony algorithm Let any v-th group of feasible solution vectors be The fitness function corresponding to the v-th group of feasible solution vectors is established by using formula (9) and the penalty function : (9) In formula (9), denotes the v-th feasible solution vector denotes the j-th feasible solution in the set of feasible solutions and corresponds to the j-th control parameter; denotes the penalty factor in the penalty function; denotes the j-th feasible solution for the i-th dangerous mode penalty term; Step 5-4, the feasible solution vector of control parameters is composed As the artificial bee colony algorithm honey, thus using the fitness function based artificial bee colony algorithm for solving the penalty function based optimization model, and get the optimal control parameters.

5. An electronic device comprising a memory and a processor, characterized in that The memory is used for storing a program supporting the processor to execute the low-frequency oscillation suppression method in any one of claims 1-4, and the processor is configured to execute the program stored in the memory.

6. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to perform the steps of the low-frequency oscillation suppression method in any one of claims 1-4.

Images