Optimal proportion configuration method for network tracking / constructing type parallel system based on GRA-VIKOR method

Abstract

The invention discloses a GRA-VIKOR method-based optimal proportion configuration method for a network tracking / constructing type parallel system, and the method comprises the steps: carrying out the standardization processing of the original data of a new energy station under different network constructing type converter capacity proportions, and obtaining a standardized data matrix; determining subjective weights and objective weights of the evaluation indexes, and determining a final combination weight according to a game combination weighting method; determining a weighted evaluation matrix, and obtaining positive and negative ideal solutions of each evaluation index of the grid-connected performance of the network construction type control station; through the GRA thought, the grey correlation degree between each evaluation index value and the positive and negative ideal solutions is obtained; obtaining a group utility value and an individual regret value to determine a final group utility value and an individual regret value, and finally obtaining a comprehensive evaluation index of the new energy station under different capacity proportions of the network construction type converters; and finally determining the optimal capacity ratio of the network-forming converter. According to the method, the proportion problem between the following network type converter and the construction network type converter in the new energy station can be accurately measured.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid control, and in particular relates to an optimal proportion configuration method for a grid-connected / grid-forming parallel system based on a GRA-VIKOR method. Background Art

[0002] As the proportion of new energy sources connected to the power grid gradually increases, power electronic converters are increasingly used in power systems, gradually replacing a large number of traditional synchronous generators. However, this change has brought significant challenges to the stability of the power system. The reduction in the number of traditional generators has resulted in them being insufficient to continue to maintain the stable operation of the power system, which has brought great challenges to the stable operation of the system.

[0003] At present, most new energy power generation units adopt the grid following (GFL) strategy, which follows the frequency and phase angle of the AC grid voltage through a phase locked loop (PLL) to achieve synchronization with the grid. However, GFL control is unstable in a weak grid environment and does not have the ability to actively support the grid. In contrast, the new energy power generation unit based on grid forming (GFM) can achieve autonomous grid synchronization without relying on PLL, which can solve the problem of stable operation of traditional power generation units under weak grids, and has the ability to support grid inertia, frequency and voltage. However, the stability of the grid forming converter is poor under a strong grid, which is prone to cause low-frequency oscillation problems similar to traditional synchronous generator sets.

[0004] Although the new energy converters in the power system are currently mainly operated by grid-following control, with the increase in the number of grid-building converters, it is becoming increasingly important to analyze the interaction between grid-building converters and PLL-based grid-following converters. At present, domestic and foreign scholars have conducted a lot of research on determining the proportion of grid-building converters. Some scholars have analyzed the mechanism of GFM control to improve subsynchronous oscillations, pointing out that GFM control becomes a "positive resistor" under reactive loop control, which can offset part of the negative damping effect caused by GFL control. Some scholars have also established the relationship between the number of GFM devices and the short-circuit ratio from the perspective of short-circuit ratio, but have not solved the problem of the optimal proportion of GFM devices in the system. Some scholars have also studied the impact of the proportion of GFM devices on the voltage and frequency of the power system, as well as the impact on the power exchange between the distribution network and the transmission network, but only gave the proportion range of GFM devices, and did not accurately determine the specific value.

[0005] In summary, although the number of grid-connected converters in the current system is gradually increasing, the ratio between GFL and grid-connected converter equipment in new energy stations still needs further research. Summary of the invention

[0006] In view of the shortcomings of existing research, the purpose of the present invention is to provide an optimal proportion configuration method for a grid-following / grid-forming parallel system based on the GRA-VIKOR method. Through the grid-following / grid-forming transient stability analysis and the GRA-VIKOR method, two aspects of stability and supporting capacity test evaluation indicators are selected to accurately measure the ratio between grid-following converters and grid-forming converters in new energy stations.

[0007] In order to achieve the above technical objectives, the present invention adopts the following technical scheme: a method for configuring the optimal proportion of a grid-type parallel system based on the GRA-VIKOR method, comprising the following steps:

[0008] Step 1: The total capacity of the hybrid new energy station converter is P W , where the capacity of the grid-following and grid-forming converters are P1 and P2 respectively, then P W =P1+P2; Output power P of grid-connected converter ref Less than its maximum output active power limit P max Under the premise of increasing the capacity of grid-type converters, the original data of new energy stations under different grid-type converter capacity ratios are standardized to eliminate the impact of dimensions and obtain the normalized data matrix Y m×n , where m is the evaluation object and n is the evaluation index;

[0009] Step 2: Combine the Bayesian BWM method to obtain the subjective weight w1 of the evaluation index, and the critical method (CRITIC method) to determine the objective weight w2 of the evaluation index, and determine the final combination weight w3 based on the game combination weighting method;

[0010] Step 3: Determine the weighted evaluation matrix and obtain the positive and negative ideal solutions of each evaluation index of the grid-connected performance of the grid-connected control station; obtain the grey correlation between each evaluation index value and the positive and negative ideal solutions through the GRA idea; obtain the group utility value and individual regret value to determine the final group utility value Q i and individual regret value R i Finally, the comprehensive evaluation index T of the new energy station under different grid-type converter capacity ratios is obtained. i ;

[0011] Step 4: Determine the optimal grid-type converter capacity ratio: From step 3, we can get the T under each grid-type converter capacity ratio. i Assuming the best proportion is P1, the second best proportion is P2, and according to the maximum group utility value Q i The capacity proportions of different grid-type converters are ranked, and the best grid-type converter capacity proportion scheme is selected as the best solution when it meets two conditions.

[0012] More preferably, the evaluation indicators include:

[0013] Stability evaluation indicators: angular frequency change rate, transient angular frequency deviation

[0014] Support capacity evaluation indicators: damping active power and inertia active power.

[0015] More preferably, the step 1 includes the following sub-steps:

[0016] Step S101: Conduct simulation experiments on parallel systems with different grid-type converter capacity ratios to collect data:

[0017] Step S102: The grid-connected performance evaluation indicators of the grid-forming stations are homogenized and standardized.

[0018] More preferably, the step 2 includes the following sub-steps:

[0019] Step S201: Conduct simulation experiments on the grid-connected / grid-connected parallel systems under different grid-connected converter capacity ratios to collect data:

[0020] Step S202: Calculate the evaluation indicators of stability and support capability in the experimental scenarios with different proportions of grid-type converters, and then calculate the subjective weights of the proposed evaluation indicators according to the Bayesian method:

[0021] Step S203: Calculate the objective weight of the proposed evaluation index according to the CRITIC method:

[0022] Step S204: Use the game weighting method to determine the comprehensive weight.

[0023] Further preferably, the step three includes the following sub-steps:

[0024] Step S301: construct a weighted evaluation matrix and determine positive and negative ideal solutions under different grid-type converter proportion experimental scenarios;

[0025] Step S302: Determine the grey correlation between each evaluation index value and the positive and negative ideal solutions, and the final group utility value Q i and individual regret value R i ;

[0026] Step S303: Obtain the comprehensive evaluation index (VIKOR value) of the new energy station under different grid-type converter capacity ratios T i .

[0027] The present invention has the following advantages:

[0028] 1) The evaluation method for determining the proportion of different grid-type converters based on the GRA-VIKOR method proposed in the present invention quantifies the stability and dynamic support capability of the grid-type control station to the following grid-type control station; stability and support capability are used as two aspects of test evaluation indicators to accurately measure the ratio between following grid-type and grid-type converters in new energy stations.

[0029] 2) Based on the electromagnetic transient model of the grid-type parallel system, the present invention analyzes the influence mechanism of the grid-type converter on the transient stability of the grid-type converter, and obtains the following results for the grid-type converter (GFL) under pure active current: GFM It is the key variable that affects its transient stability.

[0030] 3) This method can be used to more conveniently determine the proportion of grid-connected converters.

[0031] 4) The present invention tests and evaluates different grid-type converter proportions in experimental scenarios, and conducts in-depth research on its characteristic evaluation indicators and performance experimental schemes. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a topological structure diagram of the parallel system with network / network.

[0033] Figure 2 It is the equivalent circuit of the parallel system.

[0034] Figure 3 It is a transient analysis model of the system in parallel with the grid.

[0035] Figure 4 It is the phase plane diagram of GFL under different GFM inertia J.

[0036] Figure 5 It is the phase plane diagram of GFL under different GFM inertia D.

[0037] Figure 6 It is a grid-connected system of grid-connected converter.

[0038] Figure 7 This is a flow chart of the optimal ratio configuration method for the grid-type parallel system based on the GRA-VIKOR method. DETAILED DESCRIPTION

[0039] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0040] In this embodiment, first, an electromagnetic transient model of a grid-type parallel system is constructed, and the influence mechanism of the grid-type active loop control parameters (J, D) on the transient stability of the grid-type parallel system is analyzed, including the following steps:

[0041] Step 1-1: Follow the grid / build the grid parallel system topology and mathematical model.

[0042] The topology of the parallel network system is as follows: Figure 1 As shown in Figure 2, the two converters are connected to the point of common coupling (PCC) through LC filters and transmission line impedance, and then connected to the grid through transmission lines. Figure 1 Medium: i abc1,2 are the currents in the filter inductor of the grid-following converter and the grid-forming converter respectively; u abc1,2 are the output voltages of the grid-following converter and the grid-building converter respectively; L f1 , C f1 , L f2 , C f2 are the filter inductors and capacitors of the grid-following converter and grid-forming converter respectively; L1, R1, L2, and R2 are the equivalent line inductors and equivalent line resistors of the grid-following converter and grid-forming converter connected to the PCC respectively; U dc is the constant voltage source on the DC side; U pcc is the voltage at the PCC point; R g and L g are the resistance and inductance connected to the grid respectively.

[0043] Since the grid-following converter behaves as a parallel high-impedance controllable current source, the grid-building converter behaves as a series low-impedance controllable voltage source. Figure 1 The topological structure diagram is equivalent to Figure 2 The equivalent circuit model shown. Figure 2 In the grid voltage vector U g The phase angle of I is taken as the reference phase angle, GFL and I GFM are the output current vectors of the grid-following converter and the grid-forming converter respectively; U GFL and U GFM are the output voltage vectors of the grid-following converter and the grid-forming converter, δ GFL and δ GFM They are the grid-following converter output voltage vector U GFL The output voltage vector U of the grid-connected converter GFM With the grid voltage vector V g The phase angle difference. pcc is the voltage vector U at PCC point PCC With the grid voltage vector V g The phase angle difference; is the grid-following converter output voltage vector U GFL With the output current vector I GFL The phase angle difference is Since the inductance in the circuit is large, the resistance elements in the circuit are ignored. It is worth noting that ignoring the resistance has no substantial effect on the analysis results. Therefore, the following will focus on the transient stability analysis of the parallel system with the grid under the pure inductive network.

[0044] X2 and X g are the line impedances between nodes 2 and 1 and between nodes 1 and 3, respectively, and X1 is the impedance between the grid-following converter and node 1. According to the node voltage equation, the circuit equation of the grid-following / grid-forming parallel system is:

[0045]

[0046] In the formula, is the output current vector of the grid-following converter; is the output current vector of the grid-type converter; is the grid current vector; is the grid voltage vector; is the PCC voltage vector; is the grid-type converter output voltage vector, the injected current of node 1 is The injected current at node 2 j is the imaginary unit.

[0047] According to formula (1), we can get U pcc ∠δ pcc satisfy:

[0048]

[0049] Among them, ω N is the rated angular frequency, is the grid-following converter output voltage vector U GFL With the output current vector I GFL The phase angle difference, δ GFL and δ GFM They are the grid-following converter output voltage vector U GFL The output voltage vector U of the grid-connected converter GFM With the grid voltage vector V g The phase angle difference;

[0050] Then, the expressions of the voltage amplitude and phase angle at the grid-following converter terminal can be obtained as follows:

[0051]

[0052] To simplify the analysis, the following definitions are made:

[0053]

[0054] Among them, k0, k1, k2, and k3 are all defined to simplify the analysis and are still impedances.

[0055] Combined with formula (4), formula (3) can be rewritten as:

[0056]

[0057] Taking the PLL phase angle as the reference coordinate system for coordinate transformation, the q-axis component of the grid-connected converter output voltage can be obtained from equation (5):

[0058] U GFLq =k1i GFLd -k3U g sinδ GFL +k2U GFM sin(δ GFM -δ GFL ) (6)

[0059] In the formula, i GFLd is the d-axis component of the output current of the grid-following converter, usually δ GFM ,δ GFL ∈[0,π / 2]. According to formula (6), the last term of the formula can be understood as the coupling term of the grid-connected converter connected to the renewable energy station in the GFL transient model, which is defined as the coupling impedance voltage drop term. This coupling term is related to k2 and U GFM , δ GFM and δ GFL related.

[0060] Substituting equation (6) into the phase-locked loop expression of the grid-following converter, the quasi-synchronous machine expression can be obtained as follows:

[0061]

[0062] J eq Defined as the equivalent inertia coefficient; δ GFL is the equivalent power angle of the grid-type converter, dt 2 is the second derivative with respect to time, Δω GFL is the difference between the output angular frequency of the grid-following converter and the rated angular frequency, δ GFM is the output phase angle of the grid-connected converter, ω GFL is the output angular frequency of the grid-following converter, ω GFM is the output angular frequency of the grid-type converter, i GFLd is the d-axis component of the grid-following converter output current, k i_pll is the phase-locked loop integral coefficient, k p_pll is the phase-locked loop proportional coefficient; D eq1 Defined as the equivalent damping coefficient; D eq2 It is related to the deviation between the angular frequencies of the grid-type converters and is defined as the equivalent mutual damping coefficient; Pref_eq and P GFL_eq They are defined as the equivalent mechanical power and equivalent output power of the grid-following converter respectively, but they are essentially the line impedance voltage drop and the q-axis component of the grid voltage.

[0063] The output power of the grid-type converter is:

[0064]

[0065] Substituting equation (2) into equation (8), we can obtain the active power and reactive power output by the grid-connected converter in the parallel system:

[0066]

[0067] According to formula (9), the output active power and reactive power of the grid-connected converter in the grid-connected / grid-connected parallel system have an additional expression related to the output current and equivalent power angle of the grid-connected converter, which are defined as the active power coupling term and the reactive power coupling term, respectively.

[0068] Therefore, the transient analysis model of the grid-type parallel system can be obtained as follows: Figure 3 shown.

[0069] Step 1-2: Analysis of the impact of network parameters on transient stability.

[0070] The control parameters of the active loop of the grid-type converter include inertia and damping. Figure 4 As shown, increasing the inertia of the grid-connected converter can simultaneously slow down ω GFM and ω GFL During the acceleration process during the fault, the greater the inertia ω GFM and ω GFL The smaller the acceleration, the better the transient stability of the grid-type converter and the grid-following converter, and the damping is similar. It can be concluded that whether the active power command value of the grid-type converter or the J and D values ​​are increased, the influence of the grid-type converter on the grid-following converter is through ω GFM The specific reasons are as follows: when the grid-type converter injects pure active current, the active power command value of the grid-type converter is reduced and the J and D values ​​are increased, resulting in ω GFM Deceleration, δ GFM The value becomes smaller, The value of will gradually become negative, and its value interval is [-π / 2, 0]. In this interval, its sine function value is negative, while its cosine function value is positive. At this time, the addition of the grid-type converter reduces the acceleration area of ​​the grid-type converter during the fault period and increases the deceleration area after the fault is removed, and the transient stability of the grid-type converter is improved. If pure reactive current is injected into the grid-type converter, The value of will gradually become positive, and its value interval is [0, π / 2]. In this interval, its sine function value is positive, while its cosine function value is negative. At this time, the addition of the grid-type converter increases the acceleration area of ​​the GFL during the fault period, reduces the deceleration area after the fault is removed, and the transient stability of the grid-type converter deteriorates. Therefore, only the optimal proportion of the grid-type converter under the condition of pure active current injected by the grid-type converter is discussed.

[0071] In a grid-type parallel system, the present application considers that the power output by the grid-type converter is generally equal to the value of its active power command value when the system is not unstable.

[0072] Analyze the relationship between the capacity and control parameters (J, D) of the grid-connected converter and construct an evaluation index system, including the following steps:

[0073] Step 2-1: Relationship between grid-connected converter capacity and control parameters (J, D).

[0074] Will Figure 2 If I1 and X1 are removed, we can get a network-type parallel system, such as Figure 6 shown.

[0075] The active power and reactive power at the internal potential are:

[0076]

[0077] Where δ e is the internal potential phase in the common coordinate system, E is the internal potential amplitude, f P is E and δ e The related work function, f Q is E and δ e Related reactive functions.

[0078] The combined GFM type synchronous machine expression can be obtained

[0079]

[0080] Among them, P ref is the active power command value, is δ GFM The differential of is δ GFM The second derivative of J GFM is the moment of inertia of the grid-type converter, D GFM is the damping coefficient of the grid-type converter;

[0081] Then replace δ in the above formula VSG Replace with δ e , we can get:

[0082]

[0083] in

[0084]

[0085] Among them, f t (δ e ) is about δ e The function of δ e is the internal potential phase in the common coordinate system, is δ e The differential of is δ e The second derivative of

[0086] Within the normal network parameter range (except for extremely weak power grids), the area where the system operating point is located under shallow disturbances meets the following conditions:

[0087]

[0088] That is, the output power and the internal potential phase show a monotonically increasing relationship.

[0089] For f in formula (13) t (δ e ) Taking the derivative we get:

[0090]

[0091] f t (δ e ) increases with the increase of the internal potential phase. Therefore, f t (δ e ) increases with the increase of power level.

[0092] According to the above theoretical analysis, since the GFM capacity is proportional to its output active power, f t (δ e ) increases with the increase of power level, and the parameters J, D and f t (δ e ), so the GFM capacity is positively correlated with J and D. GFM It is a key variable to measure the transient stability of GFL. When the active power P increases, the output angular frequency of the grid-type converter increases; when the damping coefficient D and inertia J increase, the output angular frequency decreases. By adjusting the active power, inertia and damping parameters, the capacity of the grid-type converter can be changed. When connecting to a new energy station, it is necessary to determine the appropriate capacity ratio to ensure stable operation in a weak grid environment.

[0093] Step 2-2: Evaluation indicator construction.

[0094] Although the increase in the capacity of the grid-type converter may lead to a deterioration in the transient stability of the GFL, it also enhances its support capacity for the GFL. Therefore, in order to comprehensively consider the impact of these two aspects, this application is based on ω GFM A quantitative evaluation index to characterize the transient stability of GFL is proposed to determine the optimal proportion of GFM converters.

[0095] Stability evaluation index

[0096] 1) Angular frequency change rate RoCoW

[0097] The angular frequency change rate RoCoW is defined as the inertia evaluation index. The larger the RoCoW value is, the faster the station power grid frequency changes and the worse the frequency stability is.

[0098]

[0099] Where: max is the maximum output angular frequency of the grid-type converter in the first swing cycle; ω0 is the initial angular frequency value; T C ω max rise time.

[0100] 2) Transient angular frequency deviation K ω

[0101] Define transient angular frequency deviation K ω It is an evaluation index of transient frequency stability, reflecting the ability of the system to maintain system stability when subjected to major load shocks or other abnormal situations.

[0102] K ω =ω max -ω0 (17)

[0103] Support capability evaluation indicators

[0104] 1) Damping active power

[0105] Definition of Damping Active Power It is an evaluation index to characterize the converter's ability to suppress output power oscillation.

[0106]

[0107] Where: is the average value of damping active power; P D0 is the initial damping active power, t D1 , t D2 The converter damping active power is greater than P D0 The initial and end time of the time window is 50ms.

[0108] 2) Inertia active power

[0109] Defining Inertia Active Power It is an evaluation index to characterize the short-time support capability of the converter. The larger the value, the faster the station responds to sudden changes in grid frequency.

[0110]

[0111] Where: is the average value of inertia active power, P H0 is the initial inertia active power; t H1 , t H2 The converter frequency active power operation value is greater than P H0 The initial and end time of the inertia active power, that is, the time window value is taken as 500ms.

[0112] In summary, the present invention proposes quantitative evaluation indicators to characterize the transient stability of GFL from two aspects: stability and supporting capacity, which is convenient for subsequent analysis.

[0113] Based on the above analysis, this embodiment proposes an optimal proportion configuration method for a grid-type parallel system based on the GRA-VIKOR method, such as Figure 7 As shown, the following steps are included:

[0114] Step 1: Assume that the total capacity of the hybrid renewable energy station converter is P W (This capacity is known in advance), where the capacity of the grid-type and grid-forming converters are P1 and P2 respectively, then P W =P1+P2; Output power P of grid-connected converter ref Less than its maximum output active power limit P max Under the premise, the capacity of the grid-type converter is continuously increased; the original data of the new energy station under different grid-type converter capacity ratios is standardized to eliminate the influence of the dimension and obtain the normalized data matrix Y m×n , where m is the evaluation object and n is the evaluation index;

[0115] Step 2: Combine the Bayesian BWM method to obtain the subjective weight w1 of the evaluation index, and the critical method (CRITIC method) to determine the objective weight w2 of the evaluation index, and determine the final combination weight w3 based on the game combination weighting method.

[0116] Step 3: Determine the weighted evaluation matrix according to formula (34) to obtain the positive and negative ideal solutions of each evaluation index of the grid-connected performance of the grid-connected control station; obtain the grey correlation between each evaluation index value and the positive and negative ideal solutions through the GRA idea; combine formulas (36, 37) to obtain the group utility value and individual regret value to determine the final group utility value Q i and individual regret value R i Finally, according to formula (40), the comprehensive evaluation index (VIKOR value) of the new energy station under different grid-type converter capacity ratios is obtained: i ;

[0117] Step 4: Determine the optimal GFM capacity ratio; from step 3, we can get T under each GFM ratio. i Assuming the best proportion is P1, the second best proportion is P2, and according to the maximum group utility value Q i The different GFM capacity ratios are ranked, and the best GFM capacity ratio solution is the best solution when it meets two conditions. The step 1 includes the following sub-steps:

[0118] Step S101: Conduct simulation experiments on parallel systems with different GFM capacity ratios and collect data:

[0119] Step S102: Performing homogeneous and standardized processing on the above-mentioned grid-connected performance evaluation indicators of the grid-forming type stations:

[0120] When processing evaluation index data, we will encounter the problem that the positive and negative attributes and dimensions of each evaluation index are different. Before calculation, the evaluation index needs to be oriented and standardized. The oriented method of converting negative evaluation indexes to positive evaluation indexes is:

[0121]

[0122] Where: μ is the coordination coefficient, generally taken as 0.1; x ij is the corresponding evaluation index value, x' ij is the evaluation index value after homogeneity, X i is the i-th row of matrix X; max|X i | is the maximum value of the i-th row of matrix X.

[0123] The standardized method of dimensionless processing of evaluation indicators is:

[0124]

[0125] In the formula, x i ” j is the standardized evaluation index value, and m is the number of evaluation indicators in the i-th row of matrix X.

[0126] From the above formula, we can get the standard matrix Ym×n .

[0127] The step 2 includes the following sub-steps:

[0128] Step S201: Conduct simulation experiments on the parallel systems of the grid / network construction under different GFM capacity ratios and collect data:

[0129] Step S202: Calculate the evaluation indicators of stability and support capacity under different GFM capacity ratio experimental scenarios, and then calculate the subjective weights of the proposed evaluation indicators according to the Bayesian method:

[0130] First, k experts are invited to select the evaluation index system (A = {A1, A2, ..., A n}) to determine the optimal evaluation index A B and the worst evaluation index A W , using the 1-9 scale method, A B and A W As a reference comparison object, construct the best and worst comparison matrix C B ={a B1 ,a B2 ,…,a Bn}, C W ={a 1W ,a 2W ,…,a nW},A1,A2,…,A n They are evaluation indicators, a B1 ,a B2 ,…,a Bn are the optimal evaluation index values ​​evaluated by the 1st,…nth experts, respectively, 1W ,a 2W ,…,a nW are the worst evaluation index values ​​evaluated by the 1st,…nth experts respectively.

[0131] Second, use the multinomial distribution to distribute the input A B , A W Modeling. The target layer is regarded as a random event, the evaluation index layer weight is regarded as the probability of the random event, and the worst evaluation index A W The multinomial probability distribution of is as follows:

[0132]

[0133] Where: w is the probability distribution (weight), a jw is the jth worst evaluation index. Assume that the weight w j Total number of random events Proportional, we can get the relationship between weight and expert rating:

[0134]

[0135] Among them, w w is the weight of the worst evaluation indicator;

[0136] Similarly, input A B The multinomial probability distribution of is:

[0137]

[0138] Among them, a Bj is the jth optimal evaluation index, w B is the optimal evaluation index weight.

[0139] The Dirichlet distribution corresponding to the prior distribution of output w is introduced:

[0140]

[0141] Where: B(x) is the multivariate Beta function, x is the dimensionless Dirichlet prior distribution parameter, x j is the j-th dimensionless Dirichlet prior distribution parameter. According to the conjugation property of the Dirichlet distribution, the posterior distribution of the corresponding output w can be obtained:

[0142]

[0143] Among them, C jw is the number of observations of category j and output w;

[0144] Finally, based on the input and output of the model, the subjective weight vector W1 under different GFM proportion experimental scenarios can be obtained by formulating a Bayesian hierarchical model. It can be understood that the calculation process of the subjective weight of the evaluation index through the Bayesian hierarchical model can be calculated using existing technology, which is not limited here.

[0145] Step S203: Calculate the objective weight of the proposed evaluation index according to the CRITIC method:

[0146] In the CRITIC method, standard deviation is used to reflect the variability of the evaluation index, that is, the difference in the value of the same evaluation index under different schemes is used to illustrate the evaluation strength of the evaluation index. m×n The coefficient of variation of each evaluation indicator can be obtained:

[0147]

[0148] Where: s i is the standard deviation of the i-th evaluation indicator; is the mean value of the i-th evaluation index, V i is the coefficient of variation; V i is the coefficient of variation.

[0149] 1) Obtaining the conflict of evaluation indicators

[0150] The correlation coefficient can reflect the conflict between evaluation indicators. Based on the correlation coefficient, an expression representing the conflict between evaluation indicators is constructed:

[0151]

[0152] Where: S i Indicates the conflict between the i-th evaluation indicator and other evaluation indicators, r ij is the correlation coefficient between evaluation index i and evaluation index j, x i is the i-th evaluation indicator, is the mean value of the i-th evaluation indicator, y j is the j-th evaluation indicator, is the mean value of the j-th evaluation indicator.

[0153] 2) Calculate the objective weight of evaluation indicators

[0154] V i With S i Multiplying them together gives the information amount G of the i-th evaluation index i , G i The larger the value is, the higher the amount of information contained in the evaluation index is, and the larger the weight value assigned should be. Therefore, the normalized weight of the i-th evaluation index is:

[0155]

[0156] Based on the input and output of the model, the objective weight value w corresponding to each performance evaluation indicator can be obtained through the above calculation formula i , after aggregating the objective weights corresponding to each evaluation indicator, we can obtain the objective weight vector W2 corresponding to the experimental scenarios with different GFM proportions.

[0157] Step S204: Use the game weighting method to determine the comprehensive weight, the steps are as follows:

[0158] 1) Construct a combined weight vector

[0159]

[0160] Where: is the subjective weight vector, The objective weight vector, is the optimal combination weight, A1 and A2 are the combination coefficients corresponding to each weight;

[0161] Among them, A1 and A2 are obtained through the game theory calculation formula:

[0162] 2) Introducing game theory

[0163]

[0164] Among them, A V is the combination coefficient variable, V = 1, 2, A1 is the coefficient of W1, A2 is the coefficient of W2, are the corresponding subjective or objective weights, W1 is the subjective weight, and W2 is the objective weight.

[0165] Solve the above formula to get the final combination coefficient A h (h=1,2)

[0166] 3) Normalize the coefficients:

[0167]

[0168] 4) Get the optimal combination weight:

[0169]

[0170] in, is the combination coefficient of subjective weight, is the combination coefficient of objective weights, is the combined weight.

[0171] The step three includes the following sub-steps:

[0172] Step S301: construct a weighted evaluation matrix and determine positive and negative ideal solutions under experimental scenarios with different proportions of grid-type converters.

[0173] 1) Construct a weighted evaluation matrix

[0174] Select m evaluation objects and n evaluation indicators, and determine the original standard matrix Y=(x ij ) m×n , normalize Y and multiply it by the combined weight of the evaluation indicators to obtain the constructed weighted evaluation matrix Z:

[0175] Z=(z ij ) m×n =(y ij ω i ) m×n (34)

[0176] 2) Determine positive and negative ideal solutions

[0177] In the improved TOPSIS method, the optimal and worst limit value sets in the actual data of each evaluation index are used to construct the optimal and worst value sequence, and normalized, which is used as the positive and negative ideal solution Z of each evaluation index. + , Z -.

[0178] Step S302: Determine the grey correlation between each evaluation index value and the positive and negative ideal solutions, and the final group utility value Q i and individual regret value R i .

[0179] 1) Calculate the grey correlation coefficient

[0180] The grey correlation coefficient between the jth evaluation index of the i-th GFM proportion experimental scenario and the positive and negative ideal solutions is:

[0181]

[0182] Among them, ρ is the resolution coefficient, ρ∈[0,1]; is the grey correlation coefficient between the jth evaluation index and the positive ideal solution of the i-th grid-connected converter proportion case, is the grey correlation coefficient between the jth evaluation index and the negative ideal solution of the i-th grid-connected converter proportion case, z ij represents the jth weighted evaluation index of the i-th grid-connected converter proportion case, Z + represents a positive ideal solution, Z - represents a negative ideal solution;

[0183] 2) Determine the final group utility value Q i and individual regret value R i Numeric

[0184] First, we calculate the group utility value Q in the experimental scenario without considering the different GFM proportions of GRA i and individual regret value R i Numeric

[0185]

[0186]

[0187] In the formula, ω j is the evaluation index z j Weight, is the positive ideal solution of the jth evaluation index, is the negative ideal solution of the jth evaluation index; when the group utility value Q i and individual regret value R i The smaller the value, the better the GFM ratio scheme is.

[0188] The final group utility value Q i and individual regret value R i The numerical value is obtained by the following formula.

[0189]

[0190]

[0191] Step S303: According to formula (40), the comprehensive evaluation index (VIKOR value) of the new energy station under different grid-type converter capacity ratios can be obtained. i

[0192]

[0193] In the formula, Q + =min(Q i ), Q - =max(Q i ), R + =min(R i ), R - =max(R i ), θ is the decision-making mechanism coefficient, θ∈[0,1]; when θ∈[0,0.5], it means that more emphasis is placed on pursuing individual regret value; when θ∈[0.5,1], it means that more emphasis is placed on group utility value; since this embodiment does not conduct an in-depth study on θ, the compromise coefficient value is taken as 0.5, which means that both the group utility value is maximized and the individual regret value is minimized.

[0194] In step 4, the T under each GFM capacity ratio can be obtained from step 3. i The optimal GFM capacity ratio can be determined by the following rules. Assume that the optimal ratio is P1, the second-best ratio is P2, and the maximum group utility value Q i Sort different GFM capacity ratios. The best GFM capacity ratio solution is the best solution if it meets the following two conditions:

[0195] Condition 1: T(P1)-T(P2)≥1 / n-1; T(P1) is the T of the optimal proportion solution i value, T(P2) is the T of the second best proportion solution i value;

[0196] Condition 2: The optimal proportion plan P1 is in the group utility value Q i and individual regret R i The one ranked first in the list.

[0197] The criteria for the best ratio solution based on the above two conditions are as follows:

[0198] 1. When conditions 1 and 2 are met, P1 is the best proportion solution;

[0199] 2. When condition 1 is met, P1 and P2 are the best proportion solutions.

[0200] 3. When condition 2 is met, any pre-selected proportion scheme that does not meet condition 1 is the best proportion scheme.

[0201] The above shows and describes the basic principles, main methods and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited by the above examples. The above examples and descriptions are only for explaining the principles and schemes of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention to be protected. The scope of the present invention to be protected is defined by the attached claims and their equivalents.

Claims

1. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method is characterized by: The following steps are involved: Step 1: The total capacity of the hybrid new energy station converter is P W , where the capacity of the grid-following and grid-forming converters are P1 and P2 respectively, then P W =P1+P2; Output power P of grid-connected converter ref Less than its maximum output active power limit P max Under the premise of increasing the capacity of grid-type converters, the original data of new energy stations under different grid-type converter capacity ratios are standardized to eliminate the impact of dimensions and obtain the normalized data matrix Y m×n , where m is the evaluation object and n is the evaluation index; Step 2: Combine the Bayesian BWM method to obtain the subjective weight w1 of the evaluation index, and the critical method (CRITIC method) to determine the objective weight w2 of the evaluation index, and determine the final combination weight w3 based on the game combination weighting method; Step 3: Determine the weighted evaluation matrix and obtain the positive and negative ideal solutions of each evaluation index of the grid-connected performance of the grid-connected control station; obtain the grey correlation between each evaluation index value and the positive and negative ideal solutions through the GRA idea; obtain the group utility value and individual regret value to determine the final group utility value Q i and individual regret value R i Finally, the comprehensive evaluation index T of the new energy station under different grid-type converter capacity ratios is obtained. i ; Step 4: Determine the optimal grid-type converter capacity ratio: From step 3, we can get the T under each grid-type converter capacity ratio. i Assuming the best proportion is P1, the second best proportion is P2, and according to the maximum group utility value Q i The capacity proportions of different grid-type converters are ranked, and the best grid-type converter capacity proportion scheme is selected as the best solution when it meets two conditions.

2. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 1 is characterized in that: The evaluation indicators include: Stability evaluation indicators: angular frequency change rate, transient angular frequency deviation Support capacity evaluation indicators: damping active power and inertia active power.

3. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 1 is characterized in that: The step 1 includes the following sub-steps: Step S101: Conduct simulation experiments on parallel systems with different grid-type converter capacity ratios to collect data: Step S102: The grid-connected performance evaluation indicators of the grid-forming stations are homogenized and standardized.

4. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 1 is characterized in that: The step 2 includes the following sub-steps: Step S201: Conduct simulation experiments on the grid-connected / grid-connected parallel systems under different grid-connected converter capacity ratios to collect data: Step S202: Calculate the evaluation indicators of stability and support capability in the experimental scenarios with different proportions of grid-type converters, and then calculate the subjective weights of the proposed evaluation indicators according to the Bayesian method: Step S203: Calculate the objective weight of the proposed evaluation index according to the CRITIC method: Step S204: Use the game weighting method to determine the comprehensive weight.

5. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 1 is characterized in that: The step three includes the following sub-steps: Step S301: construct a weighted evaluation matrix and determine positive and negative ideal solutions under different grid-type converter proportion experimental scenarios; Step S302: Determine the grey correlation between each evaluation index value and the positive and negative ideal solutions, and the final group utility value Q i and individual regret value R i ; Step S303: Obtain the comprehensive evaluation index T of the new energy station under different grid-type converter capacity proportions i .

6. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 4 is characterized in that: In step S202, the subjective weight of the evaluation index proposed is calculated according to the Bayesian method: First, k experts are invited to select from the evaluation index system A = {A1, A2, …, A n }Determine the optimal evaluation index A B and the worst evaluation index A W , using the 1-9 scale method, A B and A W As a reference comparison object, construct the best and worst comparison matrix C B ={a B1 ,a B2 ,…,a Bn }, C W ={a 1W ,a 2W ,…,a nW },A1,A2,…,A n They are evaluation indicators, a B1 ,a B2 ,…,a Bn are the optimal evaluation index values ​​evaluated by the 1st,…nth experts, respectively, 1W ,a 2W ,…,a nW are the worst evaluation index values ​​evaluated by the 1st,…nth experts respectively; Second, use the multinomial distribution to distribute the input A B , A W Modeling; take the target layer as a random event, the evaluation index layer weight as the probability of the random event, and the worst evaluation index A W The multinomial probability distribution of is as follows: Where: w is the weight, a jw is the jth worst evaluation index; let the weight w j Total number of random events Proportional to the expert score, we get the relationship between the weight and the expert score: Among them, w w is the weight of the worst evaluation indicator; Similarly, input A B The multinomial probability distribution of is: Among them, a Bj is the jth optimal evaluation index, w B is the optimal evaluation index weight; The Dirichlet distribution corresponding to the prior distribution of output w is introduced: Where: B(x) is the multivariate Beta function, x is the dimensionless Dirichlet prior distribution parameter, x j is the j-th dimensionless Dirichlet prior distribution parameter; from the conjugation of the Dirichlet distribution, the posterior distribution corresponding to the output w is obtained: Among them, C jw is the number of observations of category j and output w; Finally, based on the input and output of the model, the subjective weight vector W1 under the experimental scenarios of different grid-type converter capacity ratios can be obtained by formulating a Bayesian hierarchical model.

7. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 6 is characterized in that: In step S203, the process of calculating the objective weight of the proposed evaluation index according to the CRITIC method is as follows: The difference in the values ​​of the same evaluation indicator under different schemes is used to illustrate the evaluation strength of the evaluation indicator. m×n The coefficient of variation of each evaluation indicator can be obtained: Where: s i is the standard deviation of the i-th evaluation indicator; is the mean value of the i-th evaluation index, V i is the coefficient of variation; V i is the coefficient of variation; Obtain the conflict of evaluation indicators: Where: S i Indicates the conflict between the i-th evaluation indicator and other evaluation indicators, r ij is the correlation coefficient between evaluation index i and evaluation index j, x i is the i-th evaluation indicator, is the mean value of the i-th evaluation indicator, y j is the j-th evaluation indicator, is the mean value of the j-th evaluation indicator; Calculate the objective weight of the evaluation index: V i With S i Multiplying them together gives the information amount G of the i-th evaluation index i , the normalized weight of the i-th evaluation index is: Based on the input and output of the model, the objective weight value w corresponding to each performance evaluation indicator is obtained through the above calculation formula i , after aggregating the objective weights corresponding to each evaluation index, we can obtain the objective weight vector W2 corresponding to the experimental scenarios of different grid-type converter capacity proportions.

8. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 1 is characterized in that: Step S204: Use the game weighting method to determine the comprehensive weight, the steps are as follows: Construct the combined weight vector: Where: is the subjective weight vector, The objective weight vector, is the optimal combination weight, A1 and A2 are the combination coefficients corresponding to each weight; Among them, A1 and A2 are obtained through the game theory calculation formula: Introducing game theory Among them, A V is the combination coefficient variable, V = 1, 2, A1 is the coefficient of W1, A2 is the coefficient of W2, is the corresponding subjective or objective weight, W1 is the subjective weight, and W2 is the objective weight; Solve the above formula to get the final combination coefficient A h , h=1,2; Normalize the coefficients: Get the optimal combination weight: in, is the combination coefficient of subjective weight, is the combination coefficient of objective weights, is the combined weight.

9. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 1 is characterized in that: The step S301 includes: Construct a weighted evaluation matrix: select m evaluation objects and n evaluation indicators, and determine the original standard matrix Y=(x ij ) m×n , normalize Y and multiply it by the combined weight of the evaluation indicators to obtain the constructed weighted evaluation matrix Z; Determine the positive and negative ideal solutions: Use the optimal and worst limit value sets in the actual data of each evaluation indicator to construct the optimal and worst value sequence, and perform normalization processing to use this as the positive and negative ideal solutions Z of each evaluation indicator + , Z - . In step S302, the final group utility value Q is determined i and individual regret value R i The numerical process is: First, we calculate the group utility value Q in the experimental scenario without considering the capacity ratio of GRA grid-type converters of different configurations. i and individual regret value R i Value: In the formula, ω j is the evaluation index z j Weight, is the positive ideal solution of the jth evaluation index, is the negative ideal solution of each j-th evaluation index; The final group utility value Q i and individual regret value R i The numerical value is obtained by the following formula. in, is the grey correlation coefficient between the jth evaluation index and the positive ideal solution of the i-th grid-connected converter proportion case, is the grey correlation coefficient between the jth evaluation index and the negative ideal solution of the i-th grid-type converter proportion case; In step S303, the comprehensive evaluation index T of the new energy station under different grid-type converter capacity ratios is obtained according to the following formula: i : In the formula, Q + =min(Q i ), Q - =max(Q i ), R + =min(R i ), R - =max(R i ), θ is the decision mechanism coefficient, θ∈[0,1].

10. The optimal proportion configuration method of the grid-type parallel system based on the GRA-VIKOR method according to claim 1 is characterized in that: In step 4, let the best proportion scheme be P1, the second best proportion scheme be P2, and according to the maximum group utility value Q i The capacity proportions of different grid-type converters are ranked; the optimal grid-type converter capacity proportion scheme is the best solution when it meets the following two conditions: Condition 1: T(P1)-T(P2)≥1 / n-1; T(P1) is the T of the optimal proportion solution i value, T(P2) is the T of the second best proportion solution i value; Condition 2: The best proportion scheme P1 is in the group utility value Q i and individual regret R i The ranking is the highest among the The criteria for the best ratio solution based on the above two conditions are as follows:

1. When conditions 1 and 2 are met, P1 is the best proportion solution; 2. When condition 1 is met, P1 and P2 are the best proportion solutions.

3. When condition 2 is met, any pre-selected proportion scheme that does not meet condition 1 is the best proportion scheme.

Images