A reliability assessment method for distribution network containing distributed generation

Abstract

A reliability assessment method for a distribution network containing distributed power sources, firstly, adopts the kernel density estimation method to construct a probability distribution model of the output of wind turbines and photovoltaic generators; then, based on the Copula theory, selects the Copula function with the best fitting effect on the wind and solar output, and constructs a probability distribution model of the wind and solar combined output; then, introduces new indicators on the basis of the traditional distribution network reliability assessment indicators to evaluate the reliability of the new distribution network containing DG; finally, adopts the Monte Carlo sampling method to calculate the reliability of the system after different DG types are connected to the distribution network. The wind and solar combined output model adopted by the present invention not only takes into account the uncertainty of wind and solar output, but also takes into account the correlation between them, and the model is more in line with the actual situation. The proposed new reliability assessment index can more intuitively describe the impact of the distributed power source on the reliability of the distribution network after being connected to the distribution network, and thus can provide a basis for the planning of the distribution network containing new energy.

Description

Technical Field

[0001] The present invention relates to the technical field of smart distribution networks, and in particular to a reliability assessment method for a distribution network containing distributed power sources. Background Art

[0002] According to statistics, about 80% of power outages in the power system are caused by faults in the distribution network. Therefore, evaluating and improving the reliability of the distribution network is of great significance to improving the reliability of the entire power system. With the large-scale access of DG such as wind power and photovoltaic power to the distribution network, the traditional distribution network reliability assessment is no longer applicable, and a new distribution network reliability assessment method containing DG needs to be proposed.

[0003] At present, there have been studies on the reliability assessment of distribution networks containing DG. For example, in the literature [1]: "Reliability assessment of distribution systems containing distributed generation considering correlation [J]. Electric Power Construction, 2016, 37(9): 108-114", Hu Meiyu et al. proposed a model that considers the correlation between DG and load based on the rank correlation coefficient matrix theory and the Latin hypercube sampling method. In addition, they considered the uncertainty of DG output and load power. In the reliability assessment, this literature only considered wind turbines, and did not consider the output of photovoltaic generators.

[0004] Ali Arefi et al. proposed a new single-step method in the literature [2]: "A Fast Adequacy Analysis for Radial Distribution Networks Considering Reconfiguration and DGs[J].IEEE transactions on smart grid, 2020, 11(5):3896-3909." In view of the problem that the search scheme in the reliability assessment of radial distribution systems requires complex programs and occupies a large amount of CPU, a new single-step method is proposed, which uses a set of binary description matrices to replace the search and calculate the load and system reliability indicators. Although this paper adopts a new distribution network reliability assessment method - the direct method, this method is only applicable to power systems with simple topology structures, and is not suitable for reliability calculation of power systems with complex structures.

[0005] Li Rui et al. proposed a Weibull-Markov model in the literature [3]: "Simulation Analysis of the Impact of Distributed Generation Access on Distribution Network Reliability [J]. Power System Technology, 2016, 40(7): 2016-2021.", which can more accurately simulate the multiple operating states of DG. Although the literature considers the operating state of wind and solar power in output, it does not consider the correlation between the two.

[0006] Ding Ming et al. [4], "Reliability Analysis of Distribution Networks with High Penetration of Renewable Energy [J]. Acta Energiae Solaris Sinica, 2020, 41(2): 194-202.", took the high penetration of renewable energy access to distribution networks as the background and analyzed the impact of the penetration rate of distributed generation on the power supply mode and reliability of the connected system. Although this paper considers the reliability evaluation of DGs with different penetration rates after access to distribution networks, it still uses traditional distribution network reliability evaluation indicators.

[0007] In the literature [5]: "Reliability evaluation of distribution network with distributed generation based on Latin hypercube sequential sampling [A] / / IEEE, 2020: 97-99.", Liu Aibin et al. proposed a new algorithm combining Latin hypercube sampling (LHS) and MC sequential simulation to solve the problems of multiple sampling and time consumption when using Monte Carlo simulation to evaluate the reliability of distribution networks containing DG. The shortcoming of this literature is that the scenario is relatively simple when performing case analysis, and its superiority is only reflected in the algorithm, without reflecting the impact of DG access on the reliability of the distribution network.

[0008] Guo Yugang et al. [6] considered the power generation and consumption characteristics of future distribution networks and proposed a comprehensive analysis method for distribution network reliability that takes into account the uncertainty of grid topology, equipment factors, power output, and load demand. Although there are improvements in reliability indicators, the correlation between DG outputs is not involved.

[0009] Kai Zou et al. [7] used a parameter-free particle swarm optimization algorithm to solve the uncertainty problems of distributed power output, time-varying load demand, and random fault components in reliability assessment. In terms of model establishment, a probabilistic model was used to simulate the uncertainty of renewable energy availability, time-varying load demand, random prediction errors, and random faults, without considering the actual wind and solar output data.

[0010] In summary, the shortcomings of the current research are that when performing reliability calculations, only the uncertainty of wind and solar power output is considered, without considering its correlation. In addition, the reliability assessment indicators of traditional distribution networks are used when conducting reliability assessments of distribution networks containing DG. Summary of the invention

[0011] In order to solve the above technical problems, the present invention provides a reliability assessment method for a distribution network containing distributed power sources. The method takes into account the correlation between wind and solar power outputs, selects the Frank-Copula with the best fitting effect to construct a wind-solar joint output model, and introduces an evaluation index containing DG, which can more intuitively describe the impact of DG access on the reliability of the distribution network.

[0012] The technical solution adopted by the present invention is:

[0013] A method for evaluating the reliability of a distribution network including distributed power sources comprises the following steps:

[0014] Step 1: Use the kernel density estimation method to construct a probability distribution model of wind turbine and photovoltaic generator output;

[0015] Step 2: Based on the Copula theory, select the Copula function with the best fitting effect on wind and solar power output, and build a probability distribution model for wind and solar power combined output;

[0016] Step 3: Introduce new indicators based on the traditional distribution network reliability evaluation indicators to evaluate the reliability of the new distribution network containing DG;

[0017] Step 4: Monte Carlo sampling method is used to calculate the reliability of the system after different DG types are connected to the distribution network. In step 1, the probability distribution model of the output of wind turbines and photovoltaic generators is constructed as follows:

[0018] The marginal distribution of wind turbine and photovoltaic generator output power is determined by nonparametric kernel density estimation method, and P is set w1 ,P w2 ,…,P wn and P v1 ,P v2 ,…,P vn P is taken from discrete calculations. w and P v Sample space, where:

[0019] P w1 ,P w2 ,…,P wn is the sequence sample of the output power of n wind turbines; P v1 ,P v2 ,…,P vn is the sequence sample of the output power of n photovoltaic generators; P w , P v They represent the sample space of the output power of wind turbines and photovoltaic generators respectively.

[0020] Substituting it into the kernel density estimation formula, the probability density function f of the output power of wind turbines and photovoltaic generators is estimated. WT (P w ) and f PV (P v ).

[0021] f WT (P w ) and f PV (P v ) is integrated to obtain the probability distribution function of the output power of wind turbines and photovoltaic generators:

[0022]

[0023] F WT (P w ) is the probability distribution function of the wind turbine output power.

[0024]

[0025] F PV (P v ) is the probability distribution function of the output power of the photovoltaic generator.

[0026] The step 2 specifically includes the following steps:

[0027] Step 2.1: Obtain relevant data on the output of wind turbines and photovoltaic generators;

[0028] Step 2.2: Considering that wind speed and light intensity have a certain correlation within a certain range, based on the Copula theory, the Copula function with the best fit is selected as the connection function between wind and solar outputs;

[0029] Step 2.3: Based on the distribution function F of wind and solar output established in step 1 WT (P w ) and F PV (P v ), combined with the Frank-Copula function in step 2.2, the wind-solar combined output distribution function can be obtained.

[0030] In step 2, in order to select the Copula function with the best fitting effect on wind and solar output, the Spearman rank correlation coefficient ρ, Kendall rank correlation coefficient τ, and squared Euclidean distance d are introduced. 2 and other indicators, and calculate the Empirical-Copula function of wind and solar output. The larger the rank correlation coefficient of the selected Copula function, the stronger its correlation, the closer the rank correlation coefficient is to the rank correlation coefficient of the Empirical-Copula function, and the smaller the Euclidean distance is, the better the fit. The results show that the Frank-Copula function has the best fitting effect, and its probability distribution expression is as follows:

[0031]

[0032] Where θ represents the correlation coefficient between random variables u and v. C(u,v) is the probability distribution expression of the Frank-Copula function; e -θu and e -θv is the generator of the Frank-Copula function.

[0033] The distribution function expression of wind-solar combined output is:

[0034]

[0035] Where: θ is F(P w ) and F(P v ), if θ>0, the wind and solar outputs are positively correlated; if θ<0, the wind and solar outputs are negatively correlated; if θ→0, the wind and solar outputs tend to be independent of each other; are the generators of the Frank-Copula function after substituting wind and solar output respectively.

[0036] In step 3, a new distribution network reliability evaluation index containing DG is introduced. The reliability index of the load point is usually expressed as the average failure rate λ i , Annual average power outage time U i and the mean outage duration ri Three are used to represent that after DG joins, the DG outage probability λ needs to be considered DG and average power supply time U′ DG ;

[0037] The probability of DG outage λ DG :

[0038]

[0039] Where, T is the total running time; t{p t =0} is the time set in which the DG output is 0 during the total operating time. t is the probability that DG works normally; t is the time series.

[0040] Average power supply time of DG:

[0041]

[0042] In the formula, Year is the simulation period; P Li is the load power at load point i; P DG is the output power of DG; N is the number of normal operations; Represents the power supply time set during island operation. j Indicates the power supply time during island operation; R j Indicates the island operation range; i is the number of the load point; j is the number of the load point within the island range.

[0043] The reliability evaluation indicators of the system after connecting to DG are as follows:

[0044] (1): After DG is connected to the distribution network, DG itself also has a failure rate λ DG Therefore, the average power outage frequency index SAIFI of the system after DG access DG for:

[0045]

[0046] Where N i is the number of users at load point i; P DG∩N is the probability of simultaneous failure of DG and distribution network in the island. i Represents the failure rate of load point i within the island range.

[0047] (2) Average power outage duration index SAIDI of the system after DG access DG , when calculating, it is necessary to remove the number of users who are not affected by the fault due to the addition of DG and the power supply time of DG. It is usually assumed that the number of users affected when DG is working normally and when it fails is equal.

[0048]

[0049] U i represents the average power outage time of load point i within the island range; U′ i,DG It represents the average power outage time of load point i containing DG within the island range.

[0050] (3): ASAI, the power supply reliability index of the system after connecting to DG DG , and the number of users who are not affected by the fault due to the addition of DG and the power supply time of DG also need to be removed during the calculation.

[0051]

[0052] 8760 is the number of hours in a year, calculated from 24×365.

[0053] (4): Since the expected power shortage is determined by U i and load power P Li Therefore, the expected power shortage after adding DG is EENS DG for:

[0054]

[0055] After DG is connected to the distribution network, the traditional distribution network reliability evaluation indicators can no longer intuitively represent the changes in system reliability indicators after DG is connected. Therefore, it is necessary to introduce some new reliability evaluation indicators to characterize the impact of DG on the distribution network after it is connected to the distribution network:

[0056] (5): The influence coefficient K of DG on the average power outage frequency index SAIFI :

[0057]

[0058] Where △SAIFI is the average power outage frequency index change of the system before and after DG access; K SAIFI Indicates that the access rated capacity is P E The change index of the system's SAIFI after DG can reflect the impact of DG on the system's SAIFI. SAIFI represents the average power outage frequency index of the system.

[0059] (6): The influence coefficient K of DG on the average power outage duration index SAIDI :

[0060]

[0061] Where △SAIDI is the average power outage duration index change of the system before and after DG access; P E is the rated capacity of DG. K SAIDIIndicates that the access rated capacity is P E The change index of the SAIDI of the system after DG can reflect the impact of DG on the SAIDI of the system.

[0062] (7): DG's influence coefficient on power supply reliability K ASAI :

[0063]

[0064] In the formula, △ASAI is the change in the power supply reliability index of the system before and after DG access. ASAI Indicates that the access rated capacity is P E The change index of the ASAI of the system after DG can reflect the impact of DG on the ASAI of the system.

[0065] (8): The influence coefficient K of DG on the expected power shortage EENS :

[0066]

[0067] Where △EENS is the expected power shortage index change of the system before and after DG access; K EENS Indicates that the access rated capacity is P E The change index of the system's EENS after DG can reflect the impact of DG on the system's EENS.

[0068] The step 4 comprises the following steps:

[0069] S4.1: Read the basic parameters of the distribution network, set the simulation period N, and initialize the data and simulation time;

[0070] S4.2: Number each component and sample the component status through Monte Carlo simulation, that is, generate a random number at (0,1), calculate the normal working time TTF and fault repair time TTR of each component respectively, and form the operating state sequence of each component during the simulation time. The fault isolation and switching time T res .

[0071] T TTF = -1 / λ·lnσ;

[0072] T TTR = -1 / μ·lnσ;

[0073] Where λ and μ are the failure rate and repair rate of the component respectively; σ is a random number between (0,1).

[0074] S4.3: Compare the normal working time of all components and find TTFmin The smallest component i is set as the fault component;

[0075] S4.4: After finding the location of the faulty component, determine whether it is located within the planned island, and determine the scope of island operation based on the location of the DG, and then classify the load:

[0076] S4.5: For Class IV loads, the power outage time and normal power supply time are determined by the DG output and load demand in the isolated island. Therefore, the DG output P during the isolated island operation time is calculated based on the distribution function of the wind-solar combined output. DG,i (t) and load demand P L,i (t) Conduct sampling.

[0077] S4.6: Determine whether the output of the DG in the island meets the load demand in the island. If so, the load in the island continues to be supplied by the DG; otherwise, perform load shedding operations until the power balance is achieved in the island. Each time a load shedding operation is performed, the corresponding outage time and number of power outages shall be accumulated;

[0078] S4.7: Determine whether the simulation lifespan has been reached. If so, proceed to the next step; otherwise, return to step 2;

[0079] S4.8: Determine again whether the load point i is located in the island. If so, calculate the reliability index of the load point and the system according to the evaluation index calculation formula including DG; otherwise, calculate according to the calculation method of traditional indicators.

[0080] In step 4.4, the loads are classified as follows:

[0081] Class I load: The load that is not affected by a main grid fault. Since it is not affected by faulty components, it supplies power normally, so the total number of power outages and the duration of power outages remain unchanged.

[0082] Class II load: load that can resume normal operation after the faulty component is repaired. The outage time TTR is added to the total power outage time, and the total power outage times are increased by 1.

[0083] Category III load: After a main grid fault occurs, the load can be restored through switch operation. The outage time is the fault isolation and switching time T res , add it to the total power outage time, and the total power outage times increase by 1.

[0084] Category IV load: load within the planned island.

[0085] The present invention provides a method for evaluating the reliability of a distribution network containing distributed power sources, and the technical effects are as follows:

[0086] 1) The present invention takes into account the uncertainty of wind and solar power output and establishes the probability density function of wind power generation and photovoltaic power generation; in order to describe the characteristics of wind and solar power output more realistically, while also taking into account the correlation of wind and solar power output, based on Copula theory, the Frank-Copula function with the best fitting effect for wind power generation and photovoltaic power generation is selected as the connection function, and a wind-solar joint output model is established. The parameters of its output power are used for the reliability evaluation of the distribution network, which better characterizes the reliability of the distribution network containing distributed power sources.

[0087] 2) The present invention introduces a new distribution network reliability evaluation index containing DG, which can better describe the impact of adding DG on the reliability of the distribution network.

[0088] 3) The present invention can provide a theoretical basis for the connection of future large-scale renewable energy with distribution networks and the planning of distribution networks by conducting reliability assessment on new distribution networks containing distributed power sources. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 It is a schematic diagram of the process of the present invention;

[0090] Figure 2(a) shows the frequency histogram and kernel density estimation diagram of the wind turbine output power;

[0091] Figure 2(b) shows the frequency histogram and kernel density estimation diagram of the photovoltaic generator output power;

[0092] Figure 3 (a) Wind-solar combined output distribution function diagram;

[0093] Figure 3 (b) Probability density distribution of wind-solar combined output;

[0094] Figure 4 This is the flow chart of reliability assessment of distribution network including DG;

[0095] Figure 5 To improve the F4 feeder structure diagram under the IEEE-RBTS Bus6 test system;

[0096] Figure 6(a) is a graph showing the average failure rate index for some load points;

[0097] Figure 6(b) is the average power outage time indicator diagram for some load points. DETAILED DESCRIPTION

[0098] A reliability assessment method for a distribution network containing distributed power sources is based on the Copula theory, and the Copula function with the best wind-solar output fitting effect is selected as the connection function for constructing a wind-solar joint output model. In addition, considering the impact of DG access on the distribution network, a new reliability assessment index containing DG is introduced, and a distribution network reliability assessment system containing DG is constructed to characterize the impact of DG access on the reliability of the distribution network. The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0099] S1: Using the kernel density estimation method, a probability distribution model of wind turbine and photovoltaic generator output is constructed;

[0100] Kernel density estimation is a non-parametric test method. Non-parametric estimation does not incorporate any prior knowledge, but fits the distribution based on the characteristics and properties of the data itself, which can produce a better model than the parameter estimation method.

[0101] Assume x1,x2,…,x n is a sample of the random variable x. If the probability density function of the random variable x is f(x), then the kernel density estimate of f(x) is:

[0102]

[0103] Where n is the sample size; h>0 is the smoothing parameter; K(·) is the kernel function.

[0104] The output power of a wind turbine is related to the wind speed, and the mathematical expression of its output power is:

[0105]

[0106] Where: v wi 、v wo and v r are cut-in wind speed, cut-out wind speed and rated wind speed respectively; P r is the rated power of a single wind turbine.

[0107] The output power of a photovoltaic generator is determined by the light intensity on the surface of the solar panel. The mathematical expression of its output power is:

[0108] P v =A×η×r (3);

[0109] Where: A is the total area of ​​the photovoltaic generator panel; η is the photoelectric conversion efficiency; r is the actual light intensity.

[0110] The Homer software is used to obtain the historical data of sunlight intensity at each time point. The monthly average wind speed of a certain geographical location in the NASA database is used, and then converted into hourly wind speed data using the Homer software. The present invention selects one year (8760 hours) of wind speed and sunlight data. Substituting the wind speed and sunlight data into equations (2) and (3) respectively can obtain the output power P of the wind turbine and photovoltaic generator. W , P V .

[0111] The marginal distribution of wind turbine and photovoltaic generator output power is determined by nonparametric kernel density estimation method, and P is set w1 ,P w2 ,…,P wn and P v1 ,P v2 ,…,P vn P is taken from discrete calculations. w and P v Substituting the sample space into equation (1), the probability density function f of the output power of wind turbines and photovoltaic generators is estimated WT (P w ) and f PV (P v ), the frequency histogram and kernel density estimation diagram of the measured data are shown in Figure 2(a) and Figure 2(b).

[0112] f WT (P w ) and f PV (P v ) is integrated to obtain the probability distribution function of the output power of wind turbines and photovoltaic generators:

[0113]

[0114]

[0115] S2: Based on the Copula theory, select the Copula function with the best fitting effect on wind and solar power output, and build a probability distribution model for wind and solar power combined output;

[0116] S2.1: Obtain output data related to wind turbines and photovoltaic generators. Obtain output data based on the probability distribution model of wind turbine and photovoltaic generator output established in S1.

[0117] S2.2: Considering that wind speed and light intensity have a certain correlation within a certain range, based on the Copula theory, the Copula function with the best fitting is selected as the connection function between wind and solar output.

[0118] Copula theory can be used to solve high-dimensional joint distribution problems. This method transforms joint distribution problems into marginal distribution and correlation problems. The model is practical and effective, and simplifies calculations. Copula function is a function that connects the joint distribution of multiple random variables with their respective marginal distributions. Under the premise that the marginal distribution functions of two random variables x and y are known to be F(x) and G(y), a Copula function C(F(x), G(y)) can be used to represent their joint distribution function H(x, y) and density function h(x, y):

[0119] H(x, y) = C(F(x), G(y)) (6)

[0120] h(x,y)=c(F(x),G(y))f(x)g(y) (7)

[0121] Where c(F(x), G(y)), f(x), and g(y) are the density functions of C(F(x), G(y)), F(x), and G(y), respectively.

[0122] In order to select the best Copula function to fit the wind and solar output characteristics, the Spearman rank correlation coefficient ρ, Kendall rank correlation coefficient τ and squared Euclidean distance d are introduced. 2 The Empirical-Copula function of wind and solar output is calculated. The larger the rank correlation coefficient of the selected Copula function, the stronger its correlation. The closer the rank correlation coefficient is to the rank correlation coefficient of the Empirical-Copula function, and the smaller the Euclidean distance is, the better the fit is.

[0123] According to the output data of wind turbines and photovoltaic generators obtained in S2.1, the Gumbel-Copula, Frank-Copula, and Clayton-Copula functions are used to fit the wind and solar outputs and calculate the Empirical-Copula functions of the wind and solar outputs. The rank correlation coefficient and the square Euclidean distance of the Empirical-Copula function are obtained as shown in Table 1.

[0124] Table 1 Rank correlation coefficient and Euclidean distance

[0125]

[0126] From Table 1, we can see that the ρ and τ of the Frank-Copula function are the largest, and the squared Euclidean distance d 2 Minimum, therefore, the present invention selects the Frank-Copula function as the connection function of the wind-solar joint output model, and the expressions of its probability distribution function and density function are:

[0127]

[0128]

[0129] Where θ represents the correlation coefficient between random variables u and v.

[0130] S2.3: Based on the distribution function F of wind and solar output established in S1 WT (P w ) and F PV (P v ), combined with the Frank-Copula function in S2.2, the wind-solar joint output distribution function can be obtained as shown in the formula. Its probability distribution and density function diagrams are shown in Figure 3(a) and Figure 3(b).

[0131]

[0132] Where: θ is F(P w ) and F(P v ), if θ>0, the wind and solar outputs are positively correlated; if θ<0, the wind and solar outputs are negatively correlated; if θ→0, the wind and solar outputs tend to be independent of each other.

[0133] S3: Based on the traditional distribution network reliability evaluation index, a new index is introduced to evaluate the reliability of the new distribution network containing DG;

[0134] In the reliability assessment calculation of the traditional distribution network, when a fault occurs at load point i, it is generally considered that all loads at this load point are out of service. After adding DG, the loads outside the island will be processed according to the reliability assessment method of the traditional distribution network. For the loads inside the island, since DG can still supply power to them, the number of users without power will change. Therefore, it is necessary to fully consider the impact of DG access to the distribution network, improve the traditional distribution network, and form a new distribution network reliability assessment index system containing DG.

[0135] The reliability index of the load point is usually expressed as the average failure rate λ i , Fault repair time i and the annual average power outage time U i Three are used to represent that after DG is added, the average failure rate λ of DG needs to be considered DG and troubleshooting time γ DG For DG power generation, only when the distributed power source and the main feeder fail will the reliability calculation of the load point be affected. According to the parallel failure of the distributed power source and the main feeder, the reliability evaluation indicators of the load point in the island are as follows:

[0136]

[0137]

[0138]

[0139] In the formula, R j is the planning island range, λ S,i , DG,i are the failure rates of the main feeder and DG in the island respectively; γ S,i , γ DG,i are the fault repair time of the main feeder and DG in the island, respectively.

[0140] The probability of DG outage λ DG :

[0141]

[0142] Where, T is the total running time; t{p t =0} is the time set when the DG output is 0 during the total operating time.

[0143] Average power supply time of DG:

[0144]

[0145] In the formula, Year is the simulation period; P Li is the load power at load point i; P DG is the output power of DG; N is the number of normal operations; The power supply time collection when operating in an island.

[0146] The reliability evaluation indicators of the system after connecting to DG are as follows:

[0147] (1) After DG is connected to the distribution network, DG itself also has a failure rate λ DG Therefore, the average power outage frequency index SAIFI of the system after DG access DG for:

[0148]

[0149] Where N i is the number of users at load point i; P DG∩N is the probability of simultaneous failure of DG and distribution network in the island.

[0150] (2) Average power outage duration index SAIDI of the system after DG access DG , when calculating, it is necessary to remove the number of users who are not affected by the fault due to the addition of DG and the power supply time of DG. It is usually assumed that the number of users affected when DG is working normally and when it fails is equal.

[0151]

[0152] (3) System power supply reliability index ASAI after DG access DG , and the number of users who are not affected by the fault due to the addition of DG and the power supply time of DG also need to be removed during the calculation.

[0153]

[0154] (4) Since the expected power shortage is determined by U i and load power P Li Therefore, the expected power shortage after adding DG is EENS DG for:

[0155]

[0156] After DG is connected to the distribution network, the traditional distribution network reliability evaluation indicators can no longer intuitively represent the changes in system reliability indicators after DG is connected. Therefore, it is necessary to introduce some new reliability evaluation indicators to characterize the impact of DG on the distribution network after it is connected to the distribution network:

[0157] (5) The influence coefficient K of DG on the average power outage frequency index SAIFI :

[0158]

[0159] In the formula, △SAIFI is the average power outage frequency index change of the system before and after DG access. SAIFI Indicates that the access rated capacity is P E The change index of the SAIFI of the system after DG can reflect the impact of DG on the system SAIFI.

[0160] (6) The influence coefficient K of DG on the average power outage duration index SAIDI :

[0161]

[0162] Where △SAIDI is the average power outage duration index change of the system before and after DG access; P E is the rated capacity of DG. K SAIDI Indicates that the access rated capacity is P E The change index of the SAIDI of the system after DG can reflect the impact of DG on the SAIDI of the system.

[0163] (7) DG’s influence coefficient on power supply reliability K ASAI :

[0164]

[0165] In the formula, △ASAI is the change in the power supply reliability index of the system before and after DG access. ASAI Indicates that the access rated capacity is P E The change index of the ASAI of the system after DG can reflect the impact of DG on the ASAI of the system.

[0166] (8) The influence coefficient K of DG on the expected power shortage EENS :

[0167]

[0168] Where △EENS is the expected power shortage index change of the system before and after DG access. EENS Indicates that the access rated capacity is P E The change index of the system's EENS after DG can reflect the impact of DG on the system's EENS.

[0169] S4: Monte Carlo sampling method is used to calculate the reliability of the system after different DG types are connected to the distribution network;

[0170] S4.1: Read the basic parameters of the distribution network, set the simulation period N, and initialize the data and simulation time;

[0171] S4.2: Number each component and sample the component status through Monte Carlo simulation, that is, generate a random number at (0,1), calculate the normal working time TTF and fault repair time TTR of each component through equations (21) and (22) respectively, and form the operating state sequence of each component during the simulation time. The fault isolation and switching time T res .

[0172] T TTF = -1 / λ·lnσ (24)

[0173] T TTR = -1 / μ·lnσ (25)

[0174] Where λ and μ are the failure rate and repair rate of the component respectively; σ is a random number between (0,1).

[0175] S4.3: Compare the normal working time of all components and find TTF min The smallest component i is set as the fault component;

[0176] S4.4: After finding the location of the faulty component, determine whether it is located within the planned island, and determine the scope of island operation based on the location of the DG, and then classify the load:

[0177] Class I load: load that is not affected by a main grid failure. Since it is not affected by the faulty components and supplies power normally, the total number of power outages and power outage duration remain unchanged.

[0178] Class II load: load that can resume normal operation after the faulty component is repaired. The outage time TTR is added to the total power outage time, and the total power outage times are increased by 1.

[0179] Category III load: After a main grid fault occurs, the load can be restored through switch operation. The outage time is the fault isolation and switching time T res , add it to the total power outage time, and the total power outage times increase by 1.

[0180] Category IV load: load within the planned island.

[0181] S4.5: For Class IV loads, the power outage time and normal power supply time are determined by the DG output and load demand in the island. Therefore, according to the distribution function of the wind-solar combined output, the DG output P during the island operation time is calculated. DG,i (t) and load demand P L,i (t) Conduct sampling.

[0182] S4.6: Determine whether the output of the DG in the island meets the load demand in the island. If so, the load in the island continues to be supplied by the DG; otherwise, perform load shedding operations until the power balance is achieved in the island. Each time a load shedding operation is performed, the corresponding outage time and number of power outages shall be accumulated;

[0183] S4.7: Determine whether the simulation lifespan has been reached. If so, proceed to the next step; otherwise, return to step 2;

[0184] S4.8: Determine again whether the load point i is located in the island. If so, calculate the reliability index of the load point and the system according to the evaluation index calculation formula including DG; otherwise, calculate according to the calculation method of traditional indicators.

[0185] Implementation example:

[0186] The example uses the F4 feeder under the improved IEEE-RBTS Bus6 test system. The system structure diagram is shown in Figure 5 shown. Figure 5 There are 30 feeders, 23 load points, 23 transformers, 3 disconnectors, 5 circuit breakers, and 2 distributed power sources. Assume that the probability of circuit breaker operation reliability is 100% and the disconnector operation time is 0.2h. The parameters of the components are shown in Table 2 below:

[0187] Table 2 Component reliability parameters

[0188]

[0189] Taking the historical data of wind speed and light intensity in a certain place for one year, different types of DG devices are connected at the locations of nodes 1 and 2 respectively, including two wind turbines with a cut-in wind speed of 2.5m / s, a cut-out wind speed of 25m / s, a rated wind speed of 12.5m / s, and an installed capacity of 2.1MW; the photoelectric conversion efficiency of the panels of the two photovoltaic generators is 0.15, and the installed capacity is 1.1MW; the rated capacity of the wind-solar complementary power generation system is 1.1MW, and the capacity configuration ratio is 1:1.

[0190] Solution 1: No access to DG;

[0191] Solution 2: Connect wind turbines at nodes 1 and 2;

[0192] Solution 3: Connect photovoltaic generators at nodes 1 and 2;

[0193] Solution 4: Connect wind-solar complementary power generation devices at nodes 1 and 2.

[0194] The calculation results of reliability index of some load points in the system are shown in Table 3 below:

[0195] Table 3 Reliability indicators of some load points after different DG types are connected to the distribution network

[0196]

[0197]

[0198] As shown in Figure 6(a) and Figure 6(b), the reliability indexes of the load points under the four schemes are shown. The results in Figure 6(a), Figure 6(b) and Tables 4 and 5 show that the load points in island 1 include LP11, LP12 and LP13, and the load points in island 2 include LP19 to 23; compared with the calculation results of scheme 1, it can be seen that the access of DG has almost no effect on the reliability index of the load points outside the island range, while for the load points within the island range, it can be seen that the reliability indexes of these load points are reduced. This is because when a system component fails, DG can supply power to the load points within the island range, so the power supply reliability is improved. Among them, the power supply reliability of the load point closer to the DG access point is higher than that of the load point far away from the DG.

[0199] Table 4 System reliability indicators when different DG types are connected

[0200]

[0201] Table 5 Reliability influence coefficients after different DG types are connected

[0202]

[0203] Table 4 shows the reliability index of the system when different types of DG are connected. From the calculation results in the table, it can be seen that the reliability index of the system with DG access is compared with the reliability index without DG access. It is found that the SAIFI of the system DG 、SAIDI DG EENS DG The average power supply availability rate (ASAI) DG The increase from 99.9498% to 99.9598% indicates that the reliability of the system has been improved. This is because after the distributed power supply is connected, when the system components fail, power can be supplied to the load points in the island, thereby forming an island operation, thus improving the reliability of the system. Comparing the reliability indicators of the system with wind turbines and photovoltaic generators of the same capacity, it is found that the SAIFI of the system is DG The indicator dropped from 2.6052 times / year to 2.5746 times / year, SAIDI DG The indicator dropped from 5.0263h / year to 4.8881h / year. This is because wind and solar power output is affected by natural conditions. Although it has strong randomness in time and space, wind turbines have shorter downtime, longer operating time and smaller fluctuations than photovoltaic power generation. Therefore, wind turbines have a better effect on improving reliability than photovoltaic power generation.

[0204] The wind-solar hybrid power generation system established by the present invention is compared with schemes 1, 2 and 3. When the wind-solar hybrid power generation system of the present invention is adopted, the power supply availability rate reaches a maximum of 99.9698%, SAIFI DG 、SAIDI DG All indicators declined slightly, with EENS DG The index drops most obviously, indicating that the reliability of the system is the best when the complementary power generation system in the present invention is used. The reliability impact coefficient can intuitively reflect the impact of DG access on system reliability. From Table 5, it can be intuitively seen that the impact coefficient of Scheme 4 is larger than that of Scheme 2 and Scheme 3. It can be seen that compared with separate wind and photovoltaic power generation, the wind-solar combined power generation system established by the present invention has the highest power supply reliability.

Claims

1. A reliability assessment method for a distribution network containing distributed power sources, characterized in that The following steps are involved: Step 1: Use the kernel density estimation method to construct a probability distribution model of wind turbine and photovoltaic generator output; Step 2: Based on the Copula theory, select the Copula function with the best fitting effect on wind and solar power output, and build a probability distribution model for wind and solar power combined output; Step 3: Introduce new indicators based on the traditional distribution network reliability evaluation indicators to evaluate the reliability of the new distribution network containing DG; Step 4: Monte Carlo sampling method is used to calculate the reliability of the system after different DG types are connected to the distribution network; In step 1, a probability distribution model of the output of wind turbines and photovoltaic generators is constructed, as follows: The marginal distribution of wind turbine and photovoltaic generator output power is determined by nonparametric kernel density estimation method, and P is set w1 ,P w2 ,…,P wn and P v1 ,P v2 ,…,P vn P is taken from discrete calculations. w and P v Sample space, where: P w1 ,P w2 ,…,P wn is the sequence sample of the output power of n wind turbines; P v1 ,P v2 ,…,P vn is the sequence sample of the output power of n photovoltaic generators; P w , P v They represent the sample space of the output power of wind turbines and photovoltaic generators respectively; Substituting it into the kernel density estimation formula, the probability density function f of the output power of wind turbines and photovoltaic generators is estimated. WT (P w ) and f PV (P v ); f WT (P w ) and f PV (P v ) is integrated to obtain the probability distribution function of the output power of wind turbines and photovoltaic generators: F WT (P w ) is the probability distribution function of the wind turbine output power; F PV (P v ) is the probability distribution function of the output power of the photovoltaic generator; The step 2 specifically includes the following steps: Step 2.1: Obtain relevant data on the output of wind turbines and photovoltaic generators; Step 2.2: Considering that wind speed and light intensity have a certain correlation within a certain range, based on the Copula theory, the Copula function with the best fit is selected as the connection function between wind and solar outputs; Step 2.3: Based on the distribution function F of wind and solar output established in step 1 WT (P w ) and F PV (P v ), combined with the Frank-Copula function in step 2.2, the wind-solar combined output distribution function is obtained; In step 2, in order to select the Copula function with the best fitting effect for the wind and solar power output, the Frank-Copula function has the best fitting effect, and its probability distribution expression is as follows: In the formula, θ represents the correlation coefficient between random variables u and v; C(u,v) is the probability distribution expression of the Frank-Copula function; e -θu and e -θv is the generator of the Frank-Copula function; The distribution function expression of wind-solar combined output is: Where: θ is F(P w ) and F(P v ), if θ>0, the wind and solar outputs are positively correlated; if θ<0, the wind and solar outputs are negatively correlated; if θ→0, the wind and solar outputs tend to be independent of each other; are the generators of the Frank-Copula function after substituting wind and solar output respectively; In step 3, a new distribution network reliability evaluation index including DG is introduced: The reliability index of the load point is the average failure rate λ i , Annual average power outage time U i and the mean outage duration r i Three are used to represent that after DG joins, the DG outage probability λ needs to be considered DG and average power supply time U′ DG ; The probability of DG outage λ DG : Where, T is the total running time; t{p t =0} is the time set in which the DG output is 0 during the total operating time; p t is the probability of DG working normally; t is the time series; Average power supply time of DG: In the formula, Year is the simulation period; P Li is the load power at load point i; P DG is the output power of DG; N is the number of normal operations; represents the power supply time set during island operation; t j Indicates the power supply time during island operation; R j Indicates the island operation range; i is the number of the load point; j is the number of the load point within the island range; The reliability evaluation indicators of the system after connecting to DG are as follows: (1): After DG is connected to the distribution network, DG itself also has a failure rate λ DG Therefore, the average power outage frequency index SAIFI of the system after DG access DG for: Where N i is the number of users at load point i; P DG∩N is the probability of simultaneous failure of DG and distribution network in the island; i represents the failure rate of load point i within the island range; (2) Average power outage duration index SAIDI of the system after DG access DG : U i represents the average power outage time of load point i within the island range; U i ' ,DG represents the average power outage time of load point i containing DG within the island range; (3): ASAI, the power supply reliability index of the system after connecting to DG DG : (4): Since the expected power shortage is determined by U i and load power P Li Therefore, the expected power shortage after adding DG is EENS DG for: (5): The influence coefficient K of DG on the average power outage frequency index SAIFI : Where △SAIFI is the average power outage frequency index change of the system before and after DG access; K SAIFI Indicates that the access rated capacity is P E The change index of the SAIFI of the system after DG can reflect the impact of DG on the system SAIFI; SAIFI represents the average power outage frequency index of the system; (6): The influence coefficient K of DG on the average power outage duration index SAIDI : Where △SAIDI is the average power outage duration index change of the system before and after DG access; P E is the rated capacity of DG; K SAIDI Indicates that the access rated capacity is P E The change index of the SAIDI of the system after DG can reflect the impact of DG on the SAIDI of the system; (7): DG's influence coefficient on power supply reliability K ASAI : Where, △ASAI is the change in the power supply reliability index of the system before and after DG access; K ASAI Indicates that the access rated capacity is P E The change index of the ASAI of the system after DG can reflect the impact of DG on the ASAI of the system; (8): The influence coefficient K of DG on the expected power shortage EENS : Where △EENS is the expected power shortage index change of the system before and after DG access; K EENS Indicates that the access rated capacity is P E The change index of the system's EENS after DG can reflect the impact of DG on the system's EENS.

2. A method for evaluating the reliability of a distribution network containing distributed power sources according to claim 1, characterized in that: The step 4 comprises the following steps: S4.1: Read the basic parameters of the distribution network, set the simulation period N, and initialize the data and simulation time; S4.2: Number each component and sample the component status through Monte Carlo simulation, that is, generate a random number at (0,1), calculate the normal working time TTF and fault repair time TTR of each component respectively, and form the operating state sequence of each component in the simulation; obtain the fault isolation and switching time T res ; T TTF =-1 / λ·lnσ; T TTR =-1 / μ·lnσ; In the formula, λ and μ are the failure rate and repair rate of the component respectively; σ is a random number between (0,1); S4.3: Compare the normal working time of all components and find TTF min The smallest component i is set as the fault component; S4.4: After finding the location of the faulty component, determine whether it is located within the planned island, and determine the scope of island operation based on the location of the DG, and then classify the load: S4.5: For Class IV loads, the power outage time and normal power supply time are determined by the DG output and load demand in the island. Therefore, the DG output P during the island operation time is calculated based on the distribution function of the wind-solar combined output. DG,i (t) and load demand P L,i (t) conducting sampling; S4.6: Determine whether the output of the DG in the island meets the load demand in the island; if so, the load in the island continues to be supplied by the DG; Otherwise, load shedding is performed until power balance is achieved in the island. Each time a load shedding operation is performed, the corresponding outage time and number of power outages are accumulated. S4.7: Determine whether the simulation lifespan has been reached; If the simulation lifespan is reached, proceed to the next step; otherwise, return to step 2; S4.8: Determine again whether the load point i is located in the island. If so, calculate the reliability index of the load point and the system according to the evaluation index calculation formula including DG; otherwise, calculate according to the calculation method of traditional indicators.

3. A method for evaluating the reliability of a distribution network containing distributed power sources according to claim 2, characterized in that: In step 4.4, the loads are classified as follows: Class I load: load that is not affected by a main grid failure. Since it is not affected by the faulty components, it supplies power normally, so the total number of power outages and power outage duration remain unchanged. Class II load: load whose faulty components can be restored to normal working state after repair; Add the outage time TTR to the total outage time, and add 1 to the total outage times; Class III load: after a main grid fault occurs, the load can be restored to power through switch operation; the outage time is the fault isolation and switching time T res , add it to the total power outage time, and the total number of power outages increases by 1; Category IV load: load within the planned island.

Images