New energy grid-connected system operation risk assessment method, system, equipment and medium

Abstract

The invention discloses a new energy grid-connected system operation risk assessment method, system and device and a medium, and the method comprises the steps: obtaining wind power historical data, carrying out the clustering of the wind power historical data through a clustering algorithm, and fitting a time-varying probability density function; establishing a multivariable probability distribution function for each cluster by adopting a hybrid model to obtain statistical characteristics; constructing a time-varying probability density model satisfying short-term reliability evaluation by using a Gaussian mixture model, and fitting parameters of Gaussian mixture distribution by using a maximum posterior probability estimation method; integrating the evaluation model of the new energy unit into the operation risk evaluation of the electric field through a general expectation rule; and carrying out numerical value estimation on the risk index, and optimizing a calculation process by combining an importance sampling method to realize risk assessment of operation of the new energy grid-connected system. According to the method, the importance of the uncertainty of the short-term operation risk of the accurately modeled new energy to the risk assessment is verified, the calculation performance of the non-sequential Monte Carlo is improved, and the calculation of the new energy operation risk index is accelerated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of reliable operation and risk assessment of power systems, and particularly relates to a new energy grid-connected system operation risk assessment method, system, device and medium. BACKGROUND

[0002] The penetration rate of renewable energy represented by wind energy and solar energy in new power systems is continuously increasing, and advanced information communication technology, intelligent devices and big data analysis technology are widely used, which promotes the intelligentization and automation development of power systems. However, while new power systems improve energy efficiency and sustainability, they also bring more complex risk characteristics. Identifying, assessing and managing the risk characteristics of new power systems and establishing a reliable risk prevention and control system have become an important issue to ensure the safe and stable operation of the system and promote energy transformation. This requires multidisciplinary research methods, including big data analysis, artificial intelligence, probability and statistics models, and reliability assessment techniques, to comprehensively address the complex risk challenges faced by new power systems.

[0003] Existing risk assessment methods cover statistics, machine learning, physical modeling and complex networks. However, these methods still face new challenges such as spatiotemporal dynamic changes and enhanced nonlinear dependence when dealing with new power systems such as high proportion of new energy access and widespread application of distributed energy. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a new energy grid-connected system operation risk assessment method, system, device and medium to solve the problem that the continuous decarbonization of modern power systems leads to a significant increase in operational uncertainty, especially due to the large-scale integration of renewable energy generation, as the possible operating point space continues to expand, it is necessary to develop a new risk assessment method.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a new energy grid-connected system operation risk assessment method, comprising:

[0008] Obtaining wind power historical data, clustering the wind power historical data using a clustering algorithm, and fitting a time-varying probability density function;

[0009] For each cluster, a multivariate probability distribution function is established using a mixture model to obtain statistical characteristics;

[0010] A time-varying probability density model satisfying short-term reliability evaluation is constructed using a Gaussian mixture model, and parameters of the Gaussian mixture distribution are fitted using a maximum posterior probability estimation method;

[0011] The evaluation model of the new energy unit is integrated into the operation risk evaluation of the power field by a total expectation rule;

[0012] The numerical estimation of the risk index is carried out, and the importance sampling method is combined to optimize the calculation process, so that the risk evaluation of the new energy grid-connected system operation is realized.

[0013] As a preferred scheme of the new energy grid-connected system operation risk evaluation method, the operation risk basic model comprises:

[0014] The power system comprises N G Conventional power stations, N L Transmission lines, N W Wind turbine generators and N B Bus, define random vector X t =[X G,t ,X L,t ,X W,t ], wherein X G,t ,X L,t ,X W,t It is the uncertainty state caused by the unplanned outage of the conventional power generating unit, the transmission line and the wind power generating unit at t time; the operation risk index can be expressed as:

[0015] R t =E(H(X))=∫ Ω H(x)f t (x)dx

[0016] Wherein, R t Indicates the risk index, H() indicates the limit state function or the test function, E() indicates the expectation operator, Ω indicates the state space, f t () indicates the joint multivariate probability density function of random vector X, x indicates a specific state of X;

[0017] Assuming that the outage of the conventional power generating unit and the transmission line is independent of the variability of the new energy unit power, f t () can be expressed as:

[0018] f t (x)=f G,t (x G )f L,t (x L )f W,t (x W )

[0019] Wherein, fG,t (x G ) represents the probability density function of the conventional generator set, f L,t (x L ) represents the probability density function of the power transmission line, f W,t (x W ) represents the probability density function of the wind turbine generator set.

[0020] As a preferred scheme of the new energy grid-connected system operation risk assessment method, the fitting time-varying probability density function comprises:

[0021] The wind power output historical data of the target area are obtained and preprocessed;

[0022] The clustering algorithm is used to cluster and group the preprocessed data according to the time characteristics;

[0023] The time-varying probability density function is fitted for each cluster.

[0024] As a preferred scheme of the new energy grid-connected system operation risk assessment method, the fitting time-varying probability density function comprises:

[0025] The wind power output data in each cluster are analyzed to extract statistical characteristics;

[0026] The Gaussian mixture model is used to model the multi-modal characteristics of the new energy output;

[0027] The parameters of the Gaussian mixture model are solved by the maximum posterior estimation method.

[0028] As a preferred scheme of the new energy grid-connected system operation risk assessment method, the fitting time-varying probability density function comprises:

[0029] The Gaussian mixture model parameters of each cluster are input into the risk assessment framework;

[0030] The total operation risk index R considering different probability density functions of different clusters is obtained by the total expectation rule, and the formula is represented as:

[0031]

[0032] Wherein, λ c represents the weight of the cth cluster, f t,c (·) is similar to f t (·), and only the f W,t (·) is replaced by the probability density function of each cluster f W,c,t (·).

[0033] As a preferred scheme of the new energy grid-connected system operation risk assessment method, wherein: the numerical estimation of the risk index comprises:

[0034] The non-sequential Monte Carlo simulation method is adopted to sample and estimate the high-dimensional state space;

[0035] The system operation state is evaluated by the limit state function, which is expressed as:

[0036]

[0037] Wherein, S(x) represents the system state function, L represents the load demand within the lead time, P b represents the cumulative available power at the bus B, l b represents the load supplied to the bus B;

[0038] The failure sample ratio is counted, and the risk index is recalculated, which is expressed as:

[0039]

[0040] Wherein, is an independent and identically distributed sample drawn from f t,c (·), N s represents the total number of independent and identically distributed samples drawn, represents the random variable containing unit state, line state and wind power output, represents the importance weight.

[0041] As a preferred scheme of the new energy grid-connected system operation risk assessment method, wherein: the combined importance sampling method optimization calculation process comprises:

[0042] The probability density function of the clustering cluster c is The following transformation is used to obtain a new random variable χ w As follows:

[0043] χ w = Φ -1 (F w,c,t (g w,t |g w,t-1 ))

[0044] Wherein, φ(·) is the cumulative distribution function, F w,c,t (·) is the cumulative distribution function of f w,c,t (·); after transformation, the probability density function of θ w is changed by using cross-entropy optimization;

[0045] The cross-entropy parameter is obtained by applying the analytical rule, and the distorted φ* (·) transform the random variable χ w transformed back to the actual wind power random variable;

[0046] Efficient sampling is performed using the optimized distribution to reduce the computational load of Monte Carlo simulation.

[0047] In a second aspect, the present application provides a new energy grid-connected system operation risk assessment system, comprising:

[0048] A clustering fitting module is configured to obtain wind power historical data, cluster the wind power historical data using a clustering algorithm, and fit a time-varying probability density function;

[0049] A probability model construction module is configured to establish a multivariate probability distribution function using a mixture model for each cluster to obtain statistical characteristics, construct a time-varying probability density model that meets short-term reliability evaluation using a Gaussian mixture model, and fit the parameters of the Gaussian mixture distribution using a maximum posterior probability estimation method;

[0050] An integration module is configured to integrate the evaluation model of the new energy unit into the operation risk assessment of the power plant by a total expectation rule;

[0051] A risk assessment module is configured to numerically estimate the risk index and optimize the calculation process using an importance sampling method to realize risk assessment of the new energy grid-connected system operation.

[0052] In a third aspect, the present application provides an electronic device comprising a memory and a processor; the memory is configured to store computer executable instructions, and the processor realizes the steps of the new energy grid-connected system operation risk assessment method when executing the computer executable instructions.

[0053] In a fourth aspect, the present application provides a computer readable storage medium storing computer executable instructions, which realize the steps of the new energy grid-connected system operation risk assessment method when executed by a processor.

[0054] Compared with the prior art, the present application has the following beneficial effects: the present application provides a new energy grid-connected system operation risk assessment method, system, device and medium, a new energy early period uncertainty modeling method, uses clustering technology and a mixture model to construct a time-new energy generation related probability distribution, and calculates and evaluates a new energy operation risk index through a non-sequential Monte Carlo technique. The superiority of the method has been verified on a reliability test system. In particular, the present application verifies the importance of accurately modeling the short-term operation risk uncertainty of new energy for risk assessment; in addition, the importance sampling technology based on cross-entropy is improved, the calculation performance of the non-sequential Monte Carlo is improved, and the calculation of the new energy operation risk index is accelerated. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort based on these drawings.

[0056] Figure 1 The overall flow logic diagram of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0057] Figure 2 The historical clustering data diagram of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0058] Figure 3 The modified IEEE 24-bus RTD system diagram of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0059] Figure 4 The binary histogram of a given data set of one of the three clusters of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0060] Figure 5 The GMM diagram of a given data set when using maximum a posteriori estimation of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0061] Figure 6 The GMM diagram of a given data set when using maximum likelihood estimation method of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0062] Figure 7 The binary Gaussian approximation diagram of a given data set of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0063] Figure 8 The convergence diagram of the proposed method under three clustering scenarios of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application.

[0064] Figure 9 The three-cluster convergence effect diagram of the proposed method using IEEE 3-region RTS of the new energy grid-connected system operation risk assessment method according to an embodiment of the present application. DETAILED DESCRIPTION

[0065] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application are described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the protection scope of the present application.

[0066] Embodiment 1, refer to Figures 1-2 For an embodiment of the present application, a new energy grid-connected system operation risk assessment method is provided, as shown in Figure 1 specifically comprising the following steps:

[0067] S100: obtaining wind power historical data, clustering the wind power historical data by using a clustering algorithm, and fitting a time-varying probability density function;

[0068] S200: establishing a multivariate probability distribution function by using a mixture model for each cluster to obtain statistical characteristics;

[0069] S300: using a Gaussian mixture model to construct a time-varying probability density model meeting short-term reliability evaluation, and using a maximum posterior probability estimation method to fit the parameters of the Gaussian mixture distribution;

[0070] S400: integrating the evaluation model of the new energy unit into the operation risk assessment of the power plant by using a total expectation rule;

[0071] S500: numerically estimating the risk index, and combining an importance sampling method to optimize the calculation process, so as to realize the risk assessment of the new energy grid-connected system operation.

[0072] It should be noted that the present application provides a new energy grid-connected system operation risk assessment method, system, device and medium, a new energy early period uncertainty modeling method, uses clustering technology and a mixture model to construct a time-new energy generation related probability distribution, and calculates and evaluates the new energy operation risk index by using a non-sequential Monte Carlo technology. The superiority of the method is verified on a reliability test system. In particular, the present application verifies the importance of accurately modeling the short-term operation risk uncertainty of new energy for risk assessment. In addition, the importance sampling technology based on cross-entropy is improved, the calculation performance of the non-sequential Monte Carlo is improved, and the calculation of the new energy operation risk index is accelerated.

[0073] In the embodiments of the present application, the construction of the operation risk basic model includes:

[0074] The power system includes N G conventional power stations, N L power transmission lines, N W wind turbine generators and NB Barred bus, define random vector X t = [X G,t , X L,t , X W,t ], where X G,t , X L,t , X W,t are the uncertainty states caused by unscheduled outages of conventional generating units, transmission lines and wind power units at time t, respectively, where where n s,t is the number of available generating units in power plant s; where ζ l,t is the state of transmission line l, ζ l,t = 1 if the transmission line is available, and ζ l,t = 0 if the transmission line is in outage state; where g w,t is the wind of wind farm w; then the operational risk index can be expressed as:

[0075] R t = E(H(X)) = ∫ Ω H(x)f t (x)dx

[0076] where R t represents the risk index, H() represents the limit state function or test function, E() represents the expectation operator, Ω represents the state space, f t () represents the joint multivariate probability density function of random vector X, x represents a specific state of X; from the above formula, it can be seen that the selection of f t () will significantly affect the risk index, and more accurate estimation of f t () will more accurately assess the risk index.

[0077] Assuming that the outages of conventional generating units and transmission lines are independent of the variability of new energy unit power, f t () can be expressed as:

[0078] f t (x) = f G,t (x G )f L,t (x L )f W,t (x W )

[0079] where f G,t (x G ) represents the probability density function of conventional generating units, f L,t (x L ) represents the probability density function of transmission lines, and f W,t(x W ) denotes the probability density function of the wind turbine.

[0080] Specifically, for conventional generating units, it is assumed that each generating station s also includes N s identical generating units, and the failure events of these generating units are also assumed to be mutually independent, so that f G,t () can be modeled as the product of binomial distributions, and the formula is:

[0081]

[0082] where λ s represents the failure rate of the generating units in the generating station s, Δt represents the advance time length (usually 1 hour), and p s (Δt) represents the outage replacement rate.

[0083] Specifically, similar to the conventional generating units, it is assumed that the line outages are independent of each other, and f L,t () is represented by the product of Bernoulli distributions as:

[0084]

[0085] where λ l represents the failure rate of the transmission line l.

[0086] Specifically, for new energy units, it is assumed that the spatial correlation between wind farms can be ignored, which means that f W,t (·) can be represented as:

[0087]

[0088] where f W,t (g w,t ) denotes the probability density function of the new energy station w at time t.

[0089] In the embodiments of the present application, modeling the new energy output uncertainty includes:

[0090] The short-term uncertainty of the new energy output at a specific date and time is modeled by Gaussian mixture models (GMMs), and an additional random variable X W,t is defined. In order to study the time dependence, another random vector X W,t-1 is considered. It should be pointed out that t and t-1 are both defined for a specific time of a specific date. g w,t and g w,t-1 have strong correlation, where g w,t-1 is the wind of the wind farm w at the current moment. Therefore, g w,tThe uncertainty is best represented by the conditional probability density function f w,t (g w,t | g w,t-1 ). Using the concept of conditional probability, f w,t (g w,t | g w,t-1 ) can be obtained from f w,t (g w,t-1 , g w,t ), which represents the joint density on g w,t-1 and g w,t . If g w,t-1 is deterministically known, which is usually the case during power system operation, f w,t (g w,t | g w,t-1 ) can be directly used to evaluate the risk metrics, otherwise, f w,t (g w,t ) can be obtained by marginalizing g w,t-1 :

[0091] f w,t (g w,t ) = ∫f w,t (g w,t-1 , g w,t ) dg w,t-1

[0092] f w,t (g w,t-1 , g w,t ) is defined for two specific hours, therefore this probability density function should be constructed using the corresponding wind data of these specific hours. For example, if t-1 and t are 14:00 and 15:00 of January 3rd, respectively, all the historical data of these two hours of the day will be employed to estimate the probability density function.

[0093] In embodiments of the present application, clustering the historical data comprises:

[0094] By using k-means clustering, the historical data of a specific month (i.e., January) is grouped into clusters. Then, for each cluster, the required joint PDF is estimated using the historical data of the two specific time periods 14:00 and 15:00. The joint PDF of each cluster is associated with the probability λ c obtained by k-means clustering. This approach ensures that there is enough data available to fit the PDF, as shown in Figure 2 where T h is the current time instant, and [T h , T h+1 ] is the time interval for which the risk is evaluated.

[0095] In the embodiments of the present application, the mixed model is used to solve the probability density function of new energy output, which includes:

[0096] After obtaining sufficient data for each cluster at a specific hour, different methods can be used to estimate f w,c,t (g w,t-1 ,g w,t ). Using MMs, the probability density function of new energy output in two specific hours is:

[0097]

[0098] where Ψ k (·) represents the binary probability density function of parameter θ k , K represents the mixed function, π k represents the kth mixed proportion, and the following conditions are met:

[0099]

[0100] 0≤π k ≤1

[0101] By incorporating the probability density function of new energy output into the risk assessment framework, the following can be obtained:

[0102]

[0103] where φ(g w,t-1 ,g w,t |μ k ,Σ k ) represents a Gaussian PDF with mean μ k and covariance ∑ k . The GMM parameters are grouped as π=[π1,π2,...,π K ] T , and Σ k is a three-dimensional matrix ∑ k of covariance matrices k.

[0104] In the embodiments of the present application, the maximum a posteriori estimation method is used to solve the Gaussian mixed model parameters, which includes:

[0105] The popular technique for determining K includes the split-and-merge method, cross-validation, and Akaike information criterion (AIC). For a given K, the remaining GMM parameters can then be estimated by maximizing the log-likelihood. For a given set of bivariate wind data {g w,t-1,i ,g w,t ,i} where i={1,2,...,N}, the GMM parameters can be estimated using the MLE method as:

[0106]

[0107] It should be noted that the MLE method of estimating GMM parameters has two main disadvantages. First, it is prone to overfitting due to limited data volume; second, it can be singular due to variance collapse problem. Therefore, the MAP method is used to avoid these disadvantages. Using the MAP method, the above formula is modified as:

[0108]

[0109] where f D (π) and f NIW (μ,Σ) are the prior distributions of GMM parameters, f D (π) is Dirichlet distribution, and f NIW (μ,Σ) is normal-inverse-Wishart distribution. These prior distributions are used to regularize parameter fitting, thereby avoiding overfitting and singularity.

[0110] Using the expectation maximization (EM) method, the MAP estimation in the above formula can be obtained as:

[0111]

[0112] where, g i =[g w,t-1,i ,g w,t,i ] T , a k is a scalar parameter of Dirichlet distribution, m0, v0, κ0, S0 are parameters of normal-inverse-Wishart distribution, m0 is a D × 1 vector parameter, v0 and κ0 are scalar parameters, and S0 is a D × D matrix parameter, and D is the dimension of data.

[0113] After obtaining the joint PDF, the conditional probability density function of the clustering result of each can also be obtained. Due to the characteristics of Gaussian probability density function, this conditional probability density function is also a univariate Gaussian mixture model, and it is applied to the operation reliability evaluation.

[0114] In the embodiments of the present application, integrating the evaluation model of the new energy unit into the operation risk evaluation of the power plant includes:

[0115] The risk of the operation risk index formula is defined for a specific PDF f t (·), and since there are multiple PDFs for multiple clusters, the formula needs to be modified to obtain the total operation risk index R considering different probability density functions of different clusters by using the total expectation rule, and the formula is represented as:

[0116]

[0117] where λ c is the weight of the c-th cluster, f t,c (·) is the probability density function of the c-th cluster. t (·) is similar to f W,t (·), except that f W,c,t (·) is replaced by the probability density function of each cluster f

[0118] In the embodiments of the present application, the numerical estimation of the risk index comprises:

[0119] Due to the large number of states in the state space Ω and the high dimension of the integral, it is difficult to calculate the above total operating risk index by using an analytical method, and therefore a rough NSMCS is used for estimation:

[0120]

[0121] where, is an independent and identically distributed sample drawn from f t,c (·), N s is the total number of samples. The LSFH(·) is defined as:

[0122]

[0123] where S(x) represents a system state function, L represents a load demand within a lead time, P b represents a cumulative available power at a bus B, and l b represents a load supplied to the bus B.

[0124] Due to the low probability of failure during the operation of the power system, most samples correspond to H(·) = 0. Another joint The risk index is calculated using the following formula:

[0125]

[0126] where, is an independent and identically distributed sample drawn from f t,c (·), N s represents the total number of independent and identically distributed samples drawn, represents a random variable including a unit state, a line state, and a wind power output, represents an importance weight.

[0127] The IID sample can be drawn from f , which can be rewritten as:

[0128]

[0129] In the embodiments of the present application, the optimization calculation process combined with the importance sampling method comprises:

[0130] It should be noted that, Also known as importance sampling density, it can be obtained using the widely used cross entropy (CE) optimization. For CE optimization, if the probability density function belongs to a certain distribution group, a closed-form analytical update rule can be used. And The PDF of and can use a closed-form analytical update rule because these PDFs belong to the exponential distribution family. However, for the Gaussian mixture model of new energy power, such a closed-form analytical solution is not available. To solve this problem, a transformation strategy is adopted:

[0131] Consider the probability density function of the GMM PDF of the new energy power plant w for the cluster c Use the following transformation to obtain a new random variable χ w As follows:

[0132] χ w =Φ -1 (F w,c,t (g w,t |g w,t-1 ))

[0133] Where φ(·) is the cumulative distribution function of F w,c,t (·) is the cumulative distribution function of f w,c,t (·); after transformation, the cross-entropy optimization changes to the probability density function of θ w ;

[0134] Since the PDF of χ w belongs to the exponential distribution family, an analytical rule can be applied to obtain the CE parameter. When calculating LSFH(·) through direct current optimal power flow (DC OPF), through the inverse operation of the above formula, the distorted φ * (·) is used to transform the random variable χ w back to the actual wind power random variable. Therefore, χ w serves as an alternative random variable to change the GMM. The complete CE algorithm is described as Algorithm 1. After obtaining through Algorithm 1, the NSMCS is used to estimate the risk index.

[0135] Embodiment 2, referring to Figures 3-9 , based on the previous embodiment, the present embodiment provides application examples of new energy grid-connected system operation risk assessment methods, systems, devices and media, which verify and illustrate the technical effects adopted in the present method.

[0136] This example is demonstrated based on MATLAB software simulation as benchmark results, through the use of modified IEEE 24-bus RTS system and modified IEEE 73-node 3-area RTS system as shown in Figure 3 The example analysis is carried out on the original IEEE 24-bus RTS, in which a total capacity of 1000 MW wind power station is integrated in 19 nodes, and the 155 MW traditional generator set in node 16 is removed. The new energy penetration rate of the system is 23.05%. The real data of Sotavento new energy power station in Spain for 10 years is used. All simulations are carried out in January, and the operation risk assessment is carried out with 1 hour as the lead time, and the time interval of the lead time is set as 3:00-4:00. The load is set as the peak value, the stop criterion of NCSMS is set as 5%, and the number of k-means clustering is set as 3.

[0137] Figure 4 a binary histogram of the data set of a particular cluster is depicted, Figure 5 a GMM obtained for this cluster using the MAP approach to obtain the parameters is depicted. By comparing Figure 4 and Figure 5 it can be concluded that the GMM accurately captures the variation of the wind power production in the lead time. From a close observation of Figure 5 two conclusions can be drawn. First, for different initial wind power, the PDF of the wind power in the lead time has a clear difference. Second, the multimodality of the PDF is apparent. For example, when the initial wind power is in the interval [0.2, 0.4), there are three different modes of the wind power PDF in the lead time. The impact of the multimodality on the operational risk index will be discussed in the next subsection. Figure 6 a GMM model representing the same data set, however the MLE approach is used to obtain the parameters. Figure 6 Some of the Gaussian components in Figure 7 a bivariate Gaussian PDF is estimated. From this figure, it can be seen that this PDF cannot represent the distribution of the wind power in the lead time realistically.

[0138] The method proposed in this example is compared with another method (Method B), in which a bivariate Gaussian distribution as shown in Figure 7 is used to model the wind power PDF. The results are shown in Table 1. The results show that there is a clear difference in the risk obtained by the two methods. The risk index calculated by the method proposed in this example is higher than the result of Method B at all initial wind power values. By studying Figure 5 and Figure 7 , it can be inferred that the reason behind this observation is that Method B fails to capture the multiple modes of the wind power PDF, some of which occur at lower wind values. For example, from Figure 8It can be seen that there is a pattern when the next hour wind power is around 0.2 p.u. Since these patterns are missing, method B assumes higher wind power than the actual value, thus overestimating the reliability of the power system. In the proposed method, as these patterns are captured, more number of samples are extracted from the low wind power state during NSMCS, leading to higher risk indices.

[0139] Table 1: Operational risk indices for IEEE 24-Bus RTS system

[0140]

[0141] The location of the new energy power plant also has an impact on the operational risk due to different transmission line capacities of different nodes. From Table 2, it can be seen that the operational risk indices of different nodes are different, and the operational risk of node 4 is the highest because the capacity of the transmission line connected to it is the smallest.

[0142] Table 2: Operational risk indices for different locations of new energy power plant

[0143]

[0144] Therefore, in this case, the output of new energy is greatly constrained, for bus 10, the total capacity of the transmission line is 1545 MW, therefore, the operational risk index of this bus is lower than that of bus 4, finally, for bus 19, the available transmission capacity of the wind farm is 1000 MW, therefore, the operational risk index of this bus is between the operational risk indices of bus 4 and bus 10. This impact is more obvious when the wind power PDF of the lead period is more biased towards the maximum capacity, these results highlight the impact of the transmission system on the operational risk index.

[0145] Similar to the previous example study, the 155 MW conventional generator set on bus 16 of each zone is removed and a 500 MW wind farm is installed on bus 19 of each zone. The load is set to the peak value of 8550 MW. The probability model of wind power is similar to the previous study. Table 3 lists the risk indices of this test system. Compared with the IEEE 24-node RTS, the risk indices are significantly reduced because the interconnection of the three zones improves the overall reliability of the system.

[0146] Figure 9 The calculation performance of the method for this test system when the initial wind power is set to 0.5 p.u. is described. In this case, the maximum number of samples for NSMCS is 10000. Therefore, the calculation burden is expected to be higher than that of the IEEE 24-node RTS. However, it is still lower than the original NSMCS.

[0147] Table 3: Operational risk indices for IEEE 3-zone RTS

[0148] New energy initial output (p.u.) Risk index of the proposed method 0.1 8.7893 x 10 -11 ]]> 0.3 3.7503 x 10 -11 ]]> 0.5 2.1835 x 10 -11 ]]> 0.8 1.2957 x 10 -11 ]] 1.0 7.3894 x 10 -12 ]]

[0149] Therefore, according to the above analysis, the present application provides a new energy grid-connected system operation risk assessment method, system, device and medium, a new energy early stage uncertainty modeling method, a clustering technology and a mixed model are used to construct a time-new energy generation related probability distribution, and a non-sequential Monte Carlo technology is used to calculate and evaluate the new energy operation risk index. The superiority of the method is verified on the reliability test system. In particular, the present application verifies the importance of accurately modeling the short-term operation risk uncertainty of new energy for risk assessment; in addition, the importance sampling technology based on cross-entropy is improved, the calculation performance of the non-sequential Monte Carlo is improved, and the calculation of the new energy operation risk index is accelerated.

[0150] In embodiment 3, a new energy grid-connected system operation risk assessment system is provided, comprising:

[0151] A clustering fitting module is configured to obtain wind power historical data, cluster the wind power historical data by using a clustering algorithm, and fit a time-varying probability density function;

[0152] A probability model construction module is configured to establish a multivariate probability distribution function by using a mixed model for each cluster to obtain statistical characteristics; a Gaussian mixed model is used to construct a time-varying probability density model that meets the requirements of short-term reliability evaluation, and a maximum posterior probability estimation method is used to fit the parameters of the Gaussian mixed distribution;

[0153] An integration module is configured to integrate the evaluation model of the new energy unit into the operation risk assessment of the power plant by using the total expectation rule;

[0154] A risk assessment module is configured to numerically estimate the risk index and optimize the calculation process by using the importance sampling method to realize the risk assessment of the new energy grid-connected system operation.

[0155] It should be noted that the technical scheme of the new energy grid-connected system operation risk assessment system belongs to the same concept as the technical scheme of the new energy grid-connected system operation risk assessment method described above. The technical scheme of the new energy grid-connected system operation risk assessment system in this embodiment is not described in detail, and the description of the technical scheme of the new energy grid-connected system operation risk assessment method can be referred to.

[0156] The above-mentioned various modules can be embedded in or independent of the processor in the electronic device in hardware form, or can be stored in the memory in the electronic device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned various modules.

[0157] The embodiment also provides an electronic device including a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The communication interface of the electronic device is configured to perform wired or wireless communication with an external terminal. The wireless communication can be achieved through WIFI, a carrier network, NFC (Near Field Communication) or other technologies. The computer program is executed by the processor to implement the new energy grid-connected system operation risk assessment method. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen. The input device of the electronic device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the electronic device, or an external keyboard, touchpad or mouse, etc.

[0158] The embodiment also provides a computer readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the method proposed in the above embodiment.

[0159] The storage medium proposed in the embodiment belongs to the same inventive concept as the method proposed in the above embodiment. The technical details not described in the embodiment can be referred to the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0160] From the above description about the embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software and necessary universal hardware, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a floppy disk, a ROM, a RAM, a FLASH, a hard disk or an optical disc, etc., and includes a number of instructions to make an electronic device (which can be a personal computer, a server, or a network device, etc.) execute the method of the embodiments of the present application.

[0161] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application rather than limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and all of them should be covered in the scope of the claims of the present application.

Claims

1. A method for assessing the operational risk of a new energy grid-connected system, characterized in that: include: Acquiring historical wind power data, clustering the historical wind power data using a clustering algorithm, and fitting a time-varying probability density function; A multivariate probability distribution function is established for each cluster using a mixture model to obtain statistical characteristics; A time-varying probability density model that satisfies short-term reliability assessment is constructed using a Gaussian mixture model, and the maximum a posteriori probability estimation method is used to fit the parameters of the Gaussian mixture distribution. The evaluation model of new energy units is integrated into the operational risk assessment of the power plant through the total expectation law; The risk index is numerically estimated and the calculation process is optimized in combination with the importance sampling method to realize the risk assessment of the operation of the new energy grid-connected system.

2. The method for evaluating the operation risk of a new energy grid-connected system according to claim 1, wherein: The basic operational risk model includes: The power system includes N G Conventional power stations, N L Transmission lines, N W Wind turbines and N B Generate a random vector X t =[X G,t ,X L,t ,X W,t ], where X G,t ,X L,t ,X W,t are the uncertain states caused by the unplanned outage of conventional generators, transmission lines and wind turbines at time t respectively; the operation risk index can be expressed as: R t =E(H(X))=∫ Ω H(x)f t (x)dx Among them, R t represents the risk index, H() represents the limit state function or test function, E() represents the expectation operator, Ω represents the state space, f t () represents the joint multivariate probability density function of the random vector X, and x represents a specific state of X; Assuming that the outage of conventional generators and transmission lines is independent of the variability of the power of renewable energy generators, then f t () can be expressed as: f t (x)=f G,t (x G )f L,t (x L )f W,t (x W ) Among them, f G,t (x G ) represents the probability density function of conventional generator sets, f L,t (x L ) represents the probability density function of the transmission line, f W,t (x W ) represents the probability density function of the wind turbine generator set.

3. The method for evaluating the operation risk of a new energy grid-connected system according to claim 2, wherein: The fitting time-varying probability density function comprises: Obtain historical wind power output data for the target area and perform preprocessing; Using a clustering algorithm to cluster and group the pre-processed data according to time characteristics; A time-varying probability density function is fitted for each cluster separately.

4. The method for evaluating the operation risk of a new energy grid-connected system according to claim 3, wherein: The use of a hybrid model to establish a multivariate probability distribution function includes: Analyze the wind power output data within each cluster and extract statistical features; Use Gaussian mixture models to model the multimodal characteristics of renewable energy output; The parameters of the Gaussian mixture model are solved by the maximum a posteriori estimation method.

5. The method for evaluating the operation risk of a new energy grid-connected system according to claim 4, wherein: Integrating the evaluation model of new energy units into the operational risk assessment of the power plant includes: The Gaussian mixture model parameters of each cluster are input into the risk assessment framework; The total operation risk index R of different probability density functions considering different clusters is obtained by the total expectation rule. The formula is expressed as: Among them, λ c represents the weight of the cth cluster, f t,c (·) and f t (·) is similar, just change f W,t (·) is replaced by each cluster f W,c,t The probability density function of (·).

6. The method for evaluating the operation risk of a new energy grid-connected system according to claim 5, wherein: The numerical estimation of the risk index comprises: The non-sequential Monte Carlo simulation method is used to sample and estimate the high-dimensional state space; The system operating state is evaluated by the limit state function, which is expressed as: Where S(x) represents the system state function, L represents the load demand in the lead time, and P b represents the cumulative available power generation at bus B, l b Indicates the load supplied to bus B; Count the proportion of failed samples and recalculate the risk index. The formula is: in, It is from f t,c (·) is an independent and identically distributed sample drawn from s represents the total number of independent and identically distributed samples drawn, represents the random variables including unit status, line status, and wind power output, Represents importance weight.

7. The method for evaluating the operation risk of a new energy grid-connected system according to claim 6, wherein: The optimization calculation process combined with the importance sampling method includes: The probability density function for cluster c is Use the following transformation to obtain the new random variable χ w ,as follows: where φ(·) is The cumulative distribution function, F w,c,t (·) is f w,c,t The cumulative distribution function of (·) is converted to θ using cross entropy optimization. w The probability density function of Apply the analytical rules to obtain the cross entropy parameters, and use the distorted φ through the inverse operation * (·) The random variable χ w Transform back to the actual wind power random variable; The optimized distribution is used for efficient sampling to reduce the computational complexity of Monte Carlo simulation.

8. A new energy grid-connected system operation risk assessment system, applying the new energy grid-connected system operation risk assessment method according to any one of claims 1 to 7, characterized in that: include: A clustering fitting module is used to obtain historical wind power data, cluster the historical wind power data using a clustering algorithm, and fit a time-varying probability density function; The probability model building module is used to establish a multivariate probability distribution function for each cluster using a mixture model to obtain statistical characteristics; a time-varying probability density model that meets the requirements of short-term reliability assessment is constructed using a Gaussian mixture model, and the maximum a posteriori probability estimation method is used to fit the parameters of the Gaussian mixture distribution; Integration module, used to integrate the evaluation model of new energy units into the operational risk assessment of the power plant through the total expectation law; The risk assessment module is used to numerically estimate the risk index and optimize the calculation process in combination with the importance sampling method to achieve risk assessment of the operation of the new energy grid-connected system.

9. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store computer-executable instructions, and when the processor executes the computer-executable instructions, it implements the steps of the new energy grid-connected system operation risk assessment method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: When the computer executable instructions are executed by a processor, the steps of the new energy grid-connected system operation risk assessment method according to any one of claims 1 to 7 are implemented.

Images