Comprehensive confidence capacity assessment method for offshore wind farms with complex topology

Abstract

The present invention provides a comprehensive confidence capacity assessment method for offshore wind farms with complex topologies. The method collects exploration or meteorological wind speed data, analyzes the total output of the wind farm, and obtains a Copula joint distribution function. Based on the main wind turbine output scenarios within each wind farm, the component outage events are enumerated and a correction calculation is performed on the output distribution within each wind farm using a linear optimal power flow model. The total corrected output distribution of the wind farm group is obtained using the Copula joint distribution function. Based on the total corrected output distribution of the wind farm group and the outage table of the transmission system, the comprehensive confidence capacity of the system is obtained, thereby evaluating the continuity and adequacy of the supply load of the offshore wind farm group. The present invention takes into account the internal topological grid and circuit breakers of the wind farm, the transmission system of the wind farm group, the failure process and common cause outage of all components, and takes into account both accuracy and efficiency, which can provide new technical support for the planning and development of offshore wind energy resources.

Description

Technical Field

[0001] The present invention relates to a wind power confidence capacity assessment method, and in particular to a comprehensive confidence capacity assessment method for an offshore wind farm group with a complex topology structure. Background Art

[0002] With the continuous advancement of offshore wind power planning and construction, to ensure the smooth implementation of large-scale offshore wind power grid integration, it is necessary to evaluate its actual operating equivalent capacity, taking into account random wind speeds and reliability issues, that is, to perform equivalent calculations of its confidence capacity. Furthermore, as offshore wind power continues to develop towards large-scale and deep-sea development, wind farms have numerous wind turbines, and their collection system topologies are complex, likely including numerous ring networks. Due to the complex topology and flexible switch configurations within offshore wind farm clusters, analytical methods are often used to assess the reliability of collection systems. However, due to the potential for ring networks, the collection system topology is complex, with complex structures and constraints. Failure consequence analysis, topology reconstruction, and turbine removal after component outages are unique issues not encountered in traditional radial topologies and cannot be addressed using existing reliability assessment methods. Furthermore, reliability assessment of collection systems within multiple wind farms requires consideration of a reduced-dimensional wind speed distribution, while overall reliability assessment of a wind farm cluster requires consideration of the joint distribution of multiple variables. Without processing wind speed data and correlation analysis of wind farm-level data, computational efficiency is prohibitively time-consuming and computationally inefficient. Summary of the Invention

[0003] In view of this, the purpose of the invention is to provide a comprehensive confidence capacity assessment method for offshore wind farms with complex topology structures, which can accurately perform reliability assessments on any wind turbine distribution, any topology, and any switch configuration of the offshore wind farm collection system, and can perform reliability assessments on the transmission system. It also uses clustering technology and Copula functions to greatly reduce the amount of calculation, and ultimately obtains a comprehensive confidence capacity result, providing important technical support for the large-scale planning and development of offshore wind power.

[0004] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] The comprehensive confidence capacity assessment method for offshore wind farms with complex topologies includes the following steps:

[0006] Step 1: Collect exploration or meteorological wind speed data, generate wind turbine output data for the wind farm group based on the wake model, perform dimensionality reduction and feature extraction on the wind turbine output data for the wind farm group, and obtain the main wind turbine output scenarios within each wind farm;

[0007] Step 2: Analyze the total output of the wind farm to obtain the wind farm joint output distribution and the marginal output distribution of each wind farm. Perform maximum likelihood estimation on the wind farm joint output distribution and the marginal output distribution of each wind farm to obtain the Copula joint distribution function.

[0008] Step 3: Based on the main wind turbine output scenarios within each wind farm obtained in Step 1, a linear optimal power flow model is used to calculate and obtain the corrected output distribution of each discrete wind farm. Continuous fitting is then performed. Based on the Copula joint distribution function, a joint distribution is established between the fitted corrected output distributions of each wind farm to generate the total output distribution of the wind farm group. Based on the total output distribution of the wind farm group and the outage schedule of the transmission system, the grid-connected output distribution is obtained.

[0009] Step 4: Obtain a comprehensive confidence capacity based on the grid-connected output distribution obtained in step 3; and evaluate the sustainability and adequacy of the supply load of the offshore wind farm group based on the comprehensive confidence capacity.

[0010] Furthermore, the specific process of step 1 is:

[0011] Exploration or meteorological wind speed data is collected, and the wind turbine output distribution of the wind farm group is generated based on the wind turbine location, wind turbine output model and wake model. Then, the wind turbine output data of the wind farm group is reduced in dimension and clustered using K-means to extract the main wind turbine output scenarios within each wind farm.

[0012] Furthermore, the specific process of step 2 is:

[0013] The maximum likelihood estimation of the continuous output distribution is performed on the output distribution of each wind farm and the output distribution of the wind farm group. Then, the distribution with the smallest fitting error among the truncated Gaussian distribution, Beta distribution, and Gamma distribution is selected as the continuous distribution function. Based on the continuous distribution function, the marginal output distribution fitting function of each wind farm is obtained.

[0014] According to the marginal output distribution fitting function of each wind farm, the Copula joint distribution function of the output correlation of each wind farm is obtained.

[0015] Furthermore, the continuous distribution function is obtained through the following process:

[0016] 1) When the distribution mode is truncated Gaussian distribution, the maximum likelihood function Where θ is the parameter vector of the distribution function, is the sample set of wind farm f, h is the joint distribution function under the parameter vector θ of the distribution function, a fi for A sample in

[0017] 2) Taking the logarithm of the maximum likelihood function l, we get H(θ) = lnl(θ);

[0018] 3) Find the parameter θ that maximizes the probability of the sample appearing;

[0019] 4) Determine the complete fitting distribution based on the parameter θ and calculate the fitting error;

[0020] 5) Perform steps 1) to 4) for the Beta distribution and the Gamma distribution respectively, compare the fitting errors of the truncated Gaussian distribution, the Beta distribution, and the Gamma distribution, select the distribution model with the smallest fitting error, and obtain the continuous distribution function.

[0021] Furthermore, the specific process of step 3 is:

[0022] Generate fault scenarios and analyze each fault scenario of the main wind turbine output scenarios within each wind farm using a linear optimal power flow model to obtain the corrected output distribution of each discrete wind farm;

[0023] The modified output distribution of each discrete wind farm is estimated by maximum likelihood, fitted into a continuous distribution, and a joint distribution is established based on the Copula joint distribution function to generate the total output distribution of the wind farm group.

[0024] The outage table of the transmission system is calculated by connecting the booster station and the cable elements in series. The total output distribution of the wind farm group is convolved with the outage table of the transmission system to obtain the grid-connected output distribution.

[0025] Furthermore, the constraints of the linear optimal power flow model include: power flow constraints, node voltage and line current constraints, and cable operation coupling relationships in the fault area;

[0026] The power flow constraints are as follows:

[0027]

[0028]

[0029]

[0030] Where U i，s is the voltage of node i in scenario s, M is a large number, is the disconnection status of branch ij in scenario s (0 means disconnection), I i，s is the injected current of node i in scenario s, I ij，s is the current of branch ij in scenario s, R ij is the resistance of branch ij, is the maximum power that can be generated by wind turbine i in scenario s, is a branch set, A collection of scenes;

[0031] The constraints on node voltage and line current not exceeding the limit are as follows:

[0032]

[0033]

[0034] Where, U is the lower voltage limit, is the upper voltage limit, is the upper limit of the current in branch ij;

[0035] The cable operation coupling relationship in the fault area is as follows:

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042] Where r ij，s Indicates the disconnection status of branch ij in the half area close to node i in scenario s (0 means disconnection), B ji Indicates whether the node j of the ij branch is configured with a circuit breaker (1 for configuration), s (1) and s (2) represents a certain operation scenario under k faults, is the set of operating scenarios under any k faults, ij→k means the line ij is disconnected corresponding to fault k, and subscript 0 indicates the confluence station;

[0043] Furthermore, the specific process of step 4 is: integrate according to the grid-connected output distribution, establish the duration output distribution with the horizontal axis being the duration and the vertical axis being the output size, and use the confidence capacity to be determined P crd Construct expressions for the two reliability parameters as unknown quantities:

[0044]

[0045]

[0046] Where X(t) is the integration result, LOLP is the load loss probability, and EENS is the expected load loss energy.

[0047] According to the load loss probability LOLP and the load loss energy expectation EENS, the corresponding confidence capacity is obtained, and the minimum confidence capacity is the comprehensive confidence capacity.

[0048] Furthermore, the specific process of evaluating the continuity and adequacy of the supply load of the offshore wind farm group based on the comprehensive confidence capacity is as follows:

[0049] If the comprehensive confidence capacity fails to reach the expected level of the plan, the offshore wind farm group will have defects in the continuity or adequacy of the load supply; if the comprehensive confidence capacity reaches the expected level of the plan, the offshore wind farm group will meet the requirements in terms of the continuity or adequacy of the load supply.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1) The present invention takes into account the random failures of wind turbines and the random failures of cables in the collection system during the grid connection of wind power systems, evaluates the reliability issues at the wind farm level, and uses output indicators to quantitatively reflect the impact of reliability issues on the overall output of the wind farm, making the final calculation of the wind farm's confidence capacity more realistic and more reflective of actual operating conditions. According to relevant data, since the probability of wind turbine failure and shutdown can reach 95%, the actual total output of a wind farm is often significantly lower than the expected output. The existing technology ignores the outage of wind turbines and cables, introducing significant errors in the evaluation of the overall system; the present invention takes this factor into account, and the evaluation results will be more accurate and can reflect the actual operation of the system.

[0052] 2) The present invention takes into account the scenario of offshore wind farm groups being connected to the grid together under the condition of large-scale development of offshore wind power in the future. In this scenario, multiple wind farms are geographically adjacent, there is a significant mutual influence, and the output conditions are highly correlated. At the same time, due to the larger scale of the problem analysis, the reliability calculation considering wind speed and correlation will be more difficult and have a large time complexity. The present invention analyzes the output correlation of each wind farm based on the existing Copula function; then, the reliability problem is used as a correction to the output distribution of the wind farm, and the distributed calculation of the total output problem of the wind farm group is realized, which greatly reduces the calculation time; finally, the marginal distribution of each wind farm is combined through the Copula function to establish a joint distribution, and the result of the total output distribution is obtained. The present invention realizes the decomposition of the overall problem, decomposes the entire calculation amount into the calculation of several sub-problems, and then uses the correlation to combine to obtain the main problem, thereby saving calculation time.

[0053] 3) Considering that future offshore wind farms will adopt more complex collection system topologies, the present invention proposes a reliability assessment method for complex topology collection systems; due to the complexity of the power system flow model, a linear flow model is proposed to simplify the calculation; due to the complex switch configuration of the collection system, a cable disconnection coupling model is proposed to transform the overall reliability assessment into an optimization problem for rapid solution and analysis.

[0054] 4) Existing wind power confidence capacity assessment methods typically consider only one reliability metric: LOLP or EENS. Considering only one metric cannot comprehensively evaluate wind power system output fluctuations. This invention proposes to simultaneously consider both reliability metrics to derive a comprehensive confidence capacity metric, ensuring that the system meets both the Loss of Load Rate (LOLP) and the Loss of Load Severity (EENS) requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0056] Figure 1 Flowchart of the present invention.

[0057] Figure 2 Schematic diagram of comprehensive confidence capacity evaluation of the present invention. DETAILED DESCRIPTION

[0058] The present invention will be further described below in conjunction with the accompanying drawings.

[0059] like Figure 1 As shown in the figure, the present invention provides a comprehensive confidence capacity assessment method for offshore wind farms with complex topologies. The final comprehensive confidence capacity result is obtained through four processes: wind speed scenario clustering, joint distribution generation, optimization-based reliability calculation, and multi-parameter comprehensive confidence capacity calculation. Specifically, the following steps are included:

[0060] Step 1: Wind speed scenario clustering can generate wind turbine output data (discrete) of wind farm groups using the original wind speed data according to the wake model. At the same time, the wind turbine output data of the wind farm group is reduced in dimension and features are extracted to obtain the main wind turbine output scenarios within each wind farm. The specific process is as follows:

[0061] After collecting exploration or meteorological wind speed data, we first use this data to generate a discrete wind turbine output distribution for each wind farm group based on wind turbine location, wind turbine output model, and wake model. Since the output distribution data is quite large, it is completely infeasible to use all the original data for subsequent analysis within the wind farm. Therefore, we need to perform dimensionality reduction and K-means clustering in step 1 to extract the main wind turbine output scenarios within each wind farm, reduce the data level, and prepare for subsequent analysis. The specific steps are as follows:

[0062] 1) Input wind turbine output data of a wind farm group and perform normalization preprocessing. In addition, let i = 1;

[0063] 2) Start processing the i-th wind farm data;

[0064] 3) Reduce the dimension of wind farm group wind turbine output data to the dimension of the i-th wind farm, that is, remove the data of wind turbines in other wind farms from each data point;

[0065] 4) Preset the value of k, where k is the total number of main scenes after clustering;

[0066] 5) Randomly initialize k cluster centers;

[0067] 6) Calculate the distance between the i-th wind farm data and each cluster center, and assign the i-th wind farm data point to the nearest cluster;

[0068] 7) Recalculate the cluster centers of each category, cluster centers n is the number of data points, is a set of r types of data, and a is a data point vector;

[0069] 8) Repeat steps 6) and 7) until the minimum error change is less than a certain standard or the number of iterations reaches a limit;

[0070] 9) Complete data processing for the i-th wind farm, i=i+1, repeat 2) to 9) until i equals the number of wind farms, and obtain the wind turbine output distribution of the wind farm group.

[0071] Step 2: Analyze the total output of the wind farms to obtain the wind farm joint output distribution (discrete) and the marginal output distribution (discrete) of each wind farm. By performing maximum likelihood estimation on the two distributions, two continuous fitting functions are obtained: the marginal distribution of each wind farm output and the wind farm joint output distribution. The Copula joint distribution function that describes the correlation between the outputs of each wind farm is also obtained. This can significantly reduce the workload of the subsequent reliability correction process. The specific process is as follows:

[0072] The output distribution of each wind farm is regarded as the marginal distribution, and the output distribution of the wind farm group is regarded as the joint distribution. The marginal distribution and the joint distribution are subjected to probability fitting of the continuous output distribution. The candidate distributions include truncated Gaussian distribution, Beta distribution and Gamma distribution. The distribution with the highest fitting accuracy is selected as its continuous distribution function. The process includes two steps. The first step is as follows:

[0073] 1) The distribution mode is set to truncated Gaussian distribution and the maximum likelihood function is defined Where θ is the parameter vector of the distribution function, is the sample set of wind farm f, h is the joint distribution function under the parameter vector θ of the distribution function, a fi for A sample in

[0074] 2) Taking the logarithm of the maximum likelihood function l, we get H(θ) = lnl(θ);

[0075] 3) To find the parameter θ that makes the sample group have the maximum probability, we can use the method of finding the extreme point to differentiate the function H(θ) and get Solve the system of equations to get the value of parameter θ

[0076] 4) Determine the complete fitting distribution based on the parameter θ and calculate the fitting error;

[0077] 5) Perform steps 1) to 4) for the other two candidate distributions respectively, compare the fitting accuracy of the three candidate distributions, select the distribution model with the highest accuracy, and record the marginal output distribution fitting function of each wind farm under parameter θ as F i (x i ).

[0078] The second step is to obtain the output copula joint distribution function of each wind farm. The specific process is: suppose the marginal output distribution fitting functions of each wind farm are F1(x1), F2(x2),…, F n (x n ), the fitting function of the wind farm joint output distribution is G(x), where x=(x1,x2,…,x n ), calculate the Copula joint distribution function to get The Copula joint distribution function obtained at this time can describe the correlation between the outputs of each wind farm. The wind farm joint output distribution G(x) can be expressed as G=C[F1(x1), F2(x2), ..., F n (x n )].

[0079] Step 3: Based on the main wind turbine output scenarios within each wind farm obtained in step 1, optimal control is performed through the linear optimal power flow model, with the goal of minimizing output loss. The analysis model takes into account wind turbine removal, cable branch removal, and multiple wind speed scenarios, thereby being able to analyze the post-fault topology reconstruction and wind turbine removal of complex topology collection systems. Based on the calculation results of the linear optimal power flow model, the corrected output distribution of each discrete wind farm is obtained. Then, based on the Copula joint distribution function obtained in step 2, the total output distribution of the wind farm group is generated. The total output distribution of the wind farm group and the outage table of the transmission system are connected in series to finally obtain the grid-connected output distribution of the wind farm group. The specific process is as follows:

[0080] The first step is to generate fault scenarios: use the fault enumeration method to enumerate all second-order events (or first-order and third-order events), mainly considering cable faults, and generate the system operation parameter set corresponding to the fault scenario. The variable r represents whether the cable is operating normally. If the cable ij is faulty under a certain fault scenario d, then let Whether the circuit breaker exists or not is indicated by B ij Indicates that B ij =1 means that a circuit breaker is installed near the node i in the cable ij;

[0081] The second step is to construct a linear optimal power flow model: First, based on the characteristic that the active power of the collection system is much greater than the reactive power, an approximation is adopted that ignores the reactive power, reactance, and voltage phase angle. The branch power flow is expressed as:

[0082] P i =U i I i

[0083] U i =R ij I ij +U j

[0084] Where, I i Inject current into node i, U i is the voltage at node i, P i is the injected power of node i. ij is the resistance of branch ij, I ij is the current of branch ij, U j is the voltage at node j.

[0085] The node injection current adopts the Taylor expansion approximation, which is expressed as:

[0086]

[0087] The power flow equation of the collector system is established based on the linear power flow model:

[0088]

[0089]

[0090] U=RI

[0091] Where, the first and third formulas are in the form of matrix variables, I is the injected current vector, P is the actual output power vector of the wind turbine, is the Hadamard product, 2 is the full 2 ​​vector, U is the node voltage vector, and R is the resistance matrix; the second formula is in the form of specific variables, P i,s is the actual output of fan i in scene s, is the maximum output power of wind turbine i in scenario s. Among them, the wind turbine output P is obtained by extracting the output of a single wind turbine from the main wind turbine output scenarios within each wind farm. The voltage vector U 0，s =1, subscript 0 represents the converter station node;

[0092] Considering the disconnection of the line, the power flow constraints are input into the optimization model according to the power flow equation of the collection system. The power flow constraints are as follows:

[0093]

[0094]

[0095]

[0096] Where U i，s is the voltage of node i in scenario s, M is a large number, is the disconnection status of branch ij in scenario s (0 means disconnection), I i，s is the injected current of node i in scenario s, I ij，s is the current of branch ij in scenario s, R ij is the resistance of branch ij, is the maximum power that can be generated by wind turbine i in scenario s, is a branch set, A collection of scenes;

[0097] The optimization model still has the constraints that the node voltage and line current cannot exceed the limit:

[0098]

[0099]

[0100] Where, U is the lower voltage limit, is the upper voltage limit, is the upper limit of the current in branch ij;

[0101] In the third step, for the power collection system, considering that after a cable fault, the circuit breaker operation will cause an area to be cut off, a large number of constraints are used to describe the cable operation coupling relationship in this fault area:

[0102]

[0103]

[0104]

[0105] Where r ij，sIndicates the disconnection status of branch ij in the half area close to node i in scenario s (0 means disconnection), B ji Indicates whether the node j of the ij branch is configured with a circuit breaker (1 for configured);

[0106] The fourth step is to set the circuit breaker to operate only once in a certain fault scenario to protect the working life of the circuit breaker. The following constraints are added: is a set of running scenarios under any k faults:

[0107]

[0108] Where s (1) and s (2) represents a certain operation scenario under k faults, is a set of operating scenarios under any k faults;

[0109] Step 5: Set the fault for each fault scenario:

[0110]

[0111] Where ij→k indicates that the disconnection of line ij corresponds to fault k;

[0112] Step 6: Set the objective function of the optimization model to maximize the collected power under all output scenarios:

[0113]

[0114] Where, subscript 0 indicates the confluence station;

[0115] By using the above optimization model, each fault scenario of each wind farm is analyzed to obtain the corrected output distribution of each discrete wind farm;

[0116] In the seventh step, the corrected output distribution of each discrete wind farm is multiplied by the wind turbine availability, and then maximum likelihood estimation is performed to fit the continuous distribution. The copula joint distribution function obtained in step 2 is used to establish a joint distribution to generate the total output distribution of the wind farm group.

[0117] In the eighth step, the booster station and cable components are connected in series to calculate the outage table of the transmission system. The total output distribution of the wind farm group is convolved with the outage table of the transmission system to obtain the grid-connected output distribution G c (x).

[0118] Step 4: Calculate the multi-parameter comprehensive confidence capacity: Use the grid-connected output distribution G obtained in step 3 c(x) Comprehensively evaluate the confidence capacity of the wind farm group by considering multiple reliability parameters, including the LOLP and EENS parameters of the equivalent conventional units. The two reliability parameters LOLP and EENS of the confidence capacity of the equivalent wind farm group must all meet the requirements;

[0119] The specific process is:

[0120] First, according to the grid output distribution G c (x) is integrated to establish a duration-output distribution with the horizontal axis being the duration and the vertical axis being the output size. This distribution is a monotonic function, such as Figure 2 As shown, the confidence capacity P can be used crd Construct expressions for the two reliability parameters as unknown quantities:

[0121]

[0122]

[0123] Where X(t) is the integration result, LOLP is the load loss probability, and EENS is the expected load loss energy.

[0124] According to the load loss probability LOLP and the load loss energy expectation EENS, the corresponding confidence capacity is obtained, and the minimum confidence capacity is taken as the comprehensive confidence capacity of the wind power system connected to the grid.

[0125] After obtaining the comprehensive confidence capacity, the value of the comprehensive confidence capacity is the maximum load that the wind power system can carry when the two reliability parameters are met. This comprehensive confidence capacity is an important reference for wind power system planning and can be used to evaluate the system's ability to supply load independently. If the comprehensive confidence capacity does not meet the planned expected level (which can be determined based on actual conditions), it means that the system has defects in the continuity or adequacy of the load supply, and other methods need to be adopted in the planning scheme to improve the system's load-carrying capacity (such as energy storage). If the comprehensive confidence capacity reaches the planned expected level (which can be determined based on actual conditions), then this indicator can be used for horizontal comparison or quantitative analysis of the system's power supply capacity.

[0126] A simple example is constructed below to illustrate the calculation of the present invention.

[0127] The wind farm group consists of 10 wind farms, each with 30 wind turbines. A radial topology and a fully switched configuration are employed. The wind farm group's wind turbine output distribution data consists of 40 data points, and the wind turbine output distribution within each wind farm also consists of 40 data points. Reliability analysis considers only first-order events. The method of the present invention, which performs output correction on each wind farm individually and establishes a joint distribution using a Copula function, and the conventional enumeration method, which uniformly enumerates the first-order faults of the wind farm group, are used for calculation. The method of the present invention takes 350 seconds, while the conventional enumeration method takes 25,000 seconds, showing a significant difference. The example calculation demonstrates a significant improvement in the efficiency of the calculation method of the present invention, thereby improving evaluation efficiency.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention 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 invention may be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A comprehensive confidence capacity assessment method for offshore wind farms with complex topology, characterized by: The following steps are involved: Step 1: Collect exploration or meteorological wind speed data, generate wind turbine output data for the wind farm group based on the wake model, perform dimensionality reduction and feature extraction on the wind turbine output data for the wind farm group, and obtain the main wind turbine output scenarios within each wind farm; Step 2: Analyze the total output of the wind farm to obtain the wind farm joint output distribution and the marginal output distribution of each wind farm. Perform maximum likelihood estimation on the wind farm joint output distribution and the marginal output distribution of each wind farm to obtain the Copula joint distribution function. Step 3: Based on the main wind turbine output scenarios within each wind farm obtained in Step 1, a linear optimal power flow model is used to calculate and obtain the modified output distribution of each discrete wind farm. Continuous fitting is then performed. Based on the Copula joint distribution function, a joint distribution is established for the fitted modified output distributions of each wind farm to generate the total output distribution of the wind farm group. Based on the total output distribution of the wind farm group and the outage table of the transmission system, the grid-connected output distribution is obtained. The specific process of Step 3 is as follows: Generate fault scenarios and analyze each fault scenario of the main wind turbine output scenarios within each wind farm using a linear optimal power flow model to obtain the corrected output distribution of each discrete wind farm; The modified output distribution of each discrete wind farm is estimated by maximum likelihood, fitted into a continuous distribution, and a joint distribution is established based on the Copula joint distribution function to generate the total output distribution of the wind farm group. The outage table of the transmission system is calculated by connecting the booster station and the cable element in series, and the total output distribution of the wind farm group is convolved with the outage table of the transmission system to obtain the grid-connected output distribution; Step 4: Obtain a comprehensive confidence capacity based on the grid-connected output distribution obtained in step 3; and evaluate the sustainability and adequacy of the supply load of the offshore wind farm group based on the comprehensive confidence capacity.

2. The comprehensive confidence capacity assessment method for offshore wind farms with complex topology according to claim 1 is characterized in that: The specific process of step 1 is: Exploration or meteorological wind speed data is collected, and the wind turbine output distribution of the wind farm group is generated based on the wind turbine location, wind turbine output model and wake model. Then, the wind turbine output data of the wind farm group is reduced in dimension and clustered using K-means to extract the main wind turbine output scenarios within each wind farm.

3. The comprehensive confidence capacity assessment method for offshore wind farms with complex topology according to claim 1 is characterized in that: The specific process of step 2 is: The maximum likelihood estimation of the continuous output distribution is performed on the output distribution of each wind farm and the output distribution of the wind farm group. Then, the distribution with the smallest fitting error among the truncated Gaussian distribution, Beta distribution, and Gamma distribution is selected as the continuous distribution function. Based on the continuous distribution function, the marginal output distribution fitting function of each wind farm is obtained. According to the marginal output distribution fitting function of each wind farm, the Copula joint distribution function of the output correlation of each wind farm is obtained.

4. The comprehensive confidence capacity assessment method for offshore wind farms with complex topology according to claim 3 is characterized in that: The continuous distribution function is obtained through the following process: 1) When the distribution is truncated Gaussian distribution, the maximum likelihood function ,in θ is the parameter vector of the distribution function, For wind farms f The sample set, h is the parameter vector of the distribution function θ The joint distribution function under a fi for A sample in 2) For the maximum likelihood function l Taking the logarithm ; 3) Find the parameters that maximize the probability of a sample appearing θ value; 4) According to the parameters θ Determine the complete fitting distribution and calculate the fitting error; 5) Perform steps 1) to 4) for the Beta distribution and the Gamma distribution respectively. Compare the fitting errors of the truncated Gaussian distribution, the Beta distribution, and the Gamma distribution. Select the distribution model with the smallest fitting error to obtain the continuous distribution function.

5. The comprehensive confidence capacity assessment method for offshore wind farms with complex topology according to claim 1 is characterized in that: The constraints of the linear optimal power flow model include: power flow constraints, node voltage and line current constraints, and cable operation coupling in the fault area. The power flow constraints are as follows: Where, For nodes i exist s The voltage in the scenario, M is a large number, For branch ij In the scene s Under the disconnection situation, For nodes i exist s The injected current in the scenario, For branch ij exist s The current in the scene, For branch ij The resistance, For fans i exist s The maximum power that can be generated in the scenario, is a branch set, A collection of scenes; The constraints on node voltage and line current not exceeding the limit are as follows: Where, is the lower voltage limit, is the upper voltage limit, It is a branch road ij The upper limit of current; The cable operation coupling relationship in the fault area is as follows: Where, Indicates a branch ij exist s Close to the scene i The disconnection status of the node half area, express ij branch road j Whether the node is configured with a circuit breaker, and Indicates a fault k In a certain running scenario, It is arbitrary k A collection of operating scenarios under fault conditions, Indicates line ij Disconnect corresponding fault k , subscript 0 indicates the confluence station.

6. The comprehensive confidence capacity assessment method for offshore wind farms with complex topology according to claim 1 is characterized in that: The specific process of step 4 is: integrate according to the grid-connected output distribution, establish the duration output distribution with the horizontal axis being the duration and the vertical axis being the output size, and use the confidence capacity to be determined P crd Construct expressions for the two reliability parameters as unknown quantities: Where, X ( t ) is the integration result, LOLP is the load loss probability, and EENS is the expected load loss energy; According to the load loss probability LOLP and the load loss energy expectation EENS, the corresponding confidence capacity is obtained, and the minimum confidence capacity is the comprehensive confidence capacity.

7. The comprehensive confidence capacity assessment method for offshore wind farms with complex topology according to claim 1 is characterized in that: The specific process of evaluating the continuity and adequacy of the supply load of an offshore wind farm group based on the comprehensive confidence capacity is as follows: If the comprehensive confidence capacity fails to reach the expected level of the plan, the offshore wind farm group will have defects in the continuity or adequacy of the load supply; if the comprehensive confidence capacity reaches the expected level of the plan, the offshore wind farm group will meet the requirements in terms of the continuity or adequacy of the load supply.

Images