A coordinated control method, system and storage medium for an offshore wind farm

Abstract

The present invention relates to the technical field of wind farm control, and specifically to a coordinated control method, system and storage medium for an offshore wind farm. First, according to the wind speed prediction values of each wind turbine, the theoretical maximum active power is obtained, and combined with the power grid dispatching data and the actual output power of the wind farm, the upper limit of the active power demand and the reactive power range of the wind farm are obtained; then, the power reference values of each wind turbine are obtained by using a multi-objective optimization power prediction model and an improved particle swarm optimization algorithm; combined with the power reference values and the fan status information, the reinforcement learning algorithm is used to optimize the grouping of groups and the power distribution within the groups to obtain the actual power reference values; then, the actual power reference correction values are obtained by using the power compensation correction formula; finally, according to the actual power reference correction values, the maximum power point tracking control is used to control the active power and reactive power of each wind turbine. The present invention can improve the effectiveness of the coordinated control of the wind farm.

Description

Technical Field

[0001] The present invention relates to the technical field of wind farm control, and particularly to a coordinated control method, system and storage medium for an offshore wind farm. Background Technique

[0002] The coordinated control method for an offshore wind farm is to send the power reference value calculated by the wind farm controller to each wind turbine and compare it with the actual value of the wind turbine, so as to use the wind turbine controller to adjust the power difference to achieve the power reaching the power reference value; at the same time, the power reference value can also be converted into other parameter change amounts, such as voltage and current, through gain to adjust the corresponding wind farm parameters of the wind turbine.

[0003] Currently, there are still deficiencies in the coordinated control of offshore wind farms in the prior art; on the one hand, the reference value obtained by the wind farm control equipment cannot adapt to the changes in the operating states of the power grid and the wind farm, which may lead to power grid voltage fluctuations or unqualified wind farm power factors; on the other hand, the prior art does not adaptively adjust the grouping and power distribution of wind turbines according to the real-time operating state of the wind farm and the requirements of the power grid, which will reduce the effectiveness of the coordinated control of the offshore wind farm.

[0004] Therefore, a coordinated control method, system and storage medium for an offshore wind farm are proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a coordinated control method, system and storage medium for an offshore wind farm. First, according to the wind speed prediction value of each wind turbine, the theoretical maximum active power is obtained, and combined with the power grid dispatching data and the actual output power of the wind farm, the upper limit of the active power demand and the reactive power range of the wind farm are obtained; then, the power reference value of each wind turbine is obtained by using a multi-objective optimization power prediction model and an improved particle swarm algorithm; combined with the power reference value and the wind turbine state information, the reinforcement learning algorithm is used to optimize the grouping of groups and the power distribution within the group to obtain the actual power reference value; then, the actual power reference correction value is obtained by using the power compensation correction formula; finally, according to the actual power reference correction value, the maximum power point tracking control is used to control the active power and reactive power of each wind turbine.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A coordinated control method for an offshore wind farm, comprising:

[0008] Using a short-term wind speed prediction model based on Kalman filter to obtain the wind speed prediction value of each wind turbine;

[0009] Based on the predicted wind speed value, obtain the theoretical maximum active power of each wind turbine; combine the theoretical maximum active power, grid dispatching data, and the actual output power of the wind farm to obtain the upper limit of the active power demand and the reactive power range of the wind farm;

[0010] Construct a multi-objective optimization power prediction model, combine the upper limit of the active power demand and the reactive power range, and use the improved particle swarm optimization algorithm to solve the multi-objective optimization power prediction model to obtain the power reference value of each wind turbine;

[0011] Combine the power reference value and the fan status information, and use the reinforcement learning algorithm to optimize the grouping of groups and the power distribution within the group to obtain the actual power reference value of each wind turbine;

[0012] Input the actual power reference value of each wind turbine into the power compensation correction formula to obtain the actual power reference correction value of each wind turbine;

[0013] According to the actual power reference correction value, use the maximum power point tracking control to control the active power and reactive power of each wind turbine.

[0014] Furthermore, the process of obtaining the upper limit of the active power demand and the reactive power range of the wind farm includes:

[0015] Obtain the allowable maximum power and power factor range from the grid dispatching data;

[0016] Combine the predicted wind speed value and the power curve of the wind turbine to obtain the theoretical maximum active power of each wind turbine;

[0017] Sum up the theoretical maximum active power of each wind turbine to obtain the theoretical maximum active power of the wind farm;

[0018] Compare the allowable maximum power with the theoretical maximum active power of the wind farm to obtain the upper limit of the active power demand;

[0019] Combine the power factor range and the actual output power of the wind farm to obtain the reactive power range.

[0020] Furthermore, the process of obtaining the power reference value of each wind turbine includes:

[0021] Take the total power generation revenue of the wind farm, the life loss of the wind turbine, and the voltage deviation at the grid connection point of the wind farm as the optimization objectives of the multi-objective optimization power prediction model and set the total objective function;

[0022] Set the upper limit of the active power demand and the reactive power range as the total power constraint conditions of the multi-objective optimization power prediction model of the wind farm;

[0023] Solve the multi-objective optimization power prediction model by using an improved particle swarm algorithm to obtain the power reference value of each wind turbine.

[0024] Furthermore, the process of obtaining the actual power reference value of each wind turbine includes:[[]]

[0025] Define the state space, action space, and reward function of the Markov decision process.

[0026] Among them, the state space includes: the wind turbine state information, group information, and the power reference value; the action space includes: the clustering matrix, active power distribution coefficient, and reactive power distribution coefficient; the reward function includes: power deviation reward, average mechanical load reward, and group number change reward.

[0027] Train the Actor network and Critic network by using deep deterministic policy gradient and historical data, and input the current wind turbine state information and the current power reference value into the trained network to obtain the actual power reference value.

[0028] Furthermore, the expression of the power compensation correction formula is:[[]]

[0029] ;

[0030] Among them,[[]] represents the actual power reference correction value of the th wind turbine at the th moment; represents the minimum value operation; represents the actual power reference value of the th wind turbine at the th moment; represents the wind speed compensation coefficient; represents the th wind turbine at the th moment of the wind speed prediction value; among them,[[]] is the communication delay time; represents the th wind turbine at the th moment of the current wind speed value; represents the delay compensation coefficient; represents the estimated value of the power change caused by communication delay; represents the th wind turbine at the th moment of the available power.

[0031] A coordinated control system for an offshore wind farm, comprising: a system control module, a data acquisition module, a multi-objective optimization power prediction module, a group power distribution module, a power compensation correction module, and a power control module;

[0032] The system control module is used to control the startup, pause, and stop of the system;

[0033] The data acquisition module is used to obtain the wind speed prediction value and the fan status information of each wind turbine, as well as the upper limit of the active power demand and the reactive power range of the wind farm;

[0034] The multi-objective optimization power prediction module is used to combine the upper limit of the active power demand and the reactive power range, and use an improved particle swarm optimization algorithm to solve the multi-objective optimization power prediction model to obtain the power reference value of each wind turbine;

[0035] The group power distribution module is used to combine the power reference value and the fan status information, and use a reinforcement learning algorithm to optimize the group clustering and the power distribution within the group to obtain the actual power reference value of each wind turbine;

[0036] The power compensation correction module is used to input the actual power reference value of each wind turbine into the power compensation correction formula to obtain the actual power reference correction value of each wind turbine;

[0037] The power control module is used to control the active power and reactive power of each wind turbine by using the maximum power point tracking control according to the actual power reference correction value.

[0038] A storage medium, on which computer program instructions are stored, and when the computer program is executed by a processor, the steps in a coordinated control method for an offshore wind farm are implemented.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] 1. The present invention proposes a multi-objective optimization power prediction method for obtaining the globally optimal power reference value of the wind farm; this method realizes power optimization by constructing the objective function and constraint conditions of the multi-objective optimization power prediction model; this model considers the power generation benefit, life loss, and grid connection point voltage of the wind farm, and combines an improved particle swarm optimization algorithm for solution to obtain the power reference value of each wind turbine under global optimization; this process optimizes power from the overall perspective of the wind farm and the power grid, provides global reference data for subsequent local power optimization, and thus can effectively improve the effectiveness of the coordinated control of the offshore wind farm.

[0041] 2. The present invention proposes a group power distribution method to obtain the actual power reference value that is locally optimal for a wind turbine group. This method defines the state space, action space, and reward function of the Markov decision process and combines the deep deterministic policy gradient algorithm to obtain the actual power reference value. This process not only optimizes the grouping of wind turbines but also optimizes the power distribution according to the status information of the turbines, and can improve the effectiveness of the coordinated control of an offshore wind farm by considering the local differences and actual constraints of the turbines.

[0042] 3. The present invention proposes a power compensation and correction method to compensate and correct the power. This method inputs the actual power reference value into the power compensation and correction formula to obtain the actual power reference correction value. The power compensation and correction formula considers the actual impacts of wind speed deviation, communication delay, and available capacity on the power, so that the actual power reference correction value can be more in line with the actual situation, thereby improving the effectiveness of the coordinated control of an offshore wind farm. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a schematic flow chart of a coordinated control method for an offshore wind farm according to the present invention;

[0044] Figure 2 is a schematic structural diagram of a coordinated control system for an offshore wind farm according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0046] Please refer to Figure 1 and Figure 2 , the present invention provides a coordinated control method, system, and storage medium for an offshore wind farm, and the technical solutions are as follows:

[0047] Embodiment 1

[0048] A certain power plant uses a coordinated control method for an offshore wind farm proposed by the present invention to effectively ensure the power stability of each wind turbine in the offshore wind farm. The flow of this method is schematically shown in Figure 1 as shown below:

[0049] Use a short-term wind speed prediction model based on Kalman filtering to obtain the wind speed prediction value of each wind turbine;

[0050] Furthermore, the process of obtaining the wind speed prediction value of each wind turbine includes:

[0051] Call the historical wind speed data of each wind turbine from the background monitoring system;

[0052] Clean and preprocess the historical wind speed data to obtain training data;

[0053] Construct a short-term wind speed prediction model based on Kalman filter, and define the state variables, state transition equation, observation equation and noise covariance matrix of the Kalman filter; among them, the state variables include wind speed and wind speed change rate;

[0054] Use the training data to iteratively train the Kalman filter, and use the trained Kalman filter to perform short-term wind speed prediction to obtain the wind speed prediction value of each wind turbine.

[0055] According to the wind speed prediction value, obtain the theoretical maximum active power of each wind turbine; combine the theoretical maximum active power, grid dispatching data and the actual output power of the wind farm to obtain the upper limit of the active power demand and the reactive power range of the wind farm;

[0056] Furthermore, the process of obtaining the upper limit of the active power demand and the reactive power range of the wind farm includes:

[0057] Obtain the allowable maximum power and power factor range from the grid dispatching data;

[0058] Combine the wind speed prediction value and the power curve of the wind turbine to obtain the theoretical maximum active power of each wind turbine;

[0059] Sum up the theoretical maximum active power of each wind turbine to obtain the theoretical maximum active power of the wind farm;

[0060] Compare the allowable maximum power with the theoretical maximum active power of the wind farm to obtain the upper limit of the active power demand;

[0061] Combine the power factor range and the actual output power of the wind farm to obtain the reactive power range;

[0062] Furthermore, the grid dispatching data and the actual output power of the wind farm are obtained by calling from the background monitoring system;

[0063] Furthermore, the power curve of the wind turbine is obtained by using a wind speed-power lookup table.

[0064] The following gives the upper limit of the active power demand and the reactive power range of the wind farm obtained at three different times, which are respectively denoted as Sample One, Sample Two and Sample Three, as shown in Table 1.

[0065] Table 1 Upper limit of active power demand and reactive power range

[0066]

[0067] By obtaining the upper limit of active power demand and the reactive power range of a wind farm based on the wind speed prediction value, grid dispatching data, and actual output power, accurate constraint condition data can be provided for subsequent multi-objective optimized power prediction, thereby accurately obtaining the theoretical power reference value of each wind turbine and further improving the effectiveness of coordinated control of an offshore wind farm.

[0068] Construct a multi-objective optimized power prediction model, combine the upper limit of active power demand and the reactive power range, and use an improved particle swarm optimization algorithm to solve the multi-objective optimized power prediction model to obtain the power reference value of each wind turbine;

[0069] Furthermore, the process of obtaining the power reference value of each wind turbine includes:

[0070] Take the total power generation revenue of the wind farm, the life loss of the wind turbines, and the voltage deviation at the grid connection point of the wind farm as the optimization objectives of the multi-objective optimized power prediction model and set the total objective function;

[0071] Set the upper limit of active power demand and the reactive power range as the total power constraint conditions of the multi-objective optimized power prediction model for the wind farm;

[0072] Use an improved particle swarm optimization algorithm to solve the multi-objective optimized power prediction model to obtain the power reference value of each wind turbine;

[0073] Furthermore, the expression of the total objective function is:

[0074] ;

[0075] Among them, represents the total objective function; represents the weight of the total power generation revenue; represents the total power generation revenue objective function of the wind farm; represents the weight of life loss; represents the life loss objective function of the wind turbines; represents the weight of the voltage deviation at the grid connection point; represents the voltage deviation objective function at the grid connection point of the wind farm.

[0076] Furthermore, the weights of the total power generation revenue, life loss, and voltage deviation at the grid connection point are set to 0.35, 0.3, and 0.35 respectively; of course, these weights can be flexibly set according to actual needs.

[0077] Furthermore, the total power generation revenue objective function of the wind farm can be expressed as:

[0078] ;

[0079] Among them, represents the total power generation revenue objective function of the wind farm; represents the maximization function; represents the number of wind turbines; represents the electricity price at time represents the th wind turbine's actual active power at time ; represents the time interval;

[0080] Furthermore, the wind turbine life loss objective function can be expressed as:

[0081] ;

[0082] wherein, represents the wind turbine life loss objective function; represents the minimization function; represents the number of wind turbines; represents the th wind turbine's damage value at time , and this damage value can be obtained through the wind turbine load analysis model; represents the th wind turbine's actual active power at time ; represents the time interval;

[0083] Furthermore, the wind farm connection point voltage deviation objective function can be expressed as:

[0084] ;

[0085] wherein, represents the wind farm connection point voltage deviation objective function; represents the minimization function; represents the number of wind turbines; represents the wind farm connection point voltage at time , and this wind farm connection point voltage can be obtained through power system power flow calculation; represents the reference value of the connection point voltage at time ; represents the time interval;

[0086] Furthermore, the improved particle swarm optimization algorithm adaptively adjusts the parameters of the particle swarm optimization algorithm, such as the inertia weight, acceleration coefficients, etc., to improve the convergence speed and accuracy of the algorithm; the adaptive adjustment of the inertia weight can adopt a linear decreasing strategy, a non-linear decreasing strategy, and an adjustment strategy based on the population diversity index; the adaptive adjustment of the acceleration coefficients adopts a dynamic adjustment strategy to dynamically balance the influence of the acceleration coefficients at different stages.

[0087] By constructing a multi-objective optimization power prediction model from the overall perspective of the wind farm and the power grid and using the improved particle swarm optimization algorithm to obtain the global optimal solution, global reference data is provided for subsequent local power optimization, thereby effectively improving the effectiveness of the coordinated control of the offshore wind farm.

[0088] Combining the power reference value and the fan status information, the reinforcement learning algorithm is used to optimize the grouping of groups and the power distribution within the groups to obtain the actual power reference value of each wind turbine;

[0089] Furthermore, the process of obtaining the actual power reference value of each wind turbine includes:

[0090] Defining the state space, action space, and reward function of the Markov decision process;

[0091] Among them, the state space includes: fan status information, group information, and power reference value; the action space includes: clustering matrix, active power distribution coefficient, and reactive power distribution coefficient; the reward function includes: power deviation reward, average mechanical load reward, and group number change reward;

[0092] Furthermore, the fan status information includes: the geographical location of the i-th wind turbine and the actual wind speed, available useful power, available reactive power, and mechanical load at time t;

[0093] Furthermore, the group information includes: the current number of groups and the group member indication matrix; the group member indication matrix is used to represent the belonging relationship of the wind turbines to the groups, for example: Indicating that the th wind turbine belongs to the th group;

[0094] Furthermore, the power reference value is the global optimal power allocated to each wind turbine by using the multi-objective optimization power prediction model, including the active power reference value and the reactive power reference value;

[0095] Furthermore, the clustering matrix is used to represent the clustering situation of the wind turbines, including the probability of the wind turbines being assigned to the groups; the active power distribution coefficient and the reactive power distribution coefficient are used to represent the power distribution actions, and the sum of the active power distribution coefficients and the reactive power distribution coefficients of each group is 1;

[0096] Furthermore, the reward function can be expressed as:

[0097] ;

[0098] where, represents the reward function; represents the power deviation reward weight; represents the power deviation reward function, which is used to calculate the deviation between the actual power and the reference power; represents the average mechanical load reward weight; represents the mechanical load reward function, which is used to calculate the deviation between the mechanical load of a single unit and the average mechanical load of the total units; represents the group number change reward weight; represents the group number change reward function, which is used to reduce the change of the group number;

[0099] Furthermore, the power deviation reward function can be expressed as:

[0100] ;

[0101] where, represents the power deviation reward function; represents the number of wind turbines; represents the th wind turbine's actual active power at time; represents the th wind turbine's reference active power at time; represents the th wind turbine's actual reactive power at time; represents the th wind turbine's reference reactive power at time;

[0102] Furthermore, the mechanical load reward function can be expressed as:

[0103] ;

[0104] where, represents the mechanical load reward function; represents the number of wind turbines; represents the th wind turbine's mechanical load at time; represents the average mechanical load of all wind turbines at time;

[0105] Furthermore, the group number change reward function can be expressed as:

[0106] ;

[0107] where represents the group number change reward function; represents the penalty coefficient for controlling the frequency of group number change, which can be flexibly set according to actual needs; represents the group the number of changes within;

[0108] Furthermore, the power deviation reward weight, the average mechanical load reward weight, and the group number change reward weight are respectively set to 0.35, 0.35, and 0.3; the setting of these weights can be adjusted according to the actual situation.

[0109] The Actor network and the Critic network are trained using the deep deterministic policy gradient algorithm and historical data, and the current wind turbine state information and the current power reference value are input into the trained network to obtain the actual power reference value.

[0110] The actual power reference value is obtained by defining the state space, action space, and reward function of the Markov decision process and combining the deep deterministic policy gradient algorithm. This process not only optimizes the grouping of wind turbines but also optimizes the power distribution according to the state information of the units. By considering the local differences and actual constraints of the units, the effectiveness of the coordinated control of the offshore wind farm can be improved.

[0111] The actual power reference value of each wind turbine is input into the power compensation correction formula to obtain the actual power reference correction value of each wind turbine;

[0112] Furthermore, the expression of the power compensation correction formula is:

[0113] ;

[0114] where represents the th wind turbine at the actual power reference correction value at the moment; represents the minimum value operation; represents the th wind turbine at the actual power reference value at the moment; represents the wind speed compensation coefficient; represents the th wind turbine at the wind speed prediction value at the moment; where is the communication delay time; represents the current wind speed value of the th wind turbine at moment; represents the delay compensation coefficient; represents the estimated value of power change caused by communication delay; represents the th wind turbine's available power at moment.

[0115] Furthermore, considering the differences in the dynamic characteristics of wind turbines and control systems, the wind speed compensation coefficient and the delay compensation coefficient need to be set according to the actual situation and are not unique.

[0116] To illustrate the effectiveness of the power compensation correction formula proposed by the present invention, three actual power reference values and available active powers at different moments from different wind turbines are randomly selected for power compensation correction tests, which are respectively recorded as Test One, Test Two, and Test Three; the corresponding wind speed compensation coefficients and delay compensation coefficients are set according to different turbines; combined with the calculation process of the power compensation correction formula, the actual power reference correction values of each group are obtained; among them, the actual power reference value includes the actual active power reference value (left) and the actual reactive power reference value (right); the actual power reference correction value includes the actual active power reference correction value (left) and the actual reactive power reference correction value (right); the power compensation correction test results are shown in Table 2.

[0117] Table 2 Power Compensation Correction Test Results

[0118]

[0119] By considering the actual impacts of wind speed deviation, communication delay, and available capacity on power, the actual power reference correction value can be made more in line with the actual situation, thereby improving the effectiveness of the coordinated control of offshore wind farms.

[0120] According to the actual power reference correction value, the maximum power point tracking control is adopted to control the active power and reactive power of each wind turbine.

[0121] This embodiment proposes a coordinated control method for an offshore wind farm. The method first combines the upper limit of the active power demand and the reactive power range of the wind farm, and uses a multi-objective optimization power prediction model and an improved particle swarm algorithm to obtain the power reference value of each wind turbine. Then, according to the power reference value and the wind turbine state information, the reinforcement learning algorithm is used to optimize the grouping of the groups and the power distribution within the groups to obtain the actual power reference value. Next, the actual power reference correction value is obtained by using the power compensation correction formula. Finally, according to the actual power reference correction value, the maximum power point tracking control is used to control the active power and reactive power of each wind turbine. This method adopts a global-local-compensation correction strategy, thereby improving the effectiveness of the coordinated control of the offshore wind farm.

[0122] Embodiment 2

[0123] The present invention proposes a coordinated control system for an offshore wind farm. The structure of the system can refer to Figure 2 , including: a system control module, a data acquisition module, a multi-objective optimization power prediction module, a group power distribution module, a power compensation correction module, and a power control module;

[0124] The system control module is used to control the start, pause, and stop of the system;

[0125] The data acquisition module is used to obtain the wind speed prediction value and the wind turbine state information of each wind turbine, as well as the upper limit of the active power demand and the reactive power range of the wind farm;

[0126] The multi-objective optimization power prediction module is used to combine the upper limit of the active power demand and the reactive power range, and use the improved particle swarm algorithm to solve the multi-objective optimization power prediction model to obtain the power reference value of each wind turbine;

[0127] Further, the process of the multi-objective optimization power prediction module obtaining the power reference value of each wind turbine includes:

[0128] Taking the total power generation revenue of the wind farm, the life loss of the wind turbine, and the voltage deviation at the grid connection point of the wind farm as the optimization objectives of the multi-objective optimization power prediction model and setting the total objective function;

[0129] Setting the upper limit of the active power demand and the reactive power range as the total power constraint conditions of the wind farm for the multi-objective optimization power prediction model;

[0130] Using the improved particle swarm algorithm to solve the multi-objective optimization power prediction model to obtain the power reference value of each wind turbine.

[0131] The group power distribution module is used to combine the power reference value and the wind turbine state information, and use the reinforcement learning algorithm to optimize the grouping of the groups and the power distribution within the groups to obtain the actual power reference value of each wind turbine;

[0132] Furthermore, the process by which the group power distribution module obtains the actual power reference value of each wind turbine includes:

[0133] Define the state space, action space, and reward function of the Markov decision process;

[0134] Among them, the state space includes: wind turbine state information, group information, and power reference value; the action space includes: clustering matrix, active power distribution coefficient, and reactive power distribution coefficient; the reward function includes: power deviation reward, average mechanical load reward, and group number change reward;

[0135] Use deep deterministic policy gradient and historical data to train the Actor network and the Critic network, and input the current wind turbine state information and the current power reference value into the trained network to obtain the actual power reference value.

[0136] The power compensation and correction module is used to input the actual power reference value of each wind turbine into the power compensation and correction formula to obtain the actual power reference correction value of each wind turbine;

[0137] Furthermore, the expression of the power compensation and correction formula is:

[0138] ;

[0139] Among them, represents the actual power reference correction value of the th wind turbine at time; represents the minimum value operation; represents the th wind turbine at time; represents the wind speed compensation coefficient; represents the th wind turbine at time; among them, is the communication delay time; represents the th wind turbine at time; represents the delay compensation coefficient; represents the estimated power change due to communication delay; represents the th wind turbine at time;

[0140] The power control module is used to control the active power and reactive power of each wind turbine by means of maximum power point tracking control according to the actual power reference correction value.

[0141] The system of the present invention adopts a global-local-compensation correction scheme. To further verify the effectiveness of the proposed scheme of the present invention, the same test data is randomly selected from the data acquisition module for testing and verification of different schemes, which are respectively recorded as Test 1, Test 2, and Test 3; the test schemes specifically include: the scheme proposed by the present invention, that is, the multi-objective optimization power prediction module, the group power distribution module, and the power compensation correction module, which is recorded as Scheme 1; removing the group power distribution module in the system and retaining the multi-objective global optimization and compensation correction strategy, which is recorded as Scheme 2; only retaining the multi-objective optimization power prediction module in the system and removing the group local optimization and compensation correction strategy, which is recorded as Scheme 3;

[0142] Input the test data into different schemes, obtain the final power reference values of the wind turbines under different schemes, and obtain the proportion of the final power reference values within a reasonable range through manual verification. The comparison test results of the schemes can be referred to Table 3.

[0143] Table 3 Comparison Test Results of Schemes

[0144]

[0145] As shown in Table 3, compared with other schemes, the scheme proposed by the present invention, that is, Scheme 1, can allocate more reasonable power reference values for each wind turbine, thereby effectively improving the power control of the offshore wind farm.

[0146] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirits of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A coordinated control method for an offshore wind farm, characterized in that: include: The short-term wind speed prediction model based on Kalman filter is used to obtain the wind speed prediction value of each wind turbine; According to the wind speed prediction value, the theoretical maximum active power of each wind turbine is obtained; and the upper limit of the active power demand and the reactive power range of the wind farm are obtained by combining the theoretical maximum active power, the grid dispatching data and the actual output power of the wind farm; Constructing a multi-objective optimization power prediction model, combining the active power demand upper limit and the reactive power range, and using an improved particle swarm algorithm to solve the multi-objective optimization power prediction model to obtain a power reference value for each wind turbine; Combining the power reference value and wind turbine status information, using a reinforcement learning algorithm to optimize grouping and power allocation within the group, to obtain an actual power reference value for each wind turbine set; Inputting the actual power reference value of each wind turbine generator set into a power compensation correction formula to obtain an actual power reference correction value of each wind turbine generator set; According to the actual power reference correction value, the active power and reactive power of each wind turbine generator set are controlled by adopting the maximum power point tracking control.

2. A coordinated control method for an offshore wind farm according to claim 1, characterized in that: The process of obtaining the upper limit of the active power demand and the reactive power range of the wind farm includes: Obtaining the maximum allowable power and power factor range from the power grid dispatching data; Combining the wind speed prediction value with the wind turbine power curve, obtaining the theoretical maximum active power of each wind turbine; The theoretical maximum active power of each wind turbine is summed to obtain the theoretical maximum active power of the wind farm; Comparing the maximum allowable power with the theoretical maximum active power of the wind farm to obtain the upper limit of the active power demand; The reactive power range is obtained by combining the power factor range with the actual output power of the wind farm.

3. A coordinated control method for an offshore wind farm according to claim 1, characterized in that: The process of obtaining the power reference value of each wind turbine generator set includes: The total power generation income of the wind farm, the life loss of the wind turbines and the voltage deviation of the wind farm grid connection point are taken as the optimization objectives of the multi-objective optimization power prediction model and the total objective function is set; Setting the active power demand upper limit and the reactive power range as wind farm total power constraint conditions of the multi-objective optimization power prediction model; The multi-objective optimization power prediction model is solved by using an improved particle swarm algorithm to obtain the power reference value of each wind turbine generator set.

4. A coordinated control method for an offshore wind farm according to claim 1, characterized in that: The process of obtaining the actual power reference value of each wind turbine generator set includes: Define the state space, action space, and reward function of a Markov decision process; The state space includes: the wind turbine state information, group information and the power reference value; the action space includes: the grouping matrix, the active power allocation coefficient and the reactive power allocation coefficient; the reward function includes: power deviation reward, average mechanical load reward and group quantity change reward; The Actor network and the Critic network are trained using deep deterministic policy gradients and historical data, and the current wind turbine state information and the current power reference value are input into the trained network to obtain the actual power reference value.

5. The coordinated control method of an offshore wind farm according to claim 1, characterized in that: The power compensation correction formula is expressed as: ； in, Indicates Wind turbines in The actual power reference correction value at the moment; Indicates the minimum value operation; Indicates Wind turbines in The actual power reference value at the moment; Indicates the wind speed compensation coefficient; Indicates Wind turbines in The wind speed forecast value at time; wherein, is the communication delay time; Indicates Wind turbines in The current wind speed value at the moment; represents the delay compensation coefficient; represents the estimated value of power variation due to communication delay; Indicates Wind turbines in The available power at the moment.

6. A coordinated control system for an offshore wind farm, characterized in that: include: System control module, data acquisition module, multi-objective optimization power prediction module, group power allocation module, power compensation correction module and power control module; The system control module is used to control the start, pause and stop of the system; The data acquisition module is used to obtain the wind speed forecast value and wind turbine status information of each wind turbine group and the upper limit of active power demand and reactive power range of the wind farm; The multi-objective optimization power prediction module is used to combine the active power demand upper limit and the reactive power range, and use an improved particle swarm algorithm to solve the multi-objective optimization power prediction model to obtain a power reference value for each wind turbine; The group power allocation module is used to combine the power reference value and the wind turbine status information, and optimize the grouping and power allocation within the group by using a reinforcement learning algorithm to obtain an actual power reference value for each wind turbine group; The power compensation correction module is used to input the actual power reference value of each wind turbine group into the power compensation correction formula to obtain the actual power reference correction value of each wind turbine group; The power control module is used to control the active power and reactive power of each wind turbine generator set by adopting maximum power point tracking control according to the actual power reference correction value.

7. A storage medium having computer program instructions stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the coordinated control method of an offshore wind farm as claimed in any one of claims 1 to 5 are implemented.

Images