Wind power prediction method, device and equipment for floating type offshore wind plant

Abstract

The invention discloses a wind power prediction method, device and equipment for a floating type offshore wind plant, relates to the technical field of wind power prediction, and can solve the technical problem that the wind power is difficult to accurately predict in a floating type offshore wind plant scene. Comprising the steps that a three-dimensional time wake flow model of the floating type wind turbine is constructed according to the inflow wind speed, inflow environment turbulence intensity and working conditions, and the wake flow wind speed is obtained; the wind power at the previous moment is obtained, and the environment state is determined according to the inflow wind speed, the inflow environment turbulence intensity, the working condition, the wind power at the previous moment and the wake flow wind speed; and wind power prediction is carried out on the environment state based on a trained strategy network, target wind power at the current moment is obtained, and the strategy network is constructed based on a trained convolutional neural network and a trained long-short term memory network.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power prediction, and in particular to a method, device and equipment for predicting wind power of a floating offshore wind farm. Background Art

[0002] As an important member of clean energy, wind energy has become the focus of development and utilization due to its large reserves and renewable advantages. However, wind energy is an intermittent resource with poor stability and continuity and high prediction difficulty, which makes it difficult to meet the grid connection requirements. Improving the accuracy of wind power prediction is crucial to ensure the safe and stable operation of the power system.

[0003] As the exploitable areas of onshore wind resources gradually become saturated, wind farms are gradually developing in offshore areas. Compared with fixed onshore wind turbines, offshore floating wind turbines have a more complex working environment. The unique movement of floating platforms makes the flow field distribution more difficult to predict. Therefore, there is an urgent need for a method to accurately predict wind power in floating offshore wind farm scenarios. Summary of the invention

[0004] In view of this, the present invention provides a method, device and equipment for predicting wind power of a floating offshore wind farm, which can solve the technical problem that it is difficult to accurately predict wind power in a floating offshore wind farm scenario.

[0005] According to a first aspect of the present invention, a method for predicting wind power of a floating offshore wind farm is provided, the method comprising:

[0006] A three-dimensional time wake model of a floating wind turbine is constructed according to the inflow wind speed, the turbulence intensity of the inflow environment and the working conditions, and the wake wind speed is obtained;

[0007] Obtaining the wind power at the last moment, and determining the environmental state according to the inflow wind speed, the inflow environmental turbulence intensity, the operating condition, the wind power at the last moment, and the wake wind speed;

[0008] Based on the trained strategy network, the wind power of the environmental state is predicted to obtain the target wind power at the current moment, wherein the strategy network is constructed based on the trained convolutional neural network and the trained long short-term memory network.

[0009] Preferably, the three-dimensional time wake model of the floating wind turbine is constructed according to the inflow wind speed, the inflow environment turbulence intensity and the working conditions, including:

[0010] A three-dimensional double cosine wake model is constructed based on the inflow wind speed and the inflow environment turbulence intensity;

[0011] Constructing a three-dimensional wake model of a floating wind turbine under the action of wind and waves according to the working conditions and the three-dimensional double cosine wake model;

[0012] The wake delay time is calculated and added to the three-dimensional wake model to obtain a three-dimensional time wake model of the floating wind turbine.

[0013] Preferably, constructing a three-dimensional double cosine wake model according to the inflow wind speed and the inflow environment turbulence intensity includes:

[0014] The Jensen wake model is constructed according to the turbulence intensity of the inflow environment;

[0015] Construct a double cosine model when the wake wind speed distribution of a floating wind turbine is in a double cosine shape;

[0016] A three-dimensional double cosine wake model is constructed according to the Jensen wake model, the double cosine model and the inflow wind speed.

[0017] Preferably, the three-dimensional wake model of the floating wind turbine under the action of wind and waves is constructed according to the working conditions and the three-dimensional double cosine wake model, including:

[0018] Construct the average additional wind speed of the wind rotor plane longitudinal motion according to the working conditions;

[0019] The additional wind speed average value is added to the three-dimensional double cosine wake model to obtain a three-dimensional wake model of a floating wind turbine under the action of wind and waves.

[0020] Preferably, the wind power prediction based on the trained strategy network for the environmental state to obtain the target wind power at the current moment includes:

[0021] Inputting the environmental state into a convolutional neural network to obtain an output feature value;

[0022] Based on the gating mechanism of the long short-term memory network, the cell state at the previous moment is updated according to the output characteristic value to obtain the cell state at the current moment, and the target wind power at the current moment is output according to the cell state at the current moment.

[0023] Preferably, before performing wind power prediction on the environmental state based on the trained strategy network, the method further comprises:

[0024] Construct a strategy network to be trained according to the convolutional neural network to be trained and the long and short-term memory network to be trained;

[0025] Determine the wind power at a first historical moment and the historical environmental state, train the to-be-trained convolutional neural network according to the historical environmental state, and obtain a historical output feature value, wherein the historical environmental state includes the wind power at a second historical moment, and the second moment is the moment before the first moment;

[0026] Based on the gating mechanism of the long short-term memory network to be trained, the cell state at the second historical moment is updated according to the historical output characteristic value to obtain the cell state at the first historical moment, and the predicted wind power at the first historical moment is output according to the cell state at the first historical moment;

[0027] A loss function is calculated according to the wind power at the first historical moment and the predicted wind power at the first historical moment until the loss function is less than a preset threshold, thereby obtaining a trained strategy network, wherein the strategy network includes a trained convolutional neural network and a trained long short-term memory network.

[0028] Preferably, the method further comprises:

[0029] Acquire the inflow wind speed, the inflow environment turbulence intensity and the update information in the working condition, and obtain the updated wake wind speed according to the update information and the three-dimensional time wake model;

[0030] Acquire the wind power at the last moment, and determine the updated environmental state according to the updated information, the wind power at the last moment, and the updated wake wind speed;

[0031] Based on the trained strategy network, the wind power is predicted for the updated environmental state to obtain the updated target wind power at the current moment.

[0032] According to a second aspect of the present invention, there is provided a wind power prediction device for a floating offshore wind farm, the device comprising:

[0033] A construction module is used to construct a three-dimensional time wake model of a floating wind turbine according to the inflow wind speed, the turbulence intensity of the inflow environment and the working conditions to obtain the wake wind speed;

[0034] A determination module, used to obtain the wind power at the last moment, and determine the environmental state according to the inflow wind speed, the inflow environment turbulence intensity, the working condition, the wind power at the last moment, and the wake wind speed;

[0035] A prediction module is used to predict the wind power of the environmental state based on a trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on a trained convolutional neural network and a trained long short-term memory network.

[0036] According to the third aspect of the present application, there is provided a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned wind power prediction method for a floating offshore wind farm is implemented.

[0037] According to the fourth aspect of the present application, there is provided a computer device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor. When the processor executes the program, the above-mentioned wind power prediction method for a floating offshore wind farm is implemented.

[0038] By means of the above technical solution, a wind power prediction method, device and equipment for a floating offshore wind farm provided by the present invention first constructs a three-dimensional time wake model of a floating wind turbine according to the inflow wind speed, the inflow environmental turbulence intensity and the working conditions to obtain the wake wind speed; then, obtains the wind power at the previous moment, and determines the environmental state according to the inflow wind speed, the inflow environmental turbulence intensity, the working conditions, the wind power at the previous moment and the wake wind speed; finally, performs wind power prediction on the environmental state based on the trained policy network to obtain the target wind power at the current moment, wherein the policy network is constructed based on the trained convolutional neural network and the trained long short-term memory network. Through the technical solution of the present invention, since the working conditions include information in the oscillating motion process in the floating offshore wind farm scenario, the determined wake wind speed is more in line with the floating offshore wind farm scenario. Taking such a wake wind speed as a feature of the environmental state and inputting it into the policy network, the convolutional neural network extracts the spatial features in the environmental state, and the long short-term memory network extracts the time series features in the environmental state, so that the target wind power can be accurately predicted from the feature information of the floating wind turbine wake in the two dimensions of time and space. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the local application. In the drawings:

[0040] Figure 1 The flowchart of a wind power prediction method for a floating offshore wind farm provided by an embodiment of the present invention is shown;

[0041] Figure 2 The flowchart of another wind power prediction method for a floating offshore wind farm provided by an embodiment of the present invention is shown;

[0042] Figure 3 The structural diagram of a wind power prediction device for a floating offshore wind farm provided by an embodiment of the present invention is shown;

[0043] Figure 4 A schematic structural diagram of another wind power prediction device for a floating offshore wind farm provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0044] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present invention can be combined with each other without conflict.

[0045] This embodiment provides a method for predicting wind power of a floating offshore wind farm. Figure 1 As shown, the method includes:

[0046] 101. A three-dimensional time wake model of a floating wind turbine is constructed according to the inflow wind speed, the turbulence intensity of the inflow environment and the working conditions to obtain the wake wind speed.

[0047] 102. Obtain the wind power at the last moment, and determine the environmental state according to the inflow wind speed, the inflow environmental turbulence intensity, the working condition, the wind power at the last moment, and the wake wind speed.

[0048] 103. Predict the wind power of the environmental state based on the trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on the trained convolutional neural network and the trained long short-term memory network.

[0049] For the embodiment steps 101-103, in order to accurately predict the wind power in the floating offshore wind farm scenario, it is necessary to extract multiple features to determine the environmental state, and then input this environmental state into the trained strategy network, and predict the wind power of the environmental state according to the strategy network. Therefore, the embodiment step 101 determines the wake wind speed according to the real-time measured inflow wind speed, the inflow environment turbulence intensity and the working condition. Since the working condition includes the information in the oscillation motion process in the floating offshore wind farm scenario, the determined wake wind speed is more in line with the floating offshore wind farm scenario. Take the wake wind speed as a feature of the environmental state, and then determine other features of the environmental state by the embodiment step 102. In the deep reinforcement learning framework (that is, the strategy network), through the convolution operation, the convolutional neural network is responsible for extracting the spatial features in the environmental state. After the convolutional neural network extracts the spatial features, the long short-term memory network is used to extract the time series features in the environmental state input. Based on the spatial features extracted by the convolutional neural network and the time series information processed by the long short-term memory network, the strategy network outputs the target wind power at the current moment.

[0050] The present invention provides a method, device and equipment for predicting wind power of a floating offshore wind farm. First, a three-dimensional time wake model of a floating wind turbine is constructed according to the inflow wind speed, the inflow environment turbulence intensity and the working condition to obtain the wake wind speed; then, the wind power at the previous moment is obtained, and the environmental state is determined according to the inflow wind speed, the inflow environment turbulence intensity, the working condition, the wind power at the previous moment and the wake wind speed; finally, the wind power is predicted for the environmental state based on a trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on a trained convolutional neural network and a trained long short-term memory network. Through the technical solution of the present invention, since the working conditions include information on the oscillation motion process in the floating offshore wind farm scenario, the determined wake wind speed is more in line with the floating offshore wind farm scenario. Such wake wind speed is used as a feature of the environmental state and input into the strategy network. The spatial features in the environmental state are extracted by the convolutional neural network, and the time series features in the environmental state are extracted by the long short-term memory network. Therefore, the target wind power can be accurately predicted from the characteristic information of the wake of the floating wind turbine in the two dimensions of time and space.

[0051] Further, as a refinement and extension of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, another method for predicting wind power of a floating offshore wind farm is provided, such as Figure 2 As shown, the method includes:

[0052] 201. A three-dimensional double cosine wake model is constructed based on the inflow wind speed and the inflow environment turbulence intensity.

[0053] Among them, drones are used to conduct real-time measurements of the inflow wind speed and inflow environment turbulence intensity in the wind farm.

[0054] The method of constructing a three-dimensional double cosine wake model according to the inflow wind speed and the inflow environment turbulence intensity includes: constructing a Jensen wake model according to the inflow environment turbulence intensity; constructing a double cosine model when the wake wind speed distribution of the floating wind turbine is in a double cosine shape; and constructing a three-dimensional double cosine wake model according to the Jensen wake model, the double cosine model and the inflow wind speed.

[0055] For this embodiment, the floating wind turbine is taken as a reference when it is not affected by waves and surges and is at the starting position of a wave cycle, the center of the hub is used as the coordinate origin, the x direction represents the direction of the inflow wind speed (also the x direction of the inflow wind speed), the y direction represents the horizontal direction radially perpendicular to the x-axis (also the y direction of the inflow wind speed), and the z direction represents the vertical direction perpendicular to the xoy plane (also the z direction of the inflow wind speed).

[0056] For this embodiment, the Jensen wake model is constructed according to the turbulence intensity of the inflow environment:

[0057]

[0058] Where a is the axial induction factor of the floating wind turbine, r w is the wake radius of the floating wind turbine, r d is the rotor radius, S is the normalized distance downstream of the floating wind turbine, C t is the thrust coefficient of the floating wind turbine, I 0 is the turbulence intensity of the inflow environment at the hub height.

[0059] For this embodiment, a double cosine model is constructed when the wake wind speed distribution of the floating wind turbine is in a double cosine shape:

[0060]

[0061] Among them, u 0 is the inflow wind speed at the hub height of the floating wind turbine, z represents the vertical direction perpendicular to the xoy plane, z h is the hub height of the floating wind turbine, α is the wind shear index, r is the radial distance from the center of the wind rotor, θ is the distance between the centerline of the wake and the point where the wake velocity has the maximum loss, θ = 0.6r d , r d is the rotor radius, r′ w is the distance between the point with the maximum wake velocity loss and the wake boundary, r′ w =r w -θ, r w is the wake radius of the floating wind turbine, k is the period of the double cosine function, k = π / 2r′ w .

[0062] For this embodiment, a three-dimensional double cosine wake model is constructed according to the Jensen wake model, the double cosine model and the inflow wind speed, including: the mass flux of the double cosine model is equal to the mass flux of the Jensen wake model, and the following first relationship can be obtained through the law of conservation of mass:

[0063]

[0064] Combining the Jensen wake model, the double cosine model and the first relationship, we get the three-dimensional double cosine wake model:

[0065]

[0066] where u(x,y,z) is the initial wake wind speed.

[0067] 202. Construct a three-dimensional wake model of a floating wind turbine under the action of wind and waves according to the working conditions and the three-dimensional double cosine wake model.

[0068] The method of constructing a three-dimensional wake model of a floating wind turbine under the action of wind and waves according to the working conditions and the three-dimensional double cosine wake model includes: constructing an additional wind speed average value of the longitudinal oscillation motion of the wind wheel plane according to the working conditions; adding the additional wind speed average value to the three-dimensional double cosine wake model to obtain the three-dimensional wake model of the floating wind turbine under the action of wind and waves.

[0069] For this embodiment, the additional wind speed average value of the wind rotor plane longitudinal motion is constructed according to the working conditions:

[0070] It should be noted that the longitudinal oscillation motion is caused by the longitudinal oscillation motion of the wind wheel plane caused by wave motion in the floating offshore wind farm scenario. The three-dimensional wake model of the floating wind turbine under the action of wind and waves is obtained by adding the average wind speed to the three-dimensional double cosine wake model. The wake wind speed obtained by step 203 of the embodiment is more suitable for the floating offshore wind farm scenario, and the wake wind speed is used as part of the environmental state of step 204 of the embodiment. After the trained strategy network, the target wind power at the current moment is more accurate.

[0071] The working conditions include: the amplitude of the longitudinal motion, the frequency of the longitudinal motion, the phase angle in the initial stage, the average value of the longitudinal motion displacement and the period of the longitudinal motion.

[0072]

[0073] Among them, A s is the amplitude of the oscillatory motion, f s is the frequency of the oscillatory motion, is the phase angle at the initial stage, is the average displacement of the longitudinal motion, T s is the period of the oscillatory motion, T 1 and T 2 They represent the start and end time of the oscillation motion in the selected time period, n is the number of cycles, is the average additional wind speed.

[0074] For this embodiment, the additional wind speed average value is added to the three-dimensional double cosine wake model to obtain a three-dimensional wake model of a floating wind turbine under the action of wind and waves:

[0075]

[0076] 203. Calculate the wake delay time, add the wake delay time to the three-dimensional wake model, obtain the three-dimensional time wake model of the floating wind turbine, and obtain the wake wind speed.

[0077] For this embodiment, the wake delay time is calculated as:

[0078]

[0079] Among them, τ ij is the wake delay time of wind from the i-th unit to the j-th unit.

[0080] For this embodiment, the wake delay time is added to the three-dimensional wake model to obtain a three-dimensional time wake model of the floating wind turbine:

[0081]

[0082] Where T is the propagation time of wind from the upstream wind turbine to the downstream wind turbine at time t, u(x,y,z,T) is the wake wind speed, which takes into account τ compared to u(x,y,z) (u(x,y,z) is the initial wake wind speed). ij , that is, the wake delay time of the wind from the i-th unit to the j-th unit, that is, the three-dimensional time wake model is obtained by correcting the three-dimensional wake model using the wake delay time.

[0083] It should be noted that the obtained wake wind speed u(x, y, z, T) represents the flow field characteristics of wind farms with different time and space dimensions.

[0084] 204. Obtain the wind power at the last moment, and determine the environmental state according to the inflow wind speed, the inflow environmental turbulence intensity, the working condition, the wind power at the last moment, and the wake wind speed.

[0085] For this embodiment, as an implementation method, a multi-source data set is constructed, wherein the multi-source data set includes real-time measured data (the real-time measured data is the inflow wind speed u 0 and the inflow ambient turbulence intensity I 0 ), historical data (historical data is the wind power P at the last moment t-1 ) and operating condition φ t (Working condition is the amplitude A of the longitudinal motion s , the frequency of the oscillatory motion f s , the phase angle at the initial stage The average value of the oscillatory motion displacement and the period of the oscillatory motion T s ), the wake wind speed calculated by the multi-source data set and steps 201-203 of the embodiment is determined as the environmental state:

[0086] S t =[u 0 ,I 0 ,φ t ,P t-1 ,u(x,y,z,T)]

[0087] It should be noted that before constructing a multi-source dataset, the data in the multi-source dataset needs to be preprocessed, including data cleaning, format conversion, and time alignment.

[0088] 205. Predict the wind power of the environmental state based on the trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on the trained convolutional neural network and the trained long short-term memory network.

[0089] Before the trained strategy network predicts the wind power for the environmental state, the method further includes: constructing a strategy network to be trained based on the convolutional neural network to be trained and the long short-term memory network to be trained; determining the wind power at the first historical moment and the historical environmental state, training the convolutional neural network to be trained based on the historical environmental state, and obtaining a historical output characteristic value, wherein the historical environmental state includes the wind power at the second historical moment, and the second moment is the previous moment of the first moment; based on the gating mechanism of the long short-term memory network to be trained, updating the cell state at the second historical moment according to the historical output characteristic value to obtain the cell state at the first historical moment, and outputting the predicted wind power at the first historical moment according to the cell state at the first historical moment; calculating the loss function based on the wind power at the first historical moment and the predicted wind power at the first historical moment, until the loss function is less than a preset threshold, and obtaining a trained strategy network, wherein the strategy network includes the trained convolutional neural network and the trained long short-term memory network.

[0090] Among them, the historical environmental status includes the historical inflow wind speed, the historical inflow environmental turbulence intensity, the historical second moment wind power and the historical operating conditions. The historical second moment wind power is the actual wind power at the second moment, and the historical first moment wind power is the actual wind power at the first moment. The convolutional neural network to be trained and the long short-term memory network to be trained are trained based on the historical environmental status including the historical second moment wind power. The predicted wind power at the first moment in history is the predicted wind power at the first moment. By calculating the loss function of the wind power at the first moment in history and the predicted wind power at the first moment in history, if the loss function is greater than or equal to the preset threshold, then continue training until the loss function is less than the preset threshold, and a trained strategy network is obtained.

[0091] The specific training process is as follows: (1) training the convolutional neural network to be trained according to the historical environmental state to obtain historical output feature values. Specifically:

[0092]

[0093] In the formula, Z i,j,k is the historical output feature value of the kth convolution kernel at position (i, j); S is the historical environment state; W is the convolution kernel, whose size is M×N; b k is the bias term of the kth convolution kernel.

[0094] (2) Based on the gating mechanism of the long short-term memory network to be trained, the cell state at the second historical moment is updated according to the historical output characteristic value to obtain the cell state at the first historical moment, and the predicted wind power at the first historical moment is output according to the cell state at the first historical moment. Specifically:

[0095] The cell state update formula is as follows:

[0096] C t =f t ·C t-1 +i t ·C tx

[0097] In the formula, C t is the cell state at the first moment in history, C t It is through the gating mechanism that C is dynamically adjusted according to the historical output feature value. t-1 Updated; C t-1 is the cell state at the second moment in history; f t is the forget gate; i t is the input gate; C tx is a new candidate cell state.

[0098] The predicted wind power at the first historical moment is output according to the cell state at the first historical moment. Specifically, the cell state at the first historical moment captures and stores long-term time-dependent information, so as to assist the hidden state of the long short-term memory network to be trained to output the predicted wind power at the first historical moment.

[0099] The wind power prediction for the environmental state based on the trained strategy network is performed to obtain the target wind power at the current moment, including: inputting the environmental state into a convolutional neural network to obtain an output feature value; based on the gating mechanism of the long short-term memory network, updating the cell state at the previous moment according to the output feature value to obtain the cell state at the current moment, and outputting the target wind power at the current moment according to the cell state at the current moment.

[0100]

[0101] In the formula, Z i,j,kis the output feature value of the kth convolution kernel at position (i, j); S is the environment state; W is the convolution kernel, whose size is M×N; b k is the bias term of the kth convolution kernel.

[0102] The cell state update formula is as follows:

[0103] C t =f t ·C t-1 +i t ·C tx

[0104] In the formula, C t is the cell state at the current moment, C t It is through the gating mechanism that C is dynamically adjusted according to the output eigenvalue. t-1 Updated; C t-1 is the cell state at the previous moment; f t is the forget gate; i t is the input gate; C tx is a new candidate cell state.

[0105] Preferably, after using the strategy network to predict the wind power of the environmental state and obtaining the target wind power at the current moment, the environmental state can also be input into the value network. The value network can provide the strategy network with a reward V(s) about the current state by optimizing the decision path. t ), which is the value of the state, to help the policy network make better choices in the decision-making process. In the process of policy optimization, the experience replay technology is used to store historical data in the experience pool. By randomly extracting samples for training, the model is prevented from over-relying on the latest data, thereby enhancing the stability of learning. In addition, the "ε-greedy strategy" is used to help the model strike a balance between exploring new strategies and using the currently known optimal strategy. The formula is as follows:

[0106]

[0107] In the formula, a t is the action of the model at time t; Q(s t ,a) is the current state s t and the action value function for action a; ε is the exploration rate of the model; argmaxQ(s t ,a) means selecting the action that can bring the maximum expected reward in the current state; random action means selecting a random action with probability ε.

[0108] The policy network optimizes the decision-making strategy by maximizing long-term returns, while the value network minimizes the error between the current state value and the target value through temporal difference learning, thereby continuously optimizing the estimation of state returns. Through the above methods, the policy network and the value network can be continuously optimized to achieve higher prediction accuracy and stronger robustness.

[0109] 206. Obtain the inflow wind speed, the inflow environment turbulence intensity, and update information in the working condition, and obtain an updated wake wind speed according to the update information and the three-dimensional time wake model.

[0110] 207. Obtain the wind power at the last moment, and determine an updated environmental state according to the update information, the wind power at the last moment, and the updated wake wind speed.

[0111] 208. Perform wind power prediction on the updated environmental state based on the trained strategy network to obtain an updated target wind power at the current moment.

[0112] For example embodiment steps 201-205, the inflow wind speed, inflow environment turbulence intensity and operating conditions are all measured in real time, and the determined wake wind speed is also in real time. For example embodiment steps 206-208, if the inflow wind speed, inflow environment turbulence intensity and operating conditions have not changed, then the determined wake wind speed can be used directly. As long as at least one of the inflow wind speed, inflow environment turbulence intensity and operating conditions has changed, then it is necessary to redetermine the wake wind speed to obtain an updated wake wind speed. Accordingly, the environmental state is redetermined to obtain an updated environmental state. When performing wind power prediction, it is necessary to predict the updated environmental state, so that the updated target wind power obtained is in line with the actual situation and has high accuracy.

[0113] The present invention provides a method, device and equipment for predicting wind power of a floating offshore wind farm. First, a three-dimensional time wake model of a floating wind turbine is constructed according to the inflow wind speed, the inflow environment turbulence intensity and the working condition to obtain the wake wind speed; then, the wind power at the previous moment is obtained, and the environmental state is determined according to the inflow wind speed, the inflow environment turbulence intensity, the working condition, the wind power at the previous moment and the wake wind speed; finally, the wind power is predicted for the environmental state based on a trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on a trained convolutional neural network and a trained long short-term memory network. Through the technical solution of the present invention, since the working conditions include information on the oscillation motion process in the floating offshore wind farm scenario, the determined wake wind speed is more in line with the floating offshore wind farm scenario. Such wake wind speed is used as a feature of the environmental state and input into the strategy network. The spatial features in the environmental state are extracted by the convolutional neural network, and the time series features in the environmental state are extracted by the long short-term memory network. Therefore, the target wind power can be accurately predicted from the characteristic information of the wake of the floating wind turbine in the two dimensions of time and space.

[0114] Further, as Figure 1 and Figure 2 The specific implementation of the method shown in the figure, the embodiment of the present invention provides a wind power prediction device for a floating offshore wind farm, such as Figure 3 As shown, the device includes: a construction module 31, a determination module 32, and a prediction module 33;

[0115] A construction module 31 is used to construct a three-dimensional time wake model of the floating wind turbine according to the inflow wind speed, the inflow environment turbulence intensity and the working conditions to obtain the wake wind speed;

[0116] A determination module 32, configured to obtain the wind power at the last moment, and determine the environmental state according to the inflow wind speed, the inflow environmental turbulence intensity, the operating condition, the wind power at the last moment, and the wake wind speed;

[0117] The prediction module 33 is used to predict the wind power of the environmental state based on the trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on the trained convolutional neural network and the trained long short-term memory network.

[0118] Accordingly, in order to construct a three-dimensional time wake model of a floating wind turbine according to the inflow wind speed, the inflow environment turbulence intensity and the operating conditions, a construction module 31 can be specifically used to construct a three-dimensional double cosine wake model according to the inflow wind speed and the inflow environment turbulence intensity; construct a three-dimensional wake model of a floating wind turbine under the action of wind and waves according to the operating conditions and the three-dimensional double cosine wake model; calculate the wake delay time, add the wake delay time to the three-dimensional wake model, and obtain the three-dimensional time wake model of the floating wind turbine.

[0119] Accordingly, in order to construct a three-dimensional double cosine wake model according to the inflow wind speed and the inflow environment turbulence intensity, a construction module 31 can be specifically used to construct a Jensen wake model according to the inflow environment turbulence intensity; construct a double cosine model when the wake wind speed distribution of the floating wind turbine is in a double cosine shape; and construct a three-dimensional double cosine wake model according to the Jensen wake model, the double cosine model and the inflow wind speed.

[0120] Accordingly, in order to construct a three-dimensional wake model of a floating wind turbine under the action of wind and waves according to the operating conditions and the three-dimensional double cosine wake model, a construction module 31 can be specifically used to construct an additional wind speed average value of the longitudinal oscillation motion of the wind wheel plane according to the operating conditions; the additional wind speed average value is added to the three-dimensional double cosine wake model to obtain a three-dimensional wake model of a floating wind turbine under the action of wind and waves.

[0121] Correspondingly, in order to predict the wind power of the environmental state based on the trained strategy network and obtain the target wind power at the current moment, the prediction module 33 can be specifically used to input the environmental state into the convolutional neural network to obtain the output characteristic value; based on the gating mechanism of the long short-term memory network, the cell state at the previous moment is updated according to the output characteristic value to obtain the cell state at the current moment, and the target wind power at the current moment is output according to the cell state at the current moment.

[0122] In specific application scenarios, such as Figure 4As shown, a wind power prediction device for a floating offshore wind farm, the device also includes: a training module 34, which can be specifically used to construct a strategy network to be trained according to the convolutional neural network to be trained and the long short-term memory network to be trained; determine the wind power at the first historical moment and the historical environmental state, train the convolutional neural network to be trained according to the historical environmental state, and obtain a historical output characteristic value, wherein the historical environmental state includes the wind power at the second historical moment, and the second moment is the previous moment of the first moment; based on the gating mechanism of the long short-term memory network to be trained, update the cell state at the second historical moment according to the historical output characteristic value to obtain the cell state at the first historical moment, and output the predicted wind power at the first historical moment according to the cell state at the first historical moment; calculate the loss function according to the wind power at the first historical moment and the predicted wind power at the first historical moment, until the loss function is less than a preset threshold, and obtain a trained strategy network, wherein the strategy network includes the trained convolutional neural network and the trained long short-term memory network.

[0123] In specific application scenarios, such as Figure 4 As shown, a wind power prediction device for a floating offshore wind farm, the device also includes: an updating module 35, which can be specifically used to obtain the inflow wind speed, the inflow environment turbulence intensity and the update information in the working condition, and obtain the updated wake wind speed according to the update information and the three-dimensional time wake model; obtain the wind power at the previous moment, and determine the updated environment state according to the update information, the wind power at the previous moment and the updated wake wind speed; predict the wind power of the updated environment state based on the trained strategy network to obtain the updated target wind power at the current moment.

[0124] It should be noted that for other corresponding descriptions of the functional units involved in the wind power prediction device for a floating offshore wind farm provided in this embodiment, reference can be made to Figure 1 to Figure 2 The corresponding description will not be repeated here.

[0125] Based on the above Figure 1 to Figure 2 The method shown in the embodiment, accordingly, also provides a storage medium, which can be volatile or non-volatile, and stores a computer program, which is executed by a processor to implement the above-mentioned Figure 1 to Figure 2 The wind power prediction method for floating offshore wind farms is shown.

[0126] Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of various implementation scenarios of the present invention.

[0127] Based on the above Figure 1 to Figure 2 The method shown and Figure 3 , Figure 4 In order to achieve the above-mentioned purpose, the present embodiment further provides a computer device, which includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 to Figure 2 The wind power prediction method for floating offshore wind farms is shown.

[0128] Optionally, the computer device may further include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0129] Those skilled in the art will appreciate that the computer device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different component arrangements.

[0130] The storage medium may also include an operating network communication module. The operating is a program that manages the hardware and software resources of the above-mentioned computer device, supporting the operation of the information processing program and other software and / or programs. The network communication module is used to communicate between the components inside the storage medium, and to communicate with other hardware and software in the information processing entity device.

[0131] Through the description of the above implementation modes, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform, or by hardware.

[0132] The present invention provides a method, device and equipment for predicting wind power of a floating offshore wind farm. First, a three-dimensional time wake model of a floating wind turbine is constructed according to the inflow wind speed, the inflow environment turbulence intensity and the working condition to obtain the wake wind speed; then, the wind power at the previous moment is obtained, and the environmental state is determined according to the inflow wind speed, the inflow environment turbulence intensity, the working condition, the wind power at the previous moment and the wake wind speed; finally, the wind power is predicted for the environmental state based on a trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on a trained convolutional neural network and a trained long short-term memory network. Through the technical solution of the present invention, since the working conditions include information on the oscillation motion process in the floating offshore wind farm scenario, the determined wake wind speed is more in line with the floating offshore wind farm scenario. Such wake wind speed is used as a feature of the environmental state and input into the strategy network. The spatial features in the environmental state are extracted by the convolutional neural network, and the time series features in the environmental state are extracted by the long short-term memory network. Therefore, the target wind power can be accurately predicted from the characteristic information of the wake of the floating wind turbine in the two dimensions of time and space.

[0133] Those skilled in the art will appreciate that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily necessary for implementing the present invention. Those skilled in the art will appreciate that the modules in the devices in the implementation scenario can be distributed in the devices of the implementation scenario according to the description of the implementation scenario, or can be changed accordingly and located in one or more devices different from the present implementation scenario. The modules of the above-mentioned implementation scenario can be combined into one module, or can be further split into multiple submodules.

[0134] The above serial numbers of the present invention are only for description and do not represent the advantages and disadvantages of the implementation scenarios. The above disclosure is only a few specific implementation scenarios of the present invention, but the present invention is not limited thereto, and any changes that can be thought of by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A method for predicting wind power of a floating offshore wind farm, characterized in that: The method comprises: A three-dimensional time wake model of a floating wind turbine is constructed according to the inflow wind speed, the turbulence intensity of the inflow environment and the working conditions, and the wake wind speed is obtained; Obtaining the wind power at the last moment, and determining the environmental state according to the inflow wind speed, the inflow environmental turbulence intensity, the operating condition, the wind power at the last moment, and the wake wind speed; Based on the trained strategy network, the wind power of the environmental state is predicted to obtain the target wind power at the current moment, wherein the strategy network is constructed based on the trained convolutional neural network and the trained long short-term memory network.

2. The method according to claim 1, characterized in that The three-dimensional time wake model of the floating wind turbine is constructed according to the inflow wind speed, the inflow environment turbulence intensity and the working conditions, including: A three-dimensional double cosine wake model is constructed based on the inflow wind speed and the inflow environment turbulence intensity; Constructing a three-dimensional wake model of a floating wind turbine under the action of wind and waves according to the working conditions and the three-dimensional double cosine wake model; The wake delay time is calculated and added to the three-dimensional wake model to obtain a three-dimensional time wake model of the floating wind turbine.

3. The method according to claim 2, characterized in that The three-dimensional double cosine wake model is constructed according to the inflow wind speed and the inflow environment turbulence intensity, including: The Jensen wake model is constructed according to the turbulence intensity of the inflow environment; Construct a double cosine model when the wake wind speed distribution of a floating wind turbine is in a double cosine shape; A three-dimensional double cosine wake model is constructed according to the Jensen wake model, the double cosine model and the inflow wind speed.

4. The method according to claim 2, characterized in that: The three-dimensional wake model of the floating wind turbine under the action of wind and waves is constructed according to the working conditions and the three-dimensional double cosine wake model, including: Construct the average additional wind speed of the wind rotor plane longitudinal motion according to the working conditions; The additional wind speed average value is added to the three-dimensional double cosine wake model to obtain a three-dimensional wake model of a floating wind turbine under the action of wind and waves.

5. The method according to claim 1, characterized in that The trained strategy network predicts the wind power of the environmental state to obtain the target wind power at the current moment, including: Inputting the environmental state into a convolutional neural network to obtain an output feature value; Based on the gating mechanism of the long short-term memory network, the cell state at the previous moment is updated according to the output characteristic value to obtain the cell state at the current moment, and the target wind power at the current moment is output according to the cell state at the current moment.

6. The method according to claim 1, characterized in that Before performing wind power prediction on the environmental state based on the trained strategy network, the method further includes: Construct a strategy network to be trained according to the convolutional neural network to be trained and the long and short-term memory network to be trained; Determine the wind power at a first historical moment and the historical environmental state, train the to-be-trained convolutional neural network according to the historical environmental state, and obtain a historical output feature value, wherein the historical environmental state includes the wind power at a second historical moment, and the second moment is the moment before the first moment; Based on the gating mechanism of the long short-term memory network to be trained, the cell state at the second historical moment is updated according to the historical output characteristic value to obtain the cell state at the first historical moment, and the predicted wind power at the first historical moment is output according to the cell state at the first historical moment; A loss function is calculated according to the wind power at the first historical moment and the predicted wind power at the first historical moment until the loss function is less than a preset threshold, thereby obtaining a trained strategy network, wherein the strategy network includes a trained convolutional neural network and a trained long short-term memory network.

7. The method according to claim 1, characterized in that The method further comprises: Acquire the inflow wind speed, the inflow environment turbulence intensity and the update information in the working condition, and obtain the updated wake wind speed according to the update information and the three-dimensional time wake model; Acquire the wind power at the last moment, and determine the updated environmental state according to the updated information, the wind power at the last moment, and the updated wake wind speed; Based on the trained strategy network, the wind power is predicted for the updated environmental state to obtain the updated target wind power at the current moment.

8. A wind power prediction device for a floating offshore wind farm, characterized in that: The device comprises: A construction module is used to construct a three-dimensional time wake model of a floating wind turbine according to the inflow wind speed, the turbulence intensity of the inflow environment and the working conditions to obtain the wake wind speed; A determination module, used to obtain the wind power at the last moment, and determine the environmental state according to the inflow wind speed, the inflow environment turbulence intensity, the working condition, the wind power at the last moment, and the wake wind speed; A prediction module is used to predict the wind power of the environmental state based on a trained strategy network to obtain the target wind power at the current moment, wherein the strategy network is constructed based on a trained convolutional neural network and a trained long short-term memory network.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting wind power for a floating offshore wind farm according to any one of claims 1 to 7 is implemented.

10. A computer device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the wind power prediction method for the floating offshore wind farm according to any one of claims 1 to 7 is implemented.

Images