Big data side platform and method

Abstract

The invention belongs to the technical field of wind power big data, and discloses a big data side platform and method. Comprising the steps of collecting operation data and forecast environment data of a wind driven generator; constructing an environment data prediction model, and predicting the forecast environment data according to the environment data prediction model to obtain predicted environment data; adjusting the angle of the blade of the wind driven generator according to the predicted environment data; performing fault detection on the wind driven generator according to the adjusted angle of the blade, the predicted environment data and the operation data; the wind energy utilization rate of the wind driven generator is further improved, the service life of the wind driven generator is prolonged, faults of the wind driven generator are found quickly and accurately, the wind driven generator is prevented from being damaged secondarily, and the power generation efficiency of the wind driven generator is improved.

Description

Technical Field

[0001] The present invention relates to the field of wind power big data technology, and more specifically, to a big data edge platform and method. Background Art

[0002] The patent application with the publication number CN118881525A discloses a wind turbine performance detection system, which includes: a wind simulation room, which is used to provide a wind simulation environment for the wind turbine; an analysis center, which is used to analyze the state parameters, obtain the rotation performance characteristic value of the wind turbine, compare the rotation performance characteristic value of the wind turbine with the preset rotation performance qualified characteristic value, and if the wind turbine power generation performance is qualified, generate a signal that the wind turbine performance is qualified; the present invention provides a wind simulation environment through a wind simulation room, provides different sizes of wind to the wind turbine, analyzes the state parameters through the analysis center, determines whether the wind turbine power generation performance is qualified, and fully intelligently detects its performance, so as to ensure that all performances of the wind turbine are qualified and avoid the problem of shortened service life caused by failure of a certain performance to meet the standard.

[0003] Big data includes social media data, log data, business data, wind power data, etc. The present invention mainly processes and analyzes wind power data. The forecast environment data obtained through weather forecasts may differ from the actual environment data, thereby affecting the accuracy of wind turbine fault monitoring. Wind speed and wind direction will affect the power generation efficiency and service life of wind turbines. Failure to adjust the angle of wind turbine blades or improper adjustment will reduce the efficiency of wind turbines in utilizing wind energy, thereby reducing the power generation efficiency of wind turbines and shortening the service life of wind turbines.

[0004] In view of this, the present invention proposes a big data edge platform and method to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a big data edge platform, comprising:

[0006] Data acquisition module, used to collect wind turbine operation data and forecast environmental data;

[0007] A data prediction module is used to construct an environmental data prediction model, predict the forecast environmental data according to the environmental data prediction model, and obtain the predicted environmental data;

[0008] An angle adjustment module, used to adjust the angle of the wind turbine blades according to the predicted environmental data;

[0009] The fault detection module is used to perform fault detection on the wind turbine according to the adjusted blade angle, predicted environmental data and operating data.

[0010] Furthermore, the operating data of the wind turbine generator includes rotation speed and output power;

[0011] The forecast environmental data includes forecast wind direction and forecast wind speed.

[0012] Furthermore, the method for constructing an environmental data prediction model includes:

[0013] The environmental data prediction model includes a wind speed data prediction sub-model and a wind direction data prediction sub-model, and the predicted environmental data includes predicted wind speed and predicted wind direction;

[0014] The formula of the environmental data prediction model is obtained through M iterations:

[0015] Among them, X is the input of the environmental data prediction model, F M (X) is the output of the environmental data prediction model, M is the number of iterations, m is the index of the number of iterations, η is the learning rate, h m (X) is the regression tree at the mth iteration, F0(X) is the initial prediction value, N is the number of samples, i is the index of the sample, y i is the actual environmental data of sample i, c is a constant, argmin c [] indicates that the The smallest constant c;

[0016] The loss function of the wind speed data prediction sub-model is:

[0017] Among them, v i is the actual wind speed of sample i, v pred，i The predicted wind speed of sample i output by the wind speed data prediction sub-model;

[0018] The loss function of the wind direction data prediction sub-model is:

[0019] Among them, sin(θ i ) is the actual wind direction sinusoidal component of sample i, θ i is the actual wind direction of sample i, sin(θ pred，i ) is the predicted wind direction sinusoidal component of sample i output by the wind direction data prediction sub-model, θ pred，i is the predicted wind direction of sample i, cos(θ i ) is the actual wind direction cosine component of sample i, cos(θ pred，i) is the predicted wind direction cosine component of sample i output by the wind direction data prediction sub-model.

[0020] Furthermore, the method of obtaining the formula of the environmental data prediction model through M iterations includes:

[0021] For each iteration, get the residual coefficient

[0022] in, is the residual coefficient of sample i in the mth iteration, F m-1 (X i ) is the predicted environmental data of sample i after the m-1th iteration;

[0023] Obtain the regression tree h based on the residual coefficient m (X);

[0024] in, h m (X i ) is the predicted environmental data of the regression tree for sample i at the mth iteration, argmin h [] indicates that the The smallest regression tree h m (X);

[0025] Obtain the predicted environmental data F after the mth iteration according to the regression tree m (X);

[0026] Among them, F m (X) = F m-1 (X)+η×h m (X); F m-1 (X) is the predicted environmental data after the m-1th iteration;

[0027] The formula F of the environmental data prediction model is obtained based on the regression tree and the initial prediction value at the mth iteration M (X).

[0028] Furthermore, the method of predicting the forecast environmental data according to the environmental data prediction model to obtain the predicted environmental data includes:

[0029] Step A1: collect s_p wind speed samples and wind direction samples to form a wind speed sample set and a wind direction sample set respectively, and divide the wind speed sample set into a wind speed training set and a wind speed verification set according to a ratio of 8:2, and divide the wind direction sample set into a wind direction training set and a wind direction verification set, and input the wind speed training set and the wind direction training set into the wind speed data prediction sub-model and the wind direction data prediction sub-model respectively for forward propagation;

[0030] Step A2: Obtain the output value of the environmental data prediction model, calculate the loss value using the loss function, calculate each parameter in the model using the back propagation algorithm, and update the parameters using the gradient descent algorithm;

[0031] Step A3: Repeat A1 and A2 until the loss function value of the environmental data prediction model no longer changes, import the wind speed verification set and the wind direction verification set for verification, and when the verification is successful, obtain the trained environmental data prediction model. When the verification is unsuccessful, repeat A3;

[0032] Step A4: Input the forecast wind speed into the trained wind speed data prediction sub-model to obtain the forecast wind speed v pred , according to the forecast wind direction, the forecast wind direction sine component and the forecast wind direction cosine component are obtained, and the forecast wind direction sine component and the forecast wind direction cosine component are input into the trained wind direction data prediction sub-model to obtain the forecast wind direction sine component and the forecast wind direction cosine component, and the forecast wind direction θ is obtained according to the forecast wind direction sine component and the forecast wind direction cosine component. pred ;

[0033] Among them, θ pred =arctan2(F sin (X), F cos (X)); F sin (X) is the predicted wind direction sinusoidal component, F cos (X) is the predicted wind direction cosine component.

[0034] Furthermore, the process of collecting s_p wind speed samples and wind direction samples to form a wind speed sample set and a wind direction sample set respectively includes:

[0035] Obtain s_p historical forecast wind speeds and historical forecast wind directions, and obtain the historical actual wind speeds corresponding to the s_p historical forecast wind speeds and the historical actual wind directions corresponding to the historical forecast wind directions, take a historical forecast wind speed and the corresponding historical actual wind speed as a wind speed sample, and collect all wind speed samples to form a wind speed sample set;

[0036] According to the historical forecast wind direction, the historical forecast wind direction sine component and the historical forecast wind direction cosine component are obtained. According to the historical actual wind direction, the historical actual wind direction sine component and the historical actual wind direction cosine component are obtained. A historical forecast wind direction sine component and the corresponding historical actual wind direction sine component as well as a historical forecast wind direction cosine component and the corresponding historical actual wind direction cosine component are taken as a wind direction sample, and all wind direction samples are collected to form a wind direction sample set.

[0037] Furthermore, the method for adjusting the angle of the wind turbine blades according to the predicted environmental data includes:

[0038] The plane formed by the rotation of the wind turbine blades is recorded as the wind turbine plane. If the predicted wind direction is from the front to the rear of the wind turbine plane, the operation state of the wind turbine is recorded as positive wind operation, and the first positive wind speed threshold range (v 正 0,v 正 1) and the second positive wind speed threshold range (v 正 1, v 正 2) and the positive angle threshold range (θ 正 0, θ 正 1), (θ 正 1, θ 正 2) and (θ 正 2, θ 正 3);

[0039] adjusting the angle of the wind turbine blades according to the comparison result between the predicted wind speed and the positive wind speed threshold range;

[0040] If the predicted wind direction is from the rear to the front of the wind turbine plane, the wind turbine operation state is recorded as reverse wind operation, and the reverse wind speed threshold range (v 反 0,v 反 1) and (v 反 1, v 反 2) and the reverse angle threshold range (θ 反 0, θ 反 1), (θ 反 1, θ 反 2) and (θ 反 2, θ 反 3);

[0041] adjusting the angle of the wind turbine blades according to the comparison result between the predicted wind speed and the reverse wind speed threshold range;

[0042] When v pred <v 反 0, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 反 2, θ 反 3) Inside;

[0043] When v 反 0≤v pred ≤v 反 1, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 反 1, θ 反 2) Inside;

[0044] When v 反 1 <v pred <v 反2, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 反 0, θ 反 1) Inside;

[0045] When v pred ≥v 反 At 2:00, the operation of the wind turbine is stopped;

[0046] If the predicted wind direction is parallel to the wind turbine plane, the operation state of the wind turbine is recorded as positive wind operation, and the parallel wind speed threshold range and the parallel angle threshold range are set;

[0047] The angle of the wind turbine blades is adjusted according to the comparison result between the predicted wind speed and the parallel wind speed threshold range.

[0048] Furthermore, the method for adjusting the angle of the wind turbine blade according to the comparison result between the predicted wind speed and the positive wind speed threshold range includes:

[0049] When v pred <v 正 0, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 2, θ 正 3) Inside;

[0050] When v 正 0≤v pred ≤v 正 1, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 1, θ 正 2) Inside;

[0051] When v 正 1 <v pred <v 正 2, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 0, θ 正 1) Inside;

[0052] When v pred ≥v 正 At 2:00, the operation of the wind turbine is stopped;

[0053] v 正 0 is the minimum value in the first positive wind speed threshold range, v 正 1 is the maximum value in the first positive wind speed threshold range, v 正 2 is the maximum value in the second positive wind speed threshold range.

[0054] Furthermore, the process of performing fault detection on the wind turbine generator according to the adjusted blade angle, predicted environmental data and operating data includes:

[0055] Obtaining the angle between the adjusted wind turbine plane and the predicted wind direction and the operating state of the wind turbine generator, and recording them as the adjusted angle and the adjusted operating state respectively;

[0056] Obtain the speed and output power of p_k wind turbines that have been operating normally in the past, extract the speed and output power of wind turbines with the same adjustment angle and adjustment operation state, and record them as historical speed and historical output power respectively, obtain the minimum and maximum values ​​of the historical speed and historical output power, form a speed threshold range according to the minimum and maximum values ​​of the historical speed, and form an output power threshold range according to the minimum and maximum values ​​of the historical output power;

[0057] When the speed of the wind turbine generator after adjustment is within the speed threshold range, the speed of the wind turbine generator is normal; otherwise, the speed of the wind turbine generator is abnormal, and relevant personnel are dispatched for processing;

[0058] When the output power of the wind turbine generator after adjustment is within the output power threshold range, the output power of the wind turbine generator is normal; otherwise, the output power of the wind turbine generator is abnormal, and relevant personnel are dispatched for processing.

[0059] Big data methods, including:

[0060] Step S1: collecting the operation data and forecast environment data of the wind turbine, wherein the forecast environment data includes forecast wind direction and forecast wind speed;

[0061] Step S2: constructing an environmental data prediction model, predicting the forecast environmental data according to the environmental data prediction model, and obtaining predicted environmental data, wherein the predicted environmental data includes predicted wind speed and predicted wind direction;

[0062] Step S3: adjusting the angle of the wind turbine blades according to the predicted environmental data;

[0063] Step S4: performing fault detection on the wind turbine generator according to the adjusted blade angle, predicted environmental data and operating data.

[0064] Technical effects and advantages of the big data edge platform and method of the present invention:

[0065] An environmental data prediction model was constructed, and the forecast environmental data was predicted based on the environmental data prediction model to obtain the predicted environmental data. The forecast environmental data provided by the weather forecast often covers too large a range, and the values ​​are not targeted, and cannot provide a more accurate reference for the system. Therefore, more targeted predicted environmental data are selected to ensure the accuracy of wind turbine blade angle adjustment and fault detection. The angle of the wind turbine blades is adjusted according to the predicted environmental data, thereby improving the utilization rate of wind energy by the wind turbine, thereby improving the power generation efficiency of the wind turbine, ensuring high utilization of wind energy while not overloading the wind turbine. Fault detection of the wind turbine is carried out based on the adjusted blade angle, predicted environmental data and operating data, so as to quickly and accurately discover the fault of the wind turbine, and dispatch relevant personnel for repair, thereby avoiding secondary damage to the wind turbine and increasing the life of the wind turbine. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a schematic diagram of the big data edge platform of the present invention;

[0067] Figure 2 Schematic diagram of the big data method of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] Example 1

[0070] See also Figure 1 As shown, the big data edge platform described in this embodiment includes:

[0071] Data acquisition module, used to collect wind turbine operation data and forecast environmental data;

[0072] A data prediction module is used to construct an environmental data prediction model, predict the forecast environmental data according to the environmental data prediction model, and obtain the predicted environmental data;

[0073] An angle adjustment module, used to adjust the angle of the wind turbine blades according to the predicted environmental data;

[0074] The fault detection module is used to perform fault detection on the wind turbine according to the adjusted blade angle, predicted environmental data and operating data.

[0075] The process of collecting wind turbine operating and environmental data includes:

[0076] The operation data of the wind turbine generator includes rotation speed and output power, and the forecast environment data includes forecast wind direction and forecast wind speed;

[0077] The speed of the wind turbine is obtained through a speed sensor, and the output power of the wind turbine is obtained through a power meter;

[0078] The forecast wind direction and wind speed in the area where the wind turbine is located are obtained through weather forecast.

[0079] Methods for building environmental data prediction models include:

[0080] The environmental data prediction model includes a wind speed data prediction sub-model and a wind direction data prediction sub-model, and the predicted environmental data includes predicted wind speed and predicted wind direction;

[0081] The formula of the environmental data prediction model is obtained through M iterations:

[0082] Among them, X is the input of the environmental data prediction model, F M (X) is the output of the environmental data prediction model, M is the number of iterations, m is the index of the number of iterations, η is the learning rate, h m (X) is the regression tree at the mth iteration, F0(X) is the initial prediction value, N is the number of samples, i is the index of the sample, y i is the actual environmental data of sample i, c is a constant, argmin c [] indicates that the The smallest constant c;

[0083] The method of obtaining the formula of the environmental data prediction model through M iterations includes:

[0084] For each iteration, get the residual coefficient

[0085] in, is the residual coefficient of sample i in the mth iteration, F m-1 (X i ) is the predicted environmental data of sample i after the m-1th iteration;

[0086] It should be explained that when the first iteration is performed, F m-1 (X i )=F0(X i ); and F0(X1)=F0(X2)=……=F0(X i )=F0(X); when the second iteration is performed, Fm-1 (X i )=F1(X i ); and F1(X1)=F1(X2)=……=F1(X i ) = F1(X);

[0087] Obtain the regression tree h based on the residual coefficient m (X);

[0088] in, h m (X i ) is the predicted environmental data of the regression tree for sample i at the mth iteration, argmin h [] indicates that the The smallest regression tree h m (X);

[0089] Obtain the predicted environmental data F after the mth iteration according to the regression tree m (X);

[0090] Among them, F m (X) = F m-1 (X)+η×h m (X); F m-1 (X) is the predicted environmental data after the m-1th iteration;

[0091] The formula for the environmental data prediction model is obtained based on the regression tree and initial prediction value at the mth iteration:

[0092]

[0093] The loss function of the wind speed data prediction sub-model is:

[0094] Among them, v i is the actual wind speed of sample i, v pred，i The predicted wind speed of sample i output by the wind speed data prediction sub-model;

[0095] The loss function of the wind direction data prediction sub-model is:

[0096] Among them, sin(θ i ) is the actual wind direction sinusoidal component of sample i, θ i is the actual wind direction of sample i, sin(θ pred，i ) is the predicted wind direction sinusoidal component of sample i output by the wind direction data prediction sub-model, θ pred，i is the predicted wind direction of sample i, cos(θ i ) is the actual wind direction cosine component of sample i, cos(θ pred，i) is the predicted wind direction cosine component of sample i output by the wind direction data prediction sub-model;

[0097] The forecast environmental data is predicted according to the environmental data prediction model, and the method for obtaining the predicted environmental data includes:

[0098] Step A1: collect s_p wind speed samples and wind direction samples to form a wind speed sample set and a wind direction sample set respectively, and divide the wind speed sample set into a wind speed training set and a wind speed verification set according to a ratio of 8:2, and divide the wind direction sample set into a wind direction training set and a wind direction verification set, divide the wind speed training set into batches and sequentially input them into the wind speed data prediction sub-model for forward propagation, and divide the wind direction training set into batches and sequentially input them into the wind direction data prediction sub-model for forward propagation;

[0099] Step A2: Obtain the output value of the environmental data prediction model, calculate the loss value using the loss function, calculate each parameter in the model using the back propagation algorithm, and update the parameters using the gradient descent algorithm;

[0100] Step A3: Repeat A1 and A2 until the loss function value of the environmental data prediction model no longer changes, import the wind speed verification set and the wind direction verification set for verification, and when the verification is successful, obtain the trained environmental data prediction model. When the verification is unsuccessful, repeat A3;

[0101] Step A4: Input the forecast wind speed into the trained wind speed data prediction sub-model to obtain the forecast wind speed v pred , according to the forecast wind direction, the forecast wind direction sine component and the forecast wind direction cosine component are obtained, and the forecast wind direction sine component and the forecast wind direction cosine component are input into the trained wind direction data prediction sub-model to obtain the forecast wind direction sine component and the forecast wind direction cosine component, and the forecast wind direction θ is obtained according to the forecast wind direction sine component and the forecast wind direction cosine component. pred ;

[0102] Among them, θ pred =arctan2(F sin (X), F cos (X)); F sin (X) is the predicted wind direction sinusoidal component, F cos (X) is the predicted wind direction cosine component;

[0103] The process of collecting s_p wind speed samples and wind direction samples to form a wind speed sample set and a wind direction sample set respectively includes:

[0104] Obtain s_p historical forecast wind speeds and historical forecast wind directions, and obtain the historical actual wind speeds corresponding to the s_p historical forecast wind speeds and the historical actual wind directions corresponding to the historical forecast wind directions, take a historical forecast wind speed and the corresponding historical actual wind speed as a wind speed sample, and collect all wind speed samples to form a wind speed sample set;

[0105] According to the historical forecast wind direction, a historical forecast wind direction sine component and a historical forecast wind direction cosine component are obtained; according to the historical actual wind direction, a historical actual wind direction sine component and a historical actual wind direction cosine component are obtained; a historical forecast wind direction sine component and a corresponding historical actual wind direction sine component as well as a historical forecast wind direction cosine component and a corresponding historical actual wind direction cosine component are taken as a wind direction sample; all wind direction samples are collected to form a wind direction sample set;

[0106] It needs to be explained that the predicted wind speed and predicted wind direction of the weather forecast are not directly used as a reference. Instead, a wind speed data prediction submodel and a wind direction prediction submodel are first constructed, and the predicted wind speed and predicted wind direction are obtained based on the wind speed data prediction submodel and the wind direction prediction submodel. The predicted wind speed and predicted wind direction are used as a reference. This is because the predicted wind speed and predicted wind direction provided by the weather forecast often cover too large a range, and the values ​​are not targeted, and cannot provide a more accurate reference for the system. Therefore, the present invention uses the wind speed data prediction submodel and the wind direction prediction submodel to make a more specific prediction of the wind speed and wind direction in the wind turbine area, and at the same time continuously trains and optimizes the wind speed data prediction submodel and the wind direction prediction submodel through the actual wind speed and actual wind direction to obtain more targeted predicted wind speed and predicted wind direction.

[0107] Methods for adjusting the angle of wind turbine blades based on predicted environmental data include:

[0108] The plane formed by the rotation of the wind turbine blades is recorded as the wind turbine plane. If the predicted wind direction is from the front to the back of the wind turbine plane, the operation state of the wind turbine is recorded as positive wind operation, and the positive wind speed threshold range (v 正 0,v 正 1) and (v 正 1, v 正 2) and the positive angle threshold range (θ 正 0, θ 正 1), (θ 正 1, θ 正 2) and (θ 正 2, θ 正 3);

[0109] adjusting the angle of the wind turbine blades according to the comparison result between the predicted wind speed and the positive wind speed threshold range;

[0110] When v pred<v 正 0, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 2, θ 正 3) Inside;

[0111] When v 正 0≤v pred ≤v 正 1, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 1, θ 正 2) Inside;

[0112] When v 正 1 <v pred <v 正 2, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 0, θ 正 1) Inside;

[0113] When v pred ≥v 正 At 2:00, the operation of the wind turbine is stopped;

[0114] It needs to be explained that the angle between the wind turbine plane and the predicted wind direction is less than or equal to 90 degrees. When the wind speed is too high, the wind turbine blades will be overloaded, which will greatly reduce the life of the wind turbine and even directly damage the wind turbine, so the operation of the wind turbine is stopped;

[0115] When v pred <v 正 When v is 0, the smaller the predicted wind speed, the larger the angle between the adjusted wind turbine plane and the predicted wind direction. 正 0≤v pred ≤v 正 1, the smaller the predicted wind speed, the larger the angle between the adjusted wind turbine plane and the predicted wind direction. 正 1 <v pred <v 正 2, the greater the predicted wind speed, the smaller the angle between the adjusted wind turbine plane and the predicted wind direction. The angle of the wind turbine blades is adjusted according to different wind speeds, thereby further improving the wind energy utilization rate of the wind turbine while ensuring that the wind turbine will not be overloaded.

[0116] If the predicted wind direction is from the rear to the front of the wind turbine plane, the wind turbine operation state is recorded as reverse wind operation, and the reverse wind speed threshold range (v 反 0,v 反 1) and (v 反 1, v 反 2) and the reverse angle threshold range (θ 反0, θ 反 1), (θ 反 1, θ 反 2) and (θ 反 2, θ 反 3);

[0117] adjusting the angle of the wind turbine blades according to the comparison result between the predicted wind speed and the reverse wind speed threshold range;

[0118] When v pred <v 反 0, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 反 2, θ 反 3) Inside;

[0119] When v 反 0≤v pred ≤v 反 1, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 反 1, θ 反 2) Inside;

[0120] When v 反 1 <v pred <v 反 2, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 反 0, θ 反 1) Inside;

[0121] When v pred ≥v 反 At 2:00, the operation of the wind turbine is stopped;

[0122] It needs to be explained that when v pred <v 反 When v is 0, the smaller the predicted wind speed, the larger the angle between the adjusted wind turbine plane and the predicted wind direction. 反 0≤v pred ≤v 反 1, the smaller the predicted wind speed, the larger the angle between the adjusted wind turbine plane and the predicted wind direction. 反 1 <v pred <v 反 2, the greater the predicted wind speed, the smaller the angle between the adjusted wind turbine plane and the predicted wind direction. The angle of the wind turbine blades is adjusted according to different wind speeds, thereby further improving the wind energy utilization rate of the wind turbine while ensuring that the wind turbine will not be overloaded.

[0123] If the predicted wind direction is parallel to the wind turbine plane, the operation state of the wind turbine is recorded as positive wind operation, and the parallel wind speed threshold range and the parallel angle threshold range are set;

[0124] It should be explained that the parallel wind speed threshold range is the same as the forward wind speed threshold range, and the parallel angle threshold range is the same as the forward angle threshold range;

[0125] adjusting the angle of the wind turbine blades according to the comparison result between the predicted wind speed and the parallel wind speed threshold range;

[0126] When v pred <v 正 0, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 2, θ 正 3) and the predicted wind direction after adjustment is from the front to the rear of the wind turbine plane;

[0127] When v 正 0≤v pred ≤v 正 1, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 1, θ 正 2) and the predicted wind direction after adjustment is from the front to the rear of the wind turbine plane;

[0128] When v 正 1 <v pred <v 正 2, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 0, θ 正 1) and the predicted wind direction after adjustment is from the front to the rear of the wind turbine plane;

[0129] When v pred ≥v 正 At 2 o'clock, the operation of the wind turbine is stopped.

[0130] The process of fault detection of wind turbines based on the adjusted blade angle, predicted environmental data, and operational data includes:

[0131] Obtaining the angle between the adjusted wind turbine plane and the predicted wind direction and the operating state of the wind turbine generator, and recording them as the adjusted angle and the adjusted operating state respectively;

[0132] Obtain the speed and output power of p_k wind turbines that have been operating normally in the past, extract the speed and output power of wind turbines with the same adjustment angle and adjustment operation state, and record them as historical speed and historical output power respectively, obtain the minimum and maximum values ​​of the historical speed and historical output power, form a speed threshold range according to the minimum and maximum values ​​of the historical speed, and form an output power threshold range according to the minimum and maximum values ​​of the historical output power;

[0133] When the speed of the wind turbine generator after adjustment is within the speed threshold range, the speed of the wind turbine generator is normal;

[0134] When the speed of the wind turbine generator after adjustment is not within the speed threshold range, the speed of the wind turbine generator is abnormal, and relevant personnel are dispatched to handle it;

[0135] When the output power of the wind turbine generator after adjustment is within the output power threshold range, the output power of the wind turbine generator is normal;

[0136] When the output power of the wind turbine generator after adjustment is not within the output power threshold range, the output power of the wind turbine generator is abnormal, and relevant personnel are dispatched for processing.

[0137] In this embodiment, an environmental data prediction model is constructed, and the forecast environmental data is predicted according to the environmental data prediction model to obtain the predicted environmental data. The forecast environmental data provided by the weather forecast often covers too large a range, and the values ​​are not targeted, and cannot provide a more accurate reference for the system. Therefore, more targeted predicted environmental data are selected to ensure the accuracy of wind turbine blade angle adjustment and fault detection. The angle of the wind turbine blade is adjusted according to the predicted environmental data, thereby improving the utilization rate of wind energy by the wind turbine, and then improving the power generation efficiency of the wind turbine, ensuring high utilization of wind energy while not causing the wind turbine to be overloaded. The wind turbine is detected for faults based on the adjusted blade angle, predicted environmental data and operating data, so as to quickly and accurately discover the fault of the wind turbine, and dispatch relevant personnel for repair, thereby avoiding secondary damage to the wind turbine and increasing the life of the wind turbine.

[0138] Example 2

[0139] See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, and a big data method is provided, including:

[0140] Step S1: collecting the operation data and forecast environment data of the wind turbine, wherein the forecast environment data includes forecast wind direction and forecast wind speed;

[0141] Step S2: constructing an environmental data prediction model, predicting the forecast environmental data according to the environmental data prediction model, and obtaining predicted environmental data, wherein the predicted environmental data includes predicted wind speed and predicted wind direction;

[0142] Step S3: adjusting the angle of the wind turbine blades according to the predicted environmental data;

[0143] Step S4: performing fault detection on the wind turbine generator according to the adjusted blade angle, predicted environmental data and operating data.

[0144] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0145] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0146] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0147] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Big data edge platform, characterized by: The big data edge platform includes: Data acquisition module, used to collect wind turbine operation data and forecast environmental data; A data prediction module is used to construct an environmental data prediction model, predict the forecast environmental data according to the environmental data prediction model, and obtain the predicted environmental data; An angle adjustment module, used to adjust the angle of the wind turbine blades according to the predicted environmental data; The fault detection module is used to perform fault detection on the wind turbine according to the adjusted blade angle, predicted environmental data and operating data.

2. The big data edge platform according to claim 1, characterized in that: The operation data of the wind turbine generator includes rotation speed and output power; The forecast environmental data includes forecast wind direction and forecast wind speed.

3. The big data edge platform according to claim 2, characterized in that: The method for constructing an environmental data prediction model comprises: The environmental data prediction model includes a wind speed data prediction sub-model and a wind direction data prediction sub-model, and the predicted environmental data includes predicted wind speed and predicted wind direction; The formula of the environmental data prediction model is obtained through M iterations: Among them, X is the input of the environmental data prediction model, F M (X) is the output of the environmental data prediction model, M is the number of iterations, m is the index of the number of iterations, η is the learning rate, h m (X) is the regression tree at the mth iteration, F0(X) is the initial prediction value, N is the number of samples, i is the index of the sample, y i is the actual environmental data of sample i, c is a constant, argmin c [] indicates that the The smallest constant c; The loss function of the wind speed data prediction sub-model is: Among them, v i is the actual wind speed of sample i, v pred，i The predicted wind speed of sample i output by the wind speed data prediction sub-model; The loss function of the wind direction data prediction sub-model is: Among them, sin(θ i ) is the actual wind direction sinusoidal component of sample i, θ i is the actual wind direction of sample i, sin(θ pred，i ) is the predicted wind direction sinusoidal component of sample i output by the wind direction data prediction sub-model, θ pred，i is the predicted wind direction of sample i, cos(θ i ) is the actual wind direction cosine component of sample i, cos(θ pred，i ) is the predicted wind direction cosine component of sample i output by the wind direction data prediction sub-model.

4. The big data edge platform according to claim 3, characterized in that: The method of obtaining the formula of the environmental data prediction model through M iterations includes: For each iteration, get the residual coefficient in, is the residual coefficient of sample i in the mth iteration, F m-1 (X i ) is the predicted environmental data of sample i after the m-1th iteration; Obtain the regression tree h based on the residual coefficient m (X); in, h m (X i ) is the predicted environmental data of the regression tree for sample i at the mth iteration, argmin h [] indicates that the The smallest regression tree h m (X); Obtain the predicted environmental data F after the mth iteration according to the regression tree m (X); Among them, F m (X) = F m-1 (X)+η×h m (X); F m-1 (X) is the predicted environmental data after the m-1th iteration; The formula F of the environmental data prediction model is obtained based on the regression tree and the initial prediction value at the mth iteration M (X).

5. The big data edge platform according to claim 4, characterized in that: The method of predicting the forecast environmental data according to the environmental data prediction model to obtain the predicted environmental data includes: Step A1: collect s_p wind speed samples and wind direction samples to form a wind speed sample set and a wind direction sample set respectively, and divide the wind speed sample set into a wind speed training set and a wind speed verification set according to a ratio of 8:2, and divide the wind direction sample set into a wind direction training set and a wind direction verification set, and input the wind speed training set and the wind direction training set into the wind speed data prediction sub-model and the wind direction data prediction sub-model respectively for forward propagation; Step A2: Obtain the output value of the environmental data prediction model, calculate the loss value using the loss function, calculate each parameter in the model using the back propagation algorithm, and update the parameters using the gradient descent algorithm; Step A3: Repeat A1 and A2 until the loss function value of the environmental data prediction model no longer changes, import the wind speed verification set and the wind direction verification set for verification, and when the verification is successful, obtain the trained environmental data prediction model. When the verification is unsuccessful, repeat A3; Step A4: Input the forecast wind speed into the trained wind speed data prediction sub-model to obtain the forecast wind speed v pred , according to the forecast wind direction, the forecast wind direction sine component and the forecast wind direction cosine component are obtained, and the forecast wind direction sine component and the forecast wind direction cosine component are input into the trained wind direction data prediction sub-model to obtain the forecast wind direction sine component and the forecast wind direction cosine component, and the forecast wind direction θ is obtained according to the forecast wind direction sine component and the forecast wind direction cosine component. pred ; Among them, θ pred =arctan2(F sin (X), F cos (X)); F sin (X) is the predicted wind direction sinusoidal component, F cos (X) is the predicted wind direction cosine component.

6. The big data edge platform according to claim 5, characterized in that: The process of collecting s_p wind speed samples and wind direction samples to form a wind speed sample set and a wind direction sample set respectively comprises: Obtain s_p historical forecast wind speeds and historical forecast wind directions, and obtain the historical actual wind speeds corresponding to the s_p historical forecast wind speeds and the historical actual wind directions corresponding to the historical forecast wind directions, take a historical forecast wind speed and the corresponding historical actual wind speed as a wind speed sample, and collect all wind speed samples to form a wind speed sample set; According to the historical forecast wind direction, the historical forecast wind direction sine component and the historical forecast wind direction cosine component are obtained. According to the historical actual wind direction, the historical actual wind direction sine component and the historical actual wind direction cosine component are obtained. A historical forecast wind direction sine component and the corresponding historical actual wind direction sine component as well as a historical forecast wind direction cosine component and the corresponding historical actual wind direction cosine component are taken as a wind direction sample, and all wind direction samples are collected to form a wind direction sample set.

7. The big data edge platform according to claim 6, characterized in that: The method for adjusting the angle of the wind turbine blades according to the predicted environmental data comprises: The plane formed by the rotation of the wind turbine blades is recorded as the wind turbine plane. If the predicted wind direction is from the front to the rear of the wind turbine plane, the operation state of the wind turbine is recorded as positive wind operation, and the first positive wind speed threshold range (v 正 0,v 正 1) and the second positive wind speed threshold range (v 正 1, v 正 2) and the positive angle threshold range (θ 正 0, θ 正 1), (θ 正 1, θ 正 2) and (θ 正 2, θ 正 3); adjusting the angle of the wind turbine blades according to the comparison result between the predicted wind speed and the positive wind speed threshold range; If the predicted wind direction is from the rear to the front of the wind turbine plane, the operation state of the wind turbine is recorded as reverse wind operation, and a reverse wind speed threshold range and a reverse angle threshold range are set; adjusting the angle of the wind turbine blades according to the comparison result between the predicted wind speed and the reverse wind speed threshold range; If the predicted wind direction is parallel to the wind turbine plane, the operation state of the wind turbine is recorded as positive wind operation, and the parallel wind speed threshold range and the parallel angle threshold range are set; The angle of the wind turbine blades is adjusted according to the comparison result between the predicted wind speed and the parallel wind speed threshold range.

8. The big data edge platform according to claim 7, characterized in that: The method for adjusting the angle of the wind turbine blade according to the comparison result between the predicted wind speed and the positive wind speed threshold range comprises: When v pred <v 正 0, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 2, θ 正 3) Inside; When v 正 0≤v pred ≤v 正 1, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 1, θ 正 2) Inside; When v 正 1 <v pred <v 正 2, adjust the angle of the wind turbine blades so that the angle between the wind turbine plane and the predicted wind direction is at (θ 正 0, θ 正 1) Inside; When v pred ≥v 正 At 2:00, the operation of the wind turbine is stopped; v 正 0 is the minimum value in the first positive wind speed threshold range, v 正 1 is the maximum value in the first positive wind speed threshold range, v 正 2 is the maximum value in the second positive wind speed threshold range.

9. The big data edge platform according to claim 8, characterized in that: The process of detecting faults of the wind turbine generator according to the adjusted blade angle, predicted environmental data and operating data comprises: Obtaining the angle between the adjusted wind turbine plane and the predicted wind direction and the operating state of the wind turbine generator, and recording them as the adjusted angle and the adjusted operating state respectively; Obtain the speed and output power of p_k wind turbines that have been operating normally in the past, extract the speed and output power of wind turbines with the same adjustment angle and adjustment operation state, and record them as historical speed and historical output power respectively, obtain the minimum and maximum values ​​of the historical speed and historical output power, form a speed threshold range according to the minimum and maximum values ​​of the historical speed, and form an output power threshold range according to the minimum and maximum values ​​of the historical output power; When the speed of the wind turbine generator after adjustment is within the speed threshold range, the speed of the wind turbine generator is normal; otherwise, the speed of the wind turbine generator is abnormal, and relevant personnel are dispatched for processing; When the output power of the wind turbine generator after adjustment is within the output power threshold range, the output power of the wind turbine generator is normal; otherwise, the output power of the wind turbine generator is abnormal, and relevant personnel are dispatched for processing.

10. A big data method, used to implement the big data edge platform according to any one of claims 1 to 9, characterized in that: include: Step S1: collecting the operation data and forecast environment data of the wind turbine, wherein the forecast environment data includes forecast wind direction and forecast wind speed; Step S2: constructing an environmental data prediction model, predicting the forecast environmental data according to the environmental data prediction model, and obtaining predicted environmental data, wherein the predicted environmental data includes predicted wind speed and predicted wind direction; Step S3: adjusting the angle of the wind turbine blades according to the predicted environmental data; Step S4: performing fault detection on the wind turbine generator according to the adjusted blade angle, predicted environmental data and operating data.

Images