CN110543929B - Wind speed interval prediction method and system based on Lorenz system - Google Patents
Wind speed interval prediction method and system based on Lorenz system Download PDFInfo
- Publication number
- CN110543929B CN110543929B CN201910805484.XA CN201910805484A CN110543929B CN 110543929 B CN110543929 B CN 110543929B CN 201910805484 A CN201910805484 A CN 201910805484A CN 110543929 B CN110543929 B CN 110543929B
- Authority
- CN
- China
- Prior art keywords
- wind speed
- prediction
- interval
- prediction result
- sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 230000009194 climbing Effects 0.000 claims abstract description 36
- 238000013528 artificial neural network Methods 0.000 claims abstract description 33
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims description 24
- 239000002245 particle Substances 0.000 claims description 20
- 210000002569 neuron Anatomy 0.000 claims description 18
- 238000012360 testing method Methods 0.000 claims description 15
- 238000012937 correction Methods 0.000 claims description 14
- 238000005457 optimization Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 6
- 238000005192 partition Methods 0.000 claims description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 description 21
- 238000012706 support-vector machine Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000000739 chaotic effect Effects 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 244000144992 flock Species 0.000 description 1
- 230000019637 foraging behavior Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012731 temporal analysis Methods 0.000 description 1
- 238000000700 time series analysis Methods 0.000 description 1
- 238000000489 vacuum metal deposition Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01P—MEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
- G01P5/00—Measuring speed of fluids, e.g. of air stream; Measuring speed of bodies relative to fluids, e.g. of ship, of aircraft
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Aviation & Aerospace Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a wind speed interval prediction method and system based on a Lorenz system. The prediction method comprises the following steps: acquiring an original wind speed sequence; performing Variation Modal Decomposition (VMD) on the wind speed sequence to obtain a denoising sequence and a noise remainder; establishing a long-short time neural network prediction model (LSTM), and carrying out preliminary prediction on the denoising sequence to obtain a preliminary prediction result; correcting the preliminary prediction result by defining a wind speed climbing event (WSR) and a wind speed climbing rate to obtain a corrected wind speed prediction result; describing the influence of an atmospheric power system on wind speed through a Lorenz equation, and obtaining a Lorenz Disturbance Sequence (LDS); fitting the LDS through a B spline interpolation method, fixing a confidence interval to a fitting result, and obtaining an upper limit and a lower limit of a wind speed disturbance interval; summing the corrected wind speed prediction result and a wind speed disturbance interval to obtain an interval prediction result of the wind speed; the wind speed interval prediction method or the wind speed interval prediction system obviously improves the precision and the reliability of the prediction model, and can obtain a high-precision prediction result.
Description
Technical Field
The invention relates to the field of wind speed prediction, in particular to a wind speed interval prediction method and system based on a Lorenz system.
Background
In recent years, global energy situation is becoming severe, and energy demands are increasing. As a clean renewable energy source, wind energy has been attracting attention in new energy application around the world, and wind power grid-connected technology has become a hotspot of international research. The International Energy Agency (IEA) reports that 2019 annual wind power production is expected to increase by 3.25 million kw. However, wind fluctuations tend to cause instability of the power system after wind integration. Therefore, the effective wind speed prediction can promote the intelligent development of the wind power industry, and the accurate wind speed prediction is an important precondition for large-scale development and utilization of wind power.
Disclosure of Invention
The invention aims to provide a wind speed interval prediction method and a system based on a Lorenz system, which are used for obtaining a high-precision prediction result and improving the reliability of a prediction model.
In order to achieve the above object, the present invention provides the following solutions:
a method of predicting a wind speed interval based on a Lorenz system, the method comprising:
acquiring an original wind speed sequence; performing Variation Modal Decomposition (VMD) on the wind speed sequence to obtain a denoising sequence and a noise remainder;
establishing a long-short time neural network prediction model (LSTM), and carrying out preliminary prediction on the denoising sequence to obtain a preliminary prediction result;
correcting the preliminary prediction result by defining a wind speed climbing event (WSR) and a wind speed climbing rate to obtain a corrected wind speed prediction result;
describing the influence of an atmospheric power system on wind speed through a Lorenz equation, and obtaining a Lorenz Disturbance Sequence (LDS);
fitting the LDS through a B spline interpolation method, fixing a confidence interval to a fitting result, and obtaining an upper limit and a lower limit of a wind speed disturbance interval;
and summing the corrected wind speed prediction result and the wind speed disturbance interval to obtain an interval prediction result of the wind speed.
Optionally, the acquiring the denoising sequence and the noise remainder specifically includes:
acquiring original wind speed data;
setting the number of decomposition; iteratively computing a set containing all modes and their center frequencies under the determined number of decompositions; acquiring an intrinsic mode function component and a noise remainder of a wind speed signal;
reconstructing the eigenvalue function component of the wind speed signal to obtain a denoised wind speed sequence.
Optionally, the building of the long-short time neural network prediction model performs preliminary prediction on the denoising sequence to obtain a preliminary prediction result, and the specific process includes:
dividing the denoised wind speed sequence into a training set and a testing set according to the proportion of 9:1;
setting a network structure of a long-short time neural network, wherein the network structure comprises the number of neurons of an input layer, the number of neurons of an implicit layer and the number of neurons of an output layer;
training by inputting a training set into a long-short time neural network;
and inputting the test set into a trained long-short-time neural network to obtain a preliminary prediction result of the wind speed.
Optionally, correcting the preliminary prediction result to obtain a corrected wind speed prediction result by correcting the defined wind speed climbing event and the wind speed climbing rate, which specifically includes:
calculating a wind speed gradient; introducing a definition of wind speed climbing according to a wind speed gradient;
when the absolute value of the gradient is larger than a threshold value 1, and the positive gradient increase exceeds a positive gradient threshold value 2 or the negative gradient decrease exceeds a negative gradient threshold value 3 at the current moment, correcting the wind speed predicted value at the next moment by using the error at the current moment;
for threshold 1, threshold 2 and threshold 3, they are transformed into a multi-objective optimization problem, which is solved with Particle Swarm Optimization (PSO) with the goal of minimizing root mean square error.
Optionally, describing the influence of the aerodynamic system on the wind speed by the Lorenz equation, and obtaining the LDS specifically includes:
giving initial conditions (0, 1), solving Lorenz equation, and obtaining three-dimensional LDS;
according to the Chebyshev distance, converting the three-dimensional LDS into a one-dimensional disturbance sequence.
Optionally, the fitting is performed on the LDS by a B-spline interpolation method, and a confidence interval is fixed on a fitting result, so as to obtain an upper limit and a lower limit of a wind speed disturbance interval, which specifically includes:
fitting the distribution of the LDS through the B spline difference value to obtain a B spline difference value fitting function;
fixing confidence intervals to be 90% and 98% respectively, and calculating upper and lower quantile points of the fitting function;
the upper quantile is set as the upper limit of the interval prediction and the lower quantile is set as the lower limit of the interval prediction.
Optionally, the summing the corrected wind speed prediction result and the wind speed disturbance interval to obtain an interval prediction result of the wind speed specifically includes:
adding the upper and lower dividing points under the 90% confidence interval to the corrected wind speed prediction result to obtain a wind speed interval prediction result under the 90% confidence interval;
and adding the upper and lower dividing points under the 98% confidence interval to the corrected wind speed prediction result to obtain the wind speed interval prediction result under the 98% confidence interval.
The invention also provides a wind speed interval prediction system based on the Lorenz system, which comprises:
the wind speed data acquisition and denoising module is used for acquiring an original wind speed sequence; performing variation modal decomposition on the wind speed sequence to obtain a denoising sequence and a noise remainder;
the wind speed data preliminary prediction module is used for establishing a long-short-time neural network prediction model, and carrying out preliminary prediction on the denoising sequence to obtain a preliminary prediction result;
the preliminary prediction result correction module is used for correcting the preliminary prediction result by defining a wind speed climbing event and a wind speed climbing rate to obtain a corrected wind speed prediction result;
the LDS module is used for describing the influence of the aerodynamic system on the wind speed through a Lorenz equation and obtaining the LDS;
the disturbance upper and lower limit obtaining module is used for fitting the LDS through a B spline interpolation method, fixing a confidence interval to a fitting result and obtaining an upper limit and a lower limit of a wind speed disturbance interval;
and the prediction module is used for summing the corrected wind speed prediction result and the wind speed disturbance interval to obtain an interval prediction result of the wind speed.
Optionally, the wind speed data acquisition and denoising module specifically includes:
the intrinsic mode component unit is used for acquiring original wind speed data; setting the number of decomposition; iteratively computing a set containing all modes and their center frequencies under the determined number of decompositions; acquiring an intrinsic mode function component and a noise remainder of a wind speed signal;
the denoising wind speed sequence unit is used for reconstructing the eigenvalue function component of the wind speed signal to obtain a denoised wind speed sequence.
Optionally, the wind speed data preliminary prediction module specifically includes:
dividing and dividing a wind speed sequence unit, wherein the wind speed sequence unit is used for dividing the denoised wind speed sequence into a training set and a testing set according to the proportion of 9:1;
setting a network structure unit of a long-short time neural network, wherein the network structure unit is used for comprising the number of neurons of an input layer, the number of neurons of an hidden layer and the number of neurons of an output layer;
the training model unit is used for training by inputting a training set into the long-short-time neural network;
and the prediction unit is used for inputting the test set into the trained long-short-time neural network to obtain a preliminary wind speed prediction result.
Optionally, the preliminary prediction result correction module based on wind speed climbing specifically includes:
defining a wind speed climbing unit for calculating a wind speed gradient; introducing a definition of wind speed climbing according to a wind speed gradient;
defining a correction method unit, which is used for correcting a wind speed predicted value at the next moment by using the error at the current moment when the absolute value of the gradient is larger than a threshold value A and the positive gradient increase at the current moment exceeds a positive gradient threshold value B or the negative gradient decrease exceeds a negative gradient threshold value C;
and the solving threshold unit is used for converting the threshold values A, B and C into a multi-objective optimization problem, aiming at minimizing root mean square error, and solving the optimization problem by using PSO.
Optionally, the obtaining the LDS module specifically includes:
the method comprises the steps of obtaining a three-dimensional LDS unit, wherein the three-dimensional LDS unit is used for giving initial conditions (0, 1), solving a Lorenz equation and obtaining a three-dimensional LDS;
and acquiring a one-dimensional LDS unit, wherein the one-dimensional LDS unit is used for converting the three-dimensional LDS into the one-dimensional LDS according to the Chebyshev distance.
Optionally, the obtaining the disturbance upper and lower limit module specifically includes:
the fitting unit is used for fitting the distribution of the LDS through the B spline difference value to obtain a B spline difference value fitting function;
the quantile calculating unit is used for respectively fixing 90% and 98% confidence intervals and calculating upper and lower quantile points of the fitting function;
and determining an upper and lower interval prediction limit unit, wherein the upper partition point is set as an upper limit of interval prediction, and the lower partition point is set as a lower limit of interval prediction.
Optionally, the prediction module specifically includes:
a 90% confidence interval lower interval prediction result unit, configured to obtain a wind speed interval prediction result under the 90% confidence interval by adding the upper and lower quantiles under the 90% confidence interval to the corrected wind speed prediction result;
and the interval prediction result unit under the 98% confidence interval is used for obtaining the wind speed interval prediction result under the 98% confidence interval by adding the upper and lower dividing points under the 98% confidence interval to the corrected wind speed prediction result.
Compared with the prior art, the invention has the following technical effects:
the wind speed prediction method and the system are wind speed interval prediction processes based on Lorenz systems. Firstly, adopting a signal decomposition technology VMD to perform a noise reduction process, secondly, predicting data after noise reduction processing by using a long-short-time neural network, then introducing a wind speed gradient, correcting a preliminary prediction result to obtain a corrected wind speed prediction result, on the basis, introducing a Lorentz disturbance theory description aerodynamic system, solving a Lorenz equation to obtain LDS, adopting B spline interpolation to fit the distribution of the LDS, and obtaining the upper limit and the lower limit of interval prediction by fixing different confidence intervals of a fitting function. According to the wind speed interval prediction method, the influence of the aerodynamic system on the wind speed is considered, and the accuracy of a short-term wind speed prediction model and the reliability of a prediction result are effectively improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for predicting a wind speed interval based on Lorenz system according to embodiment 1 of the present invention;
FIG. 2 is a flowchart of a prediction method according to embodiment 2 of the present invention;
FIG. 3 is a block diagram of a prediction system according to embodiment 3 of the present invention;
FIG. 4 is a graph showing the prediction results of 6 single models on dataset 1 according to the present invention;
FIG. 5 is a graph showing the predicted results of 6 single models on dataset 2 according to the present invention;
FIG. 6 is a schematic diagram of the modified prediction result on dataset 1 according to the present invention;
FIG. 7 is a schematic representation of the modified prediction results on dataset 2 according to the present invention;
FIG. 8 is a graph of the interval prediction results of the present invention on dataset 1 when the confidence interval is 90%;
FIG. 9 is a graph of the interval prediction results of the present invention on dataset 1 when the confidence interval is 98%;
FIG. 10 is a graph of the interval prediction results of the present invention on dataset 2 when the confidence interval is 90%;
FIG. 11 is a graph showing the interval prediction results of the present invention on dataset 2 when the confidence interval is 98%.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
FIG. 1 is a flow chart of a method for predicting a wind speed interval according to the present invention. As shown in fig. 1, the method for predicting the wind speed interval based on the Lorenz system comprises the following steps:
step 11: and acquiring an original wind speed sequence, performing variation modal decomposition on the wind speed sequence, and acquiring a denoising sequence and a noise remainder.
Currently, VMDs can filter out sequences with frequency characteristics, leaving a noisy portion. In the invention, a signal processing method is utilized to change a modal decomposition method to decompose and reconstruct wind speed, and denoising is carried out to obtain a denoising sequence and a noise remainder. The method comprises the following specific steps:
step 111: the denoising sequence constructs a constraint variation model under the current mode number: first, for each modal component(A K (t) and->Respectively represent u k Instantaneous amplitude and instantaneous phase of (t), +.>Non-subtraction function, t is sampling time), and obtaining an analytic signal thereof through Hilbert transformation; then, predicting the center frequency of each analytic signal, and converting the frequency spectrum of each analytic signal to a base frequency band; and finally, estimating the bandwidth of each modal component by using Gaussian smoothing indexes of the frequency shift signals, and constructing a constraint variation model shown in the formula (1).
In the method, in the process of the invention,for the component signal u k Center frequency of (t),>representing the partial derivative of the function t, j 2 = -1, x (t) is the de-noising sequence.
Step 112: the constraint problem is changed into the unconstrained problem by introducing a quadratic penalty factor alpha and Lagrange multiplication operator lambda (t), and the extended Lagrange expression is as follows:
step 113: and obtaining a formula (3) through Fourier equidistant transformation and other processes, and realizing self-adaptive decomposition of signals.
Wherein,are respectively->λ (t), x (t). The center frequency is updated by the following formula:
step 114: and (5) iteratively updating until convergence meets the following conditions:
step 12: and establishing a long and short time neural network prediction model, and carrying out preliminary prediction on the denoising sequence to obtain a preliminary prediction result.
With the continuous development of deep learning technology, the concept of time series is applied in the structural design of Recurrent Neural Networks (RNNs). Thus, RNNs have good time series analysis capabilities. The long-short time neural network is used as an improved RNN, inherits the analysis capability of the RNN on time sequence data, and makes up the defect of the RNN in long-term memory. Because long-short time neural network models can effectively keep long-time memory, certain achievements are also achieved in the field of wind speed prediction.
Step 121: dividing the denoised wind speed sequence into a training set and a testing set according to the ratio of 9:1, wherein the first 90% of data are used as the training set, and the second 10% of data are used as the testing set;
step 122: determining the structure of a long and short time neural network, wherein the number of neurons of an input layer is 3, the number of neurons of an hidden layer is 5, and the number of neurons of an output layer is 1;
step 123: the data of the training set and the test set are respectively constructed into an input matrix according to a formula (6),
v 1 ，v 2 ，v 3 .. it represents 1,2, 3..th wind speed values;
step 124: and inputting the constructed training set input matrix into the long-short time neural network according to the rows, training the long-short time neural network, inputting the constructed testing set input matrix into the trained long-short time neural network according to the rows, and obtaining a wind speed preliminary prediction result.
Step 13: the preliminary prediction result is corrected by defining a wind speed climbing event and a wind speed climbing rate, and the corrected wind speed prediction result is obtained, which specifically comprises the following steps:
step 131: a wind speed hill climbing event is defined according to equation 7,
|v(t 0 +Δ t )-v(t 0 )|/Δ t ＞v val (7)
v(t 0 ) At t 0 Wind speed value, delta at time t Is the time interval, v val Is a threshold value for wind speed climbing;
step 132: the wind speed gradient is defined according to equation 8,
k(i)＝(v(i)-v(i-1))/imterval (8)
v (i) is the ith wind speed value, interval is the time interval;
step 133: judging whether the ith or i-1 wind speed value is greater than a wind speed climbing threshold WRR val I.e., whether equation 9 or equation 10 holds,
|k(i-1)|＞WRR val (9)
|k(i)|＞WRR val (10)
if true, the wind speed climbing correction is converted into a multi-objective optimization problem, and when the forward climbing is increased to a certain extent to exceed the forward threshold WRR up I.e., equation 11,
or the negative climbing is reduced to a certain degree to exceed the negative threshold WRR down I.e., equation 12,
the predicted value of the wind speed at the next moment is corrected according to equation 13,
representing a predicted value of wind speed before the wind speed climbing correction, < >>Representing a wind speed predicted value after wind speed climbing correction;
step 134: if neither equation 9 nor equation 10 is established, the wind speed predicted value at that time is not corrected, i.e., the wind speed predicted value is kept unchanged before and after correction.
Step 135: for the determination of the three thresholds, a particle swarm algorithm (PSO) is used;
PSO was originally proposed by Eberhart and Kennedy in 1995, and its basic concept was derived from the study of the foraging behavior of the flock. The bird individuals are simulated by using particles, each particle can be regarded as a search individual in an N-dimensional search space, the current position of the particle is a candidate solution of a corresponding optimization problem, the flying process of the particle is the search process of the individual, the flying speed of the particle can be dynamically adjusted according to the optimal position of the particle history and the optimal position of the population history, and the particle has only two attributes: speed, which represents the speed of movement, and position, which represents the direction of movement. The optimal solution searched by each particle is called an individual extremum, and the optimal individual extremum in the particle swarm is used as the current global optimal solution. The iteration is continued, updating the speed and the position. And finally obtaining the optimal solution meeting the termination condition.
Step 1351: initializing, firstly setting the maximum iteration number to be 5 times, and setting the number of independent variables of an objective function to be 3 WRRs up ，WRR down And WRR val The maximum speed of the particles is 1.5, the learning factor is 1, and the position information is the whole search spaceWe randomly initialize the velocity and position over the velocity interval and search space, setting the particle swarm size to 10, and randomly initializing one flight velocity per particle.
Step 1352: solving individual extremum and global optimal solution, defining an objective function as root mean square error between predicted wind speed and real wind speed, wherein the individual extremum is an optimal solution (pbest) found by each particle, and finding a global value (gbest) from the optimal solutions, which is called the current global optimal solution. And comparing with the historical global optimum, and updating.
Step 1353: the speed and position are updated according to equations 14 and 15,
v i ＝v i +c 1 ×rand()×(pbest i -x i )+c 2 ×rand()×(gbest i -x i ) (14)
x i ＝v i +x i (15)
i=1, 2,..n, N is the total number of particles, v i Is the velocity of the particle, rand () is a random number between (0, 1), x i Is the current position of the particle, c 1 ，c 2 Is a learning factor;
step 1354: the termination condition is that the minimum limit 1e10-8 is met when the difference between the set iteration times or algebra is reached;
step 14: the influence of the aerodynamic system on the wind speed is described by Lorenz equation, and LDS is obtained.
In 1963, the pneumologist edwardlozenges calculated the aperiodic phenomenon from a definite equation (hereinafter referred to as lorentz equation). The lorentz system is the earliest chaotic motion dissipation system found in numerical experiments. Its state equation (lorentz equation) is a simplified model of weather forecast. The method comprises the following specific steps:
step 141: establishing Lorenz equation as shown in formula 14, setting parameters sigma=10, b=8/3, r=28, and obtaining three-dimensional LDS with initial values of (0, 1);
step 142: according to the Chebyshev distance formula, performing dimension reduction on the three-dimensional LDS, and converting the three-dimensional LDS into a one-dimensional LDS;
d(C n -C 0 )＝max(|x n -x 0 |，|y n -y 0 |，|z n -z 0 |) (17)
step 15: and fitting the LDS through a B spline interpolation method, fixing a confidence interval to the fitting result, and obtaining the upper limit and the lower limit of the wind speed disturbance interval.
Step 151: drawing a frequency histogram for the LDS, and setting the number of intervals of the frequency histogram to be 25;
step 152: fitting the frequency histogram by using a B spline interpolation method to obtain a fitting function;
step 153: setting a confidence interval, obtaining an upper confidence limit and a lower confidence limit according to a fitting function, wherein the upper confidence limit takes a positive value, and the lower confidence limit takes a negative value, so as to obtain the upper limit and the lower limit of the wind speed disturbance interval;
step 16: and summing the corrected wind speed prediction result and the wind speed disturbance interval to obtain an interval prediction result of the wind speed.
Step 161: adding the upper limit of the disturbance interval to the corrected wind speed prediction result to obtain the upper limit of wind speed interval prediction;
step 162: subtracting the lower limit of the disturbance interval from the corrected wind speed prediction result to obtain the lower limit of the wind speed interval prediction
In order to verify that the method has good prediction performance on wind speed data of an actual wind power plant, a data simulation experiment is carried out on wind speed of a Sotavento wind power plant by Spain and Li Xiya, and fig. 2 is a flow chart of the prediction method provided by the embodiment. As shown in fig. 2, the specific process includes:
step 21: acquiring an original wind speed sequence
In order to better verify the prediction capability of the method provided by the invention, a prediction model is established by selecting 2 data sets of a seashore wind farm in the Sacalix of Spain. The time for dataset 1 was 2018, 9, 1, 00:00 to 2018 9, 7 days 22:40. the time for the second dataset was 2019, 3, 9, 08:00 to 2019, 3 months, 16 days 06:30. they represent wind speeds in different seasons, with time intervals of 10 minutes, 1000 wind speeds in data set 1 and 996 wind speeds in data set 2. Since there are four discrete gaps in dataset 2, we use linear interpolation as shown in equation (18) to accomplish them.
x t Represents the t time, y t The wind speed value corresponding to the moment is represented, and x is between x t And x t+i Y is its corresponding unknown wind speed value at any time in between.
For each dataset, the first 900 data (90% of the total data) were used as training sets and the last 100 data were used as test sets. That is, the predicted period is 16 hours 30 minutes.
Step 22: defining error indicators
For deterministic prediction, the most common Mean Square Error (MSE), mean Absolute Error (MAE), root Mean Square Error (RMSE) and Mean Absolute Percent Error (MAPE) are chosen for evaluation herein. Their calculation formula is shown below.
y t The true wind speed at the moment t is indicated,the predicted wind speed at time t is indicated.
For the section prediction, both aspects of section coverage (ratio of the number of sections including the actual wind speed) and average section diameter should be considered at the same time. Thus, definition formula (23) and formula (24) describe the interval coverage and average interval diameter, respectively.
N t Indicating whether the t interval covers the t real wind speed value, if N t =1, indicating that the t-th true wind speed is covered, if N t =0, indicating that the t-th true wind speed is not covered, R cover Represents the coverage of the interval d average Represents the average interval diameter. Obviously, the higher the coverage, the smaller the average diameter, the better the effect of the interval prediction, which means that the prediction accuracy is still higher when the prediction interval is smaller.
Step 23: deterministic prediction results and discussion
Step 231: first, the prediction results of 6 single wind speed prediction models, namely, an autoregressive moving average model (ARMA), a Support Vector Machine (SVM), a Gradient Boost Decision Tree (GBDT), an extreme gradient boost tree (XGBoost), a back propagation neural network (BP) and LSTM, were compared, and errors on both data sets were calculated, as shown in tables 1 and 2.
Step 232: parameters of the respective models are determined. The autocorrelation coefficient parameters p and the partial autocorrelation coefficient parameters q of the ARMA are determined according to a minimum information criterion (AIC). The most commonly used radial basis functions are chosen as the core functions of the SVM. And obtaining the learning rate and the training set number of GBDT and XGBOOST by minimizing the MAE respectively, wherein the number of the lifting trees of the two models is set to be 100. The BP structure is three layers, the node number is 3-30-1, the LSTM is also arranged as three layers, and the node number is 3-5-1. The number of hidden layer nodes set in LSTM herein is smaller than BP because too many nodes increase computation time, especially for LSTM (more than one minute would be spent when using six nodes in our programming environment), which would greatly increase computation time for the whole model, and even though LSTM has fewer hidden layer nodes, its predictive performance is still better than BP.
Table 1 single model prediction error on dataset 1
Table 2 single model prediction error on dataset 2
Step 233: and (5) analyzing the table calculation result. As can be seen from tables 1 and 2, all error indicators of LSTM neural networks are minimal on any data set. As can be seen from table 1, the GBDT error is greatest. Compared to GBDT, LSTM has 69.83% lower MSE, 46.78% lower MAE, 45.08% lower RMSE, and 53.85% lower MAPE. In other words, LSTM has the highest prediction accuracy among the commonly used single prediction models. As can be seen from table 2, the error of the support vector machine is the largest. Compared to SVM, LSTM has MSE reduced by 84.63%, MAE reduced by 61.09%, RMSE reduced by 60.79%, and MAPE reduced by 67.47%. That is, LSTM has the highest prediction accuracy compared to the reference prediction model.
Step 234: the predicted results of the six models on the two data sets are shown in fig. 4 and 5, respectively. As is clear from fig. 4 and 5, there is a significant one-step lag in the LSTM's prediction, so we denoise the data and modify the prediction according to WSR, thus improving LSTM. The prediction errors for the two data sets are shown in tables 3 and 4, respectively.
Table 3 comparison of corrected prediction errors on dataset 1
Table 4 comparison of corrected prediction errors on dataset 1
In tables 4 and 5, we underline the minimum error for WD-LSTM implementation for different number of decomposition layers (lev) (lev from 2 to 5 in both tables). For different numbers of IMFs (ranges 4 to 8 in Table 4 and ranges 3 to 7 in Table 5), we bolded the minimum error achieved by VMD-LSTM. From these two tables, it can be seen that the error of WD-LSTM is minimized when WD decomposition layer number is 2. Moreover, most VMD-LSTM errors are less than WD-LSTM minimum errors prior to WSR correction. This fully demonstrates that the denoising effect of VMD is much better than that of WD. On both datasets, the error of each model is significantly reduced after correction of the prediction by WSR. The reduction ratio is between 5% and 30%. The minimum error on data set 1 is obtained by VMD-LSTM-PSOR (imf=6) and the minimum error on data set 2 is obtained by VMD-LSTM-PSOR (imf=4).
Step 235: and (5) analyzing the corrected prediction result. Fig. 6 and 7 are predicted results for LSTM, VMD-LSTM, and VMD-LSTM-PSOR for data set 1 when imf=6 and data set 2 when imf=4, respectively. From both figures, it can be seen that there is a significant one-step lag in the solid line representing the LSTM prediction. After VMD pretreatment, the fluctuation of the wind speed sequence is weakened, and the wind speed curve is stable. And finally, correcting the predicted result by using the WSR with particle swarm optimization to enable the wind field to be relatively close to an actual wind speed curve, wherein the line representing the predicted result of the VMD-LSTM-PSOR is closest to the line representing the actual wind speed.
Step 24: interval prediction results and discussion
Step 241: and respectively using KDE and B spline interpolation to fit LDS. Based on the deterministic prediction result, according to different confidence intervals, the interval prediction of wind speed is obtained. The coverage, average diameter and uncovered points are shown in table 5.
TABLE 5 analysis of interval prediction results
Step 242: and (5) analyzing the interval prediction result. Table 5 calculates the interval predictions for different fitting methods for 90% confidence interval and 98% confidence interval, respectively. For data set 1, on the one hand, the average diameters of the KDE and B-spline interpolations were nearly identical (1.5009 m/s and 1.5075m/s, respectively) for the 90% confidence interval, but the coverage of the B-spline was 2% higher than the KDE. B-spline interval prediction covers points 2 and 74, but KDE does not. The situation is similar for a 98% confidence interval. The average diameter was almost the same, but the coverage of the B-spline was 3% higher than the KDE. The B-spline interval prediction includes three points, namely points 19, 81 and 95, but the KDE does not. The result shows that when the confidence intervals are the same and the average diameters are basically the same, the fitting effect of the KDE on the LDS is not as good as that of the B-spline interpolation, and the confidence intervals of the B-spline interpolation can cover more real wind speed points. On the other hand, for the same fitting method, the larger the confidence interval, the larger the average diameter, which is consistent with the statistical principle, but does not lead to higher coverage. Because a higher confidence interval means that there may be a higher upper limit and a higher lower limit for the interval prediction, coverage will also change (for a KDE fit LDS,90% of the confidence interval covers 19, 81 and 85 points and 98% of the confidence interval does not, conversely, 98% of the confidence interval covers 84 points and 90% of the confidence interval does not. This means that the fit to the LDS can achieve higher coverage and better interval prediction results without using high confidence intervals.
Step 243: we plotted the results of the interval prediction on two datasets as shown in fig. 8, 9, 10, 11, respectively. Fig. 8 and 10 plot the interval prediction results at 90% confidence interval, and fig. 9 and 11 plot the interval prediction results at 98% confidence interval. The result of the interval prediction covers most of the actual wind speed points, which indicates that the wind speed interval prediction based on the LDS is effective, the prediction result obtained by fitting the LDS by B-spline interpolation is superior to the LDS fitted by KDE, and the latter is a more effective method, which is consistent with that shown in Table 5.
FIG. 3 is a schematic diagram of a wind speed prediction system according to the present invention. As shown in fig. 3, the prediction system includes:
the wind speed data acquisition and denoising module 31 is used for acquiring an original wind speed sequence; and carrying out variation modal decomposition on the wind speed sequence to obtain a denoising sequence and a noise remainder.
The wind speed data acquisition and denoising module 31 specifically includes:
acquiring original wind speed data;
setting the number of decomposition; iteratively computing a set containing all modes and their center frequencies under the determined number of decompositions; acquiring an intrinsic mode function component and a noise remainder of a wind speed signal;
reconstructing the eigenvalue function component of the wind speed signal to obtain a denoised wind speed sequence.
And the wind speed data preliminary prediction module 32 is used for establishing a long and short time neural network prediction model, and carrying out preliminary prediction on the denoising sequence to obtain a preliminary prediction result.
The wind speed data preliminary prediction module 32 specifically includes:
dividing the denoised wind speed sequence into a training set and a testing set according to the proportion of 9:1;
setting a network structure of a long-short time neural network, wherein the network structure comprises the number of neurons of an input layer, the number of neurons of an implicit layer and the number of neurons of an output layer;
training by inputting a training set into a long-short time neural network;
and inputting the test set into a trained long-short-time neural network to obtain a preliminary prediction result of the wind speed.
The preliminary prediction result correction module 33 is configured to correct the preliminary prediction result by defining a wind speed climbing event and a wind speed climbing rate, and obtain a corrected wind speed prediction result;
the preliminary prediction result correction module 33 based on wind speed climbing specifically includes:
calculating a wind speed gradient; introducing a definition of wind speed climbing according to a wind speed gradient;
when the absolute value of the gradient is larger than the threshold value A, and the positive gradient increase exceeds the positive gradient threshold value B or the negative gradient decrease exceeds the negative gradient threshold value C at the current moment, correcting the wind speed predicted value at the next moment by using the error at the current moment;
for thresholds A, B and C, it is transformed into a multi-objective optimization problem, which is solved with PSO with the goal of minimizing root mean square error.
The LDS module 34 is acquired to describe the influence of the aerodynamic system on the wind speed by Lorenz equation, and obtain LDS.
The obtaining the LDS module 34 specifically includes:
giving initial conditions (0, 1), solving Lorenz equation, and obtaining three-dimensional LDS;
according to the Chebyshev distance, converting the three-dimensional LDS into a one-dimensional disturbance sequence.
The disturbance upper and lower limit obtaining module 35 is configured to fit the LDS by using a B-spline interpolation method, fix a confidence interval to the fitting result, and obtain an upper limit and a lower limit of a wind speed disturbance interval.
The obtaining perturbation upper and lower limit module 35 specifically includes:
fitting the distribution of the LDS through the B spline difference value to obtain a B spline difference value fitting function;
fixing confidence intervals to be 90% and 98% respectively, and calculating upper and lower quantile points of the fitting function;
the upper quantile is set as the upper limit of the interval prediction and the lower quantile is set as the lower limit of the interval prediction.
The prediction module 36 is configured to sum the corrected wind speed prediction result and the wind speed disturbance interval to obtain an interval prediction result of the wind speed.
The prediction module 36 specifically includes:
adding the upper and lower dividing points under the 90% confidence interval to the corrected wind speed prediction result to obtain a wind speed interval prediction result under the 90% confidence interval;
and adding the upper and lower dividing points under the 98% confidence interval to the corrected wind speed prediction result to obtain the wind speed interval prediction result under the 98% confidence interval.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.
Claims (10)
1. A short-term wind speed prediction method based on wind speed characteristics, the prediction method comprising: acquiring an original wind speed sequence; performing variation modal decomposition VMD on the wind speed sequence to obtain a denoising sequence and a noise remainder; establishing a long-short time neural network prediction model LSTM, and carrying out preliminary prediction on the denoising sequence to obtain a preliminary prediction result; correcting the preliminary prediction result by defining a wind speed climbing event WSR and a wind speed climbing rate to obtain a corrected wind speed prediction result; describing the influence of an atmospheric power system on the wind speed through a Lorenz equation, and obtaining a Lorenz disturbance sequence LDS; fitting the LDS through a B spline interpolation method, fixing a confidence interval to a fitting result, and obtaining an upper limit and a lower limit of a wind speed disturbance interval; summing the corrected wind speed prediction result and a wind speed disturbance interval to obtain an interval prediction result of the wind speed;
the method for acquiring the denoising sequence and the noise remainder specifically comprises the following steps: acquiring original wind speed data; setting the number of decomposition; iteratively computing a set containing all modes and their center frequencies under the determined number of decompositions; acquiring an intrinsic mode function component and a noise remainder of a wind speed signal; reconstructing the eigenvalue function component of the wind speed signal to obtain a denoised wind speed sequence;
the defining the wind speed climbing event and the wind speed climbing rate, correcting the preliminary prediction result, and obtaining a corrected wind speed prediction result, specifically comprising: calculating a wind speed gradient; introducing a definition of wind speed climbing according to a wind speed gradient; when the absolute value of the gradient is larger than a threshold value 1, and the positive gradient increase exceeds a positive gradient threshold value 2 or the negative gradient decrease exceeds a negative gradient threshold value 3 at the current moment, correcting the wind speed predicted value at the next moment by using the error at the current moment; for threshold 1, threshold 2 and threshold 3, they are transformed into a multi-objective optimization problem, which is solved with particle swarm algorithm PSO, with the goal of minimizing root mean square error.
2. The method for predicting short-term wind speed according to claim 1, wherein the establishing a long-short time neural network prediction model performs preliminary prediction on the denoising sequence to obtain a preliminary prediction result, and the specific process comprises: dividing the denoised wind speed sequence into a training set and a testing set according to the proportion of 9:1; setting a network structure of a long-short time neural network, wherein the network structure comprises the number of neurons of an input layer, the number of neurons of an implicit layer and the number of neurons of an output layer; training by inputting a training set into a long-short time neural network; and inputting the test set into a trained long-short-time neural network to obtain a preliminary prediction result of the wind speed.
3. The short-term wind speed prediction method according to claim 1, wherein the influence of the aerodynamic system on the wind speed is described by the Lorenz equation, and the LDS is obtained, in particular comprising: giving initial conditions (0, 1), solving Lorenz equation, and obtaining three-dimensional LDS; according to the Chebyshev distance, converting the three-dimensional LDS into a one-dimensional disturbance sequence.
4. The method according to claim 1, wherein the fitting the LDS by the B-spline interpolation method, and fixing a confidence interval to the fitting result, and obtaining an upper limit and a lower limit of a wind speed disturbance interval, specifically includes: fitting the distribution of the LDS through B spline interpolation to obtain a B spline interpolation fitting function; fixing confidence intervals to be 90% and 98% respectively, and calculating upper and lower quantile points of the fitting function;
the upper quantile is set as the upper limit of the interval prediction and the lower quantile is set as the lower limit of the interval prediction.
5. The method according to claim 1, wherein the step of summing the corrected wind speed prediction result and the wind speed disturbance zone to obtain a wind speed zone prediction result comprises: adding the upper and lower dividing points under the 90% confidence interval to the corrected wind speed prediction result to obtain a wind speed interval prediction result under the 90% confidence interval; and adding the upper and lower dividing points under the 98% confidence interval to the corrected wind speed prediction result to obtain the wind speed interval prediction result under the 98% confidence interval.
6. A short-term wind speed prediction system, the prediction system comprising: the wind speed data acquisition and denoising module is used for acquiring an original wind speed sequence; performing variation modal decomposition on the wind speed sequence to obtain a denoising sequence and a noise remainder; the wind speed data preliminary prediction module is used for establishing a long-short-time neural network prediction model, and carrying out preliminary prediction on the denoising sequence to obtain a preliminary prediction result; the preliminary prediction result correction module is used for correcting the preliminary prediction result by defining a wind speed climbing event and a wind speed climbing rate to obtain a corrected wind speed prediction result; the LDS module is used for describing the influence of the aerodynamic system on the wind speed through a Lorenz equation and obtaining the LDS; the disturbance upper and lower limit obtaining module is used for fitting the LDS through a B spline interpolation method, fixing a confidence interval to a fitting result and obtaining an upper limit and a lower limit of a wind speed disturbance interval; the prediction module is used for summing the corrected wind speed prediction result and the wind speed disturbance interval to obtain an interval prediction result of the wind speed;
the wind speed data acquisition and denoising module specifically comprises: the intrinsic mode component unit is used for acquiring original wind speed data; setting the number of decomposition; iteratively computing a set containing all modes and their center frequencies under the determined number of decompositions; acquiring an intrinsic mode function component and a noise remainder of a wind speed signal; the wind speed sequence denoising unit is used for reconstructing the eigenvalue function component of the wind speed signal to obtain a wind speed sequence denoising;
the preliminary prediction result correction module based on wind speed climbing specifically comprises: defining a wind speed climbing unit for calculating a wind speed gradient; introducing a definition of wind speed climbing according to a wind speed gradient; defining a correction method unit, which is used for correcting a wind speed predicted value at the next moment by using an error at the current moment when the absolute value of the gradient is larger than a threshold value 1 and the positive gradient increase at the current moment exceeds a positive gradient threshold value 2 or the negative gradient decrease exceeds a negative gradient threshold value 3; and the solving threshold unit is used for converting the thresholds 1,2 and 3 into a multi-objective optimization problem, aiming at minimizing root mean square error, and solving the optimization problem by using PSO.
7. The short-term wind speed prediction system according to claim 6, wherein the wind speed data preliminary prediction module specifically comprises: dividing and dividing a wind speed sequence unit, wherein the wind speed sequence unit is used for dividing the denoised wind speed sequence into a training set and a testing set according to the proportion of 9:1; setting a network structure unit of a long-short time neural network, wherein the network structure unit is used for comprising the number of neurons of an input layer, the number of neurons of an hidden layer and the number of neurons of an output layer; the training model unit is used for training by inputting a training set into the long-short-time neural network; and the prediction unit is used for inputting the test set into the trained long-short-time neural network to obtain a preliminary wind speed prediction result.
8. The short-term wind speed prediction system according to claim 6, wherein the obtaining LDS module specifically comprises: the method comprises the steps of obtaining a three-dimensional LDS unit, wherein the three-dimensional LDS unit is used for giving initial conditions (0, 1), solving a Lorenz equation and obtaining a three-dimensional LDS; and acquiring a one-dimensional LDS unit, wherein the one-dimensional LDS unit is used for converting the three-dimensional LDS into the one-dimensional LDS according to the Chebyshev distance.
9. The short-term wind speed prediction system according to claim 6, wherein the obtaining perturbation upper and lower limits module specifically comprises: the fitting unit is used for fitting the distribution of the LDS through B spline interpolation to obtain a B spline interpolation fitting function; the quantile calculating unit is used for respectively fixing 90% and 98% confidence intervals and calculating upper and lower quantile points of the fitting function; and determining an upper and lower interval prediction limit unit, wherein the upper partition point is set as an upper limit of interval prediction, and the lower partition point is set as a lower limit of interval prediction.
10. The short-term wind speed prediction system according to claim 6, wherein the prediction module specifically comprises: a 90% confidence interval lower interval prediction result unit, configured to obtain a wind speed interval prediction result under the 90% confidence interval by adding the upper and lower quantiles under the 90% confidence interval to the corrected wind speed prediction result; and the interval prediction result unit under the 98% confidence interval is used for obtaining the wind speed interval prediction result under the 98% confidence interval by adding the upper and lower dividing points under the 98% confidence interval to the corrected wind speed prediction result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910805484.XA CN110543929B (en) | 2019-08-29 | 2019-08-29 | Wind speed interval prediction method and system based on Lorenz system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910805484.XA CN110543929B (en) | 2019-08-29 | 2019-08-29 | Wind speed interval prediction method and system based on Lorenz system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110543929A CN110543929A (en) | 2019-12-06 |
| CN110543929B true CN110543929B (en) | 2023-11-14 |
Family
ID=68712231
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910805484.XA Active CN110543929B (en) | 2019-08-29 | 2019-08-29 | Wind speed interval prediction method and system based on Lorenz system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110543929B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111950759B (en) * | 2020-05-18 | 2024-06-21 | 武汉科技大学 | Short-term wind speed prediction method based on two-stage decomposition, LSTM and AT |
| CN112183866A (en) * | 2020-09-29 | 2021-01-05 | 中南大学 | Short-term wind speed exceeding probability prediction method and system |
| CN112132360A (en) * | 2020-09-29 | 2020-12-25 | 中交建冀交高速公路投资发展有限公司 | Wind speed short-time confidence interval prediction method and system based on LSTM and GMM |
| CN112765544B (en) * | 2020-12-30 | 2023-09-15 | 国网湖南省电力有限公司 | Method for correcting and forecasting strong wind power under influence of typhoon of power transmission line |
| CN112785052B (en) * | 2021-01-14 | 2022-04-19 | 吉林大学 | Wind speed and wind direction prediction method based on particle filter algorithm |
| CN113516320B (en) * | 2021-09-14 | 2021-12-10 | 国能日新科技股份有限公司 | Wind speed correction and predicted wind speed optimization method and device based on multi-objective genetic algorithm |
| CN114814963B (en) * | 2022-03-03 | 2023-03-21 | 吉林大学 | B-spline channel extraction-based time domain electromagnetic data very low frequency noise suppression method |
| CN115345387B (en) * | 2022-10-18 | 2023-01-24 | 哈尔滨工业大学(深圳)(哈尔滨工业大学深圳科技创新研究院) | Wind field wind speed prediction method and device and storage medium |
| CN115688588B (en) * | 2022-11-04 | 2023-06-27 | 自然资源部第一海洋研究所 | Sea surface temperature daily variation amplitude prediction method based on improved XGB method |
| CN117217421B (en) * | 2023-11-07 | 2024-09-10 | 东北大学 | Visual analysis method and system for auxiliary decision making of thick plate production plan |
| CN117725368B (en) * | 2024-02-07 | 2024-05-31 | 浙江公路水运工程咨询集团有限公司 | Highway slope displacement prediction method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101699518A (en) * | 2009-10-30 | 2010-04-28 | 华南理工大学 | Method for beautifying handwritten Chinese character based on trajectory analysis |
| US7831419B2 (en) * | 2005-01-24 | 2010-11-09 | Smith International, Inc. | PDC drill bit with cutter design optimized with dynamic centerline analysis having an angular separation in imbalance forces of 180 degrees for maximum time |
| CN104794546A (en) * | 2015-04-29 | 2015-07-22 | 武汉大学 | Wind power climbing forecasting method based on deep confidence network classifying method |
| CN108154260A (en) * | 2017-12-15 | 2018-06-12 | 南京信息工程大学 | A kind of short-term wind power forecast method |
| CN108520269A (en) * | 2018-03-10 | 2018-09-11 | 华北电力大学（保定） | A kind of wind speed forecasting method and forecasting wind speed system |
| CN108846508A (en) * | 2018-05-30 | 2018-11-20 | 华北电力大学（保定） | A kind of wind speed forecasting method and system based on atmospheric perturbation |
-
2019
- 2019-08-29 CN CN201910805484.XA patent/CN110543929B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7831419B2 (en) * | 2005-01-24 | 2010-11-09 | Smith International, Inc. | PDC drill bit with cutter design optimized with dynamic centerline analysis having an angular separation in imbalance forces of 180 degrees for maximum time |
| CN101699518A (en) * | 2009-10-30 | 2010-04-28 | 华南理工大学 | Method for beautifying handwritten Chinese character based on trajectory analysis |
| CN104794546A (en) * | 2015-04-29 | 2015-07-22 | 武汉大学 | Wind power climbing forecasting method based on deep confidence network classifying method |
| CN108154260A (en) * | 2017-12-15 | 2018-06-12 | 南京信息工程大学 | A kind of short-term wind power forecast method |
| CN108520269A (en) * | 2018-03-10 | 2018-09-11 | 华北电力大学（保定） | A kind of wind speed forecasting method and forecasting wind speed system |
| CN108846508A (en) * | 2018-05-30 | 2018-11-20 | 华北电力大学（保定） | A kind of wind speed forecasting method and system based on atmospheric perturbation |
Non-Patent Citations (1)
| Title |
|---|
| 基于Lorenz扰动分布和VMD的神经网络风速预测研究;张亚刚 等;《华北电力大学学报》;20190731;第8-15页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110543929A (en) | 2019-12-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110543929B (en) | Wind speed interval prediction method and system based on Lorenz system | |
| CN110309603B (en) | Short-term wind speed prediction method and system based on wind speed characteristics | |
| Chen et al. | Short-term wind speed forecasting with principle-subordinate predictor based on Conv-LSTM and improved BPNN | |
| CN111860982A (en) | Wind power plant short-term wind power prediction method based on VMD-FCM-GRU | |
| CN110163429B (en) | Short-term load prediction method based on similarity day optimization screening | |
| CN109871977A (en) | Load forecasting method based on wavelet transform and the minimum support vector machines of optimization | |
| CN110942194A (en) | Wind power prediction error interval evaluation method based on TCN | |
| CN116596044B (en) | Power generation load prediction model training method and device based on multi-source data | |
| CN109255728B (en) | Photovoltaic power generation power neural network prediction method based on chaotic phase space optimization reconstruction | |
| CN113553755B (en) | Power system state estimation method, device and equipment | |
| CN111382906A (en) | Power load prediction method, system, equipment and computer readable storage medium | |
| Li et al. | Short-term prediction of the power of a new wind turbine based on IAO-LSTM | |
| CN112819238A (en) | Short-term wind power prediction method based on chaotic chicken flock optimization algorithm | |
| CN111242355A (en) | Photovoltaic probability prediction method and system based on Bayesian neural network | |
| CN115759465A (en) | Wind power prediction method based on multi-target collaborative training and NWP implicit correction | |
| CN115310669A (en) | Ultra-short-term wind power prediction method based on secondary decomposition and IWOA-LSSVM | |
| CN115239029A (en) | Wind power prediction method and system considering power time sequence and meteorological dependent characteristics | |
| CN114781692A (en) | Short-term power load prediction method and device and electronic equipment | |
| CN117494888A (en) | Distributed photovoltaic power prediction method and system based on optimized cascade | |
| CN115995810A (en) | Wind power prediction method and system considering weather fluctuation self-adaptive matching | |
| CN115907131A (en) | Method and system for building electric heating load prediction model in northern area | |
| Sun et al. | Short-term power load prediction based on VMD-SG-LSTM | |
| CN111950752A (en) | Photovoltaic power station generating capacity prediction method, device and system and storage medium thereof | |
| CN113449920A (en) | Wind power prediction method, system and computer readable medium | |
| CN116706907B (en) | Photovoltaic power generation prediction method based on fuzzy reasoning and related equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |