CN113537595A - Model training method, wind power prediction method, system, device and medium - Google Patents

Model training method, wind power prediction method, system, device and medium Download PDF

Info

Publication number
CN113537595A
CN113537595A CN202110804988.7A CN202110804988A CN113537595A CN 113537595 A CN113537595 A CN 113537595A CN 202110804988 A CN202110804988 A CN 202110804988A CN 113537595 A CN113537595 A CN 113537595A
Authority
CN
China
Prior art keywords
wind power
moment
historical
data
predicted
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.)
Pending
Application number
CN202110804988.7A
Other languages
Chinese (zh)
Inventor
汤子琪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Electric Wind Power Group Co Ltd
Original Assignee
Shanghai Electric Wind Power Group Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Electric Wind Power Group Co Ltd filed Critical Shanghai Electric Wind Power Group Co Ltd
Priority to CN202110804988.7A priority Critical patent/CN113537595A/en
Publication of CN113537595A publication Critical patent/CN113537595A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply

Abstract

The invention discloses a model training method, a wind power prediction method, a system, equipment and a medium, wherein the training method comprises the following steps: acquiring historical training data of the wind power plant, wherein the data comprises observation data of each moment in a first historical period and prediction data of each moment in a second historical period as input and observation data of each moment in the second historical period as output; and training a neural network model based on historical training data to obtain a wind power prediction model at the moment to be predicted. The wind power prediction model of the time to be predicted is obtained based on the observation data of each moment in the first historical period, the prediction data of each moment in the second historical period and the observation data of each moment in the second historical period in the historical training data, the latest observation data is input into the wind power prediction model to predict the wind power of any future moment, the prediction error of the ultra-short-period wind power is reduced, and the accuracy of the wind power prediction model is improved.

Description

Model training method, wind power prediction method, system, device and medium
Technical Field
The invention relates to the technical field of wind power generation, in particular to a model training method, a wind power prediction method, a system, equipment and a medium.
Background
With the rapid increase of the installed capacity of the new energy, the impact of the intermittency, randomness and fluctuation of the new energy power generation on the power grid becomes more obvious, and higher requirements are provided for the stable operation of a grid-connected electric field.
At present, the ultra-short-term wind power prediction method mainly comprises a physical model method, a statistical method, a combined model method and the like. However, all three methods have certain defects, for example, a physical model method, and the accuracy of modeling affects the accuracy of prediction; statistical methods, require long-term measurement data and additional training. In addition, because the wind speed prediction data provided by the numerical weather forecast is strongly uncertain due to the influences of many factors such as climate and geographic conditions, and the uncertainty of the neural network model and the input quantity, the prediction error of the neural network model used by the wind power prediction system is large.
Disclosure of Invention
The invention aims to overcome the defects of large prediction error and low accuracy rate of an ultra-short-term wind power prediction method adopted in the prior art, and provides a model training method, a wind power prediction method, a system, equipment and a medium.
The invention solves the technical problems through the following technical scheme:
the invention provides a model training method in a first aspect, which comprises the following steps:
acquiring historical training data of a wind power plant, wherein the historical training data comprises observation data of each moment in a first historical period and prediction data of each moment in a second historical period as input, and observation data of each moment in the second historical period as output; the duration of the first historical period is the same as the duration of the second historical period, the earliest moment in the first historical period is earlier than the earliest moment in the second historical period, and the observed data and the predicted data are both related to wind power calculation of the wind farm;
and training a neural network model based on the historical training data to obtain a wind power prediction model at the moment to be predicted.
Preferably, the step of training the neural network model based on the historical training data to obtain a wind power prediction model at the time to be predicted includes:
aiming at each time t + i to be predicted, training a neural network model based on the historical training data corresponding to the time t + i to correspondingly train a wind power prediction model ANNi。
Preferably, the step of obtaining historical training data of the wind farm includes: acquiring observation data of m historical moments and prediction data of m historical moments, wherein the range of the m historical moments comprises the first historical time period and the second historical time period, and the m historical moments are sequentially marked as t-m, t-m +1, … and t-1 according to the time sequence, wherein t represents the current moment;
aiming at each time t + i to be predicted, training a neural network model based on the historical training data corresponding to the time t + i to correspondingly train a wind power prediction model ANNiComprises the following steps:
training each wind power prediction model ANNiThe historical training data includes [ A ]k，Bk+i，Ck+i]Wherein A iskObservation data representing a first history period, Bk+iPrediction data representing a second history period, Ck+iObservation data representing a second historical period, wherein the moments of the first historical period are the t-m th moment, the t-m +1 th moment, … and the t-1-i th moment in sequence; the moments of the second historical period are the t-m + i th moment, the t-m +1+ i th moment, … and the t-1 th moment in sequence.
Preferably, the observation data of the first historical period and the second historical period are both wind speed and wind direction observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value, or the observation data of the first historical period and the second historical period are both wind power observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value.
The invention provides a model training system in a second aspect, which comprises a first obtaining module and a training module;
the first acquisition module is used for acquiring historical training data of the wind power plant, wherein the historical training data comprises observation data of each moment in a first historical period and prediction data of each moment in a second historical period as input, and observation data of each moment in the second historical period as output; the duration of the first historical period is the same as the duration of the second historical period, the earliest moment in the first historical period is earlier than the earliest moment in the second historical period, and the observed data and the predicted data are both related to wind power calculation of the wind farm;
the training module is used for training a neural network model based on the historical training data to obtain a wind power prediction model at the moment to be predicted.
Preferably, the training module is specifically configured to, for each time t + i to be predicted, train the neural network model based on the historical training data corresponding to the time t + i, so as to correspondingly train a wind power prediction model ANNi。
Preferably, the first obtaining module is specifically configured to obtain observed data of m historical moments and predicted data of m historical moments, where a range of the m historical moments includes the first historical period and the second historical period, and the m historical moments are sequentially marked as t-m, t-m +1, …, t-1 according to a time sequence, where t represents a current moment;
the training module is specifically used for training each wind power prediction model ANNiThe historical training data includes [ A ]k，Bk+i，Ck+i]Wherein A iskObservation data representing a first history period, Bk+iPrediction data representing a second history period, Ck+iObservation data representing a second historical period, wherein the moments of the first historical period are the t-m th moment, the t-m +1 th moment, … and the t-1-i th moment in sequence; the moments of the second historical period are the t-m + i th moment, the t-m +1+ i th moment, … and the t-1 th moment in sequence.
Preferably, the observation data of the first historical period and the second historical period are both wind speed and wind direction observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value, or the observation data of the first historical period and the second historical period are both wind power observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value.
The third aspect of the present invention provides a wind power prediction method, including:
determining a time point t + i to be predicted of wind power of a wind power plant, wherein t represents the current time;
obtaining tjThe observed data at the moment and the predicted data at the t + i moment are used as the input of a wind power prediction model, wherein t isjThe moment is the historical moment nearest to the current moment t;
predicting the wind power of the time point t + i to be predicted by using a wind power prediction model;
wherein the wind power prediction model is trained by using the training method according to the first aspect.
Preferably, if the time point t + i to be predicted is targeted, the neural network model is trained based on the historical training data corresponding to the time point t + i, so as to correspondingly train a wind power prediction model ANNi(ii) a The wind power prediction method further comprises:
obtaining a wind power prediction model ANN corresponding to a to-be-predicted time point t + i of wind power of the wind power planti；
Using the wind power prediction model ANNiAnd predicting the wind power of the time point t + i to be predicted.
Preferably, if the observed data of the first historical period and the second historical period are both wind speed and wind direction observed values, and the predicted data of the second historical period is a wind speed and wind direction predicted value, the wind power prediction model ANN is usediThe step of predicting the wind power of the time point t + i to be predicted specifically includes:
will tjInputting the wind power forecast according to the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i momentModel ANNi；
And obtaining the wind power predicted value at the t + i moment based on the corresponding relation between the wind speed and the wind direction and the wind power of the wind power plant.
Preferably, if the observed data of the first historical period and the second historical period are both wind power observed values, and the predicted data of the second historical period is a wind speed and wind direction predicted value, the wind power prediction model ANN is usediThe step of predicting the wind power of the time point t + i to be predicted specifically includes:
will tjInputting the observed wind power value at the moment and the predicted wind speed and direction value at the t + i moment into the wind power prediction model ANNiAnd obtaining the predicted value of the wind power at the t + i moment.
The invention provides a wind power prediction system in a fourth aspect, which comprises a determining module, a second obtaining module and a prediction module;
the determining module is used for determining a to-be-predicted time point t + i of wind power of the wind power plant, wherein t represents the current time;
the second acquisition module is used for acquiring tjThe observed data at time and the predicted data at time t + i are used as input of a prediction model, wherein t isjThe moment is the historical moment nearest to the current moment t;
the prediction module is used for predicting the wind power of the time point t + i to be predicted by using a wind power prediction model;
wherein the wind power prediction model is trained by using the training system according to the second aspect.
Preferably, if the time point t + i to be predicted is targeted, the neural network model is trained based on the historical training data corresponding to the time point t + i, so as to correspondingly train a wind power prediction model ANNi(ii) a The wind power prediction system further comprises a third obtaining module;
the third obtaining module is used for obtaining a wind power prediction model ANN corresponding to a to-be-predicted time point t + i of the wind power planti；
The prediction module is specifically configured to utilize the wind power prediction model ANNiAnd predicting the wind power of the time point t + i to be predicted.
Preferably, if the observation data of the first historical period and the second historical period are both wind speed and wind direction observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value, the prediction module includes an obtaining unit and a prediction unit;
the acquisition unit is used for converting tjInputting the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i moment into the wind power prediction model ANNi；
And the prediction unit is used for obtaining a wind power prediction value at the t + i moment based on the corresponding relation between the wind speed and the wind direction and the wind power of the wind power plant.
Preferably, if the observed data of the first historical period and the second historical period are both wind power observed values, and the predicted data of the second historical period is a wind speed and direction predicted value, the prediction module is specifically configured to predict tjInputting the observed wind power value at the moment and the predicted wind speed and direction value at the t + i moment into the wind power prediction model ANNiAnd obtaining the predicted value of the wind power at the t + i moment.
A fifth aspect of the invention provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the model training method according to the first aspect or the wind power prediction method according to the third aspect when executing the computer program.
A sixth aspect of the invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a model training method as described in the first aspect or performs a wind power prediction method as described in the third aspect.
On the basis of the common knowledge in the field, the preferred conditions can be combined randomly to obtain the preferred embodiments of the invention.
The positive progress effects of the invention are as follows:
the method takes the observation data of each moment in the first historical period and the prediction data of each moment in the second historical period in the obtained historical training data of the wind power plant as input, takes the observation data of each moment in the second historical period as output, trains the neural network model to obtain the wind power prediction model of the moment to be predicted, realizes that the wind power of any future moment is predicted by inputting the latest observation data into the wind power prediction model, reduces the prediction error of the ultra-short-term wind power, and trains the corresponding wind power prediction models for each future moment, thereby reducing the prediction error generated by the uncertainty of the wind power prediction model and the input quantity and improving the accuracy of the wind power prediction model.
Drawings
Fig. 1 is a flowchart of a model training method according to embodiment 1 of the present invention.
Fig. 2 is a block diagram of a model training system according to embodiment 2 of the present invention.
Fig. 3 is a schematic structural diagram of an electronic device according to embodiment 3 of the present invention.
Fig. 4 is a flowchart of a wind power prediction method according to embodiment 5 of the present invention.
Fig. 5 is a schematic block diagram of a wind power prediction system according to embodiment 6 of the present invention.
Detailed Description
The invention is further illustrated by the following examples, which are not intended to limit the scope of the invention.
Example 1
As shown in fig. 1, the present embodiment provides a model training method, including:
101, acquiring historical training data of a wind power plant, wherein the historical training data comprises observation data of each moment in a first historical period and prediction data of each moment in a second historical period as input and observation data of a second historical moment as output; the duration of the first historical period is the same as the duration of the second historical period, the earliest moment in the first historical period is earlier than the earliest moment in the second historical period, and the observation data and the prediction data are both related to wind power calculation of the wind power plant.
In the embodiment, the observation data at each time in the first history period and the prediction data at each time in the second history period are used as input, so that the observation data (i.e. the real information) at each time in the first history period before the fusion is used for correcting the prediction value at each time in the second history period.
And 102, training the neural network model based on historical training data to obtain a wind power prediction model at the moment to be predicted.
In this embodiment, the SCADA (monitoring and data acquisition) data and the wind power tower meteorological data of the wind farm are accessed to the wind power prediction system in real time, and historical training data of the wind farm is obtained from the SCADA data and the wind power tower meteorological data of the wind farm and used for subsequent model training and wind power prediction.
The SCADA data is fan operation data, and includes related data such as the power generation amount and the operation state of the fan; the meteorological data of the wind power tower refer to variables such as wind speed, wind direction, air temperature and humidity.
In one implementation, step 102 includes: aiming at each time point t + i to be predicted, training a neural network model based on historical training data corresponding to the time t + i to correspondingly train a wind power prediction model ANNi。
In one embodiment, step 101 comprises: the method comprises the steps of obtaining observation data of m historical moments and prediction data of the m historical moments, wherein the range of the m historical moments comprises a first historical time period and a second historical time period, the m historical moments are sequentially marked as t-m, t-m +1, … and t-1 according to the time sequence, and t represents the current moment.
In this embodiment, the m historical times t-m, t-m +1, …, t-1 do not represent specific time, but represent sequence points of the data sequence, each sequence point corresponds to a specific historical time, and the earlier the sequence point is, the longer the represented historical time is from the current time; for example, m is 6, t-6 denotes 1/11: 00 in 2021, t-5 denotes 1/12: 00 in 2021, t-4 denotes 1/13: 00 in 2021, t-3 denotes 1/14: 00 in 2021, t-2 denotes 1/15: 00 in 2021, t-1 denotes 1/16: 00 in 2021, t denotes 1/17: 00 in 2021.
The time intervals between two sequence points may be different or the same, preferably the same.
In this embodiment, each wind power prediction model ANN is trainediWhen, the historical training data includes [ A ]k，Bk+i，Ck+i]Wherein A iskObservation data representing a first history period, Bk+iPrediction data representing a second history period, Ck+iThe observation data representing the second historical period, and the moments of the first historical period are the t-m moment, the t-m +1 moment, … and the t-1-i moment in sequence; the moments of the second history period are the t-m + i th moment, the t-m +1+ i th moment, … and the t-1 th moment in sequence. Specifically, observed data (i.e., A) at time t-m, time t-m +1, …, time t-1-i of the first history period is obtainedt-m，At-m+1，…，At-1-i) As training input 1; obtaining the predicted data (i.e. B) of the t-m + i th time, the t-m +1+ i th time, …, the t-1 th time of the second history periodt-m+i，Bt-m+1+i，…，Bt-1) As training input 2; the observation data (namely C) at the t-m + i th time, the t-m +1+ i th time, … and the t-1 th time of the second historical period are obtainedt-m+i，Ct-m+1+i，…，Ct-1) As training output, training the neural network model to obtain a wind power prediction model ANNi。
It should be noted that, in training each wind power prediction model ANNiThe number of historical training data is related to i, i.e. the number of historical training data is m-1-i, e.g. wind power prediction model ANN1The historical training data of (1) is m-2, and the wind power prediction model ANN10The historical training data of (a) is m-11.
In this embodiment, for the time t + i to be predicted, the larger i is, the smaller the number of the historical training data of the corresponding wind power prediction model is, that is, the closer the time to be predicted is to the current time, the larger the number of the historical training data of the corresponding wind power prediction model is; and the farther the time to be predicted is from the current time, the smaller the amount of historical training data of the corresponding wind power prediction model is.
In this embodiment, the observation data of the first historical period and the second historical period are both wind speed and wind direction observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value, or the observation data of the first historical period and the second historical period are both wind power observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value.
Training each wind power prediction model ANNiIf the observation data of the first historical period and the second historical period are both the wind speed and wind direction observation values and the prediction data of the second historical period is the wind speed and wind direction prediction value, acquiring the wind speed and wind direction observation values of the t-m moment, the t-m +1 moment, … and the t-1-i moment of the first historical period as training input 1; acquiring a wind speed and direction predicted value at the t-m + i moment, the t-m +1+ i moment, … and the t-1 moment of a second historical period as a training input 2; acquiring the observed values of the wind speed and the wind direction at the t-m + i moment, the t-m +1+ i moment, … moment and the t-1 moment of the second historical period as training output, and training the neural network model to obtain a wind power prediction model ANNi。
If the observation data of the first historical period and the second historical period are both wind power observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value, acquiring the wind power observation values of the t-m moment, the t-m +1 moment, … and the t-1-i moment of the first historical period as training input 1; acquiring wind speed and direction predicted values at t-m + i moment, t-m +1+ i moment, … moment and t-1 moment of the second historical moment as training input 2; obtaining wind power observation values of the t-m + i moment, the t-m +1+ i moment, … and the t-1 moment of the second historical period as training output, and training the neural network model to obtain a wind power prediction model ANNi。
In the embodiment, the wind speed and direction observation values of a first historical time period and a second historical time period are obtained from a wind speed and direction instrument arranged on a wind tower or a fan inside the wind power plant; acquiring a wind speed and direction predicted value of a second historical time period from a weather forecast database; and acquiring wind power observed values of the first historical moment and the second historical period from the SCADA data.
The embodiment inputs the latest observed data (including data of SCADA and wind power tower) from the wind power plant into the neural network model, and the input of the latest observed data into the neural network model can reduce the prediction error 2-3 hours before prediction. The neural network model requires setting a wind power prediction model for each predicted time step, extracting each time step into the wind power prediction model that refers to it, using the observation data and prediction data time series as inputs, and obtaining a corrected time series by merging the corrected time steps. By setting a wind power prediction model for each time step, the wind power prediction model can automatically mine the relation between the current moment and the future moment, thereby reducing the prediction error caused by uncertainty of the wind power prediction model and the input quantity and also reducing the prediction error of the ultra-short-term wind power.
According to the method, the observation data of each moment in the first historical period and the prediction data of each moment in the second historical period in the obtained historical training data of the wind power plant are used as input, the observation data of each moment in the second historical period are used as output, the neural network model is trained to obtain the wind power prediction model of the moment to be predicted, and the training of the wind power prediction model corresponding to each moment in the future is realized, so that the prediction error caused by uncertainty of the wind power prediction model and the input quantity is reduced, and the accuracy of the wind power prediction model is improved.
Example 2
As shown in fig. 2, the present embodiment provides a model training system, which includes a first obtaining module 21 and a training module 22.
The first obtaining module 21 is configured to obtain historical training data of the wind farm, where the historical training data includes, as inputs, observed data at each time in a first historical period, predicted data at each time in a second historical period, and, as outputs, observed data at each time in the second historical period; the duration of the first historical period is the same as the duration of the second historical period, the earliest moment in the first historical period is earlier than the earliest moment in the second historical period, and the observation data and the prediction data are both related to wind power calculation of the wind power plant.
In the embodiment, the observation data at each time in the first history period and the prediction data at each time in the second history period are used as input, so that the observation data (i.e. the real information) at each time in the first history period before the fusion is used for correcting the prediction value at each time in the second history period.
The training module 22 is configured to train the neural network model based on the historical training data to obtain a wind power prediction model at a time to be predicted.
In the embodiment, the SCADA data and the wind power tower meteorological data of the wind power plant are accessed into the wind power prediction system in real time, and historical training data of the wind power plant are obtained from the SCADA data and the wind power tower meteorological data of the wind power plant and are used for subsequent model training and wind power prediction.
The SCADA data is fan operation data, and includes related data such as the power generation amount and the operation state of the fan; the meteorological data of the wind power tower refer to variables such as wind speed, wind direction, air temperature and humidity.
In an implementation scheme, the training module 22 is specifically configured to train, for each time point t + i to be predicted, a neural network model based on historical training data corresponding to the time point t + i, so as to correspondingly train a wind power prediction model ANNi。
In an implementable scheme, the first obtaining module 21 is specifically configured to obtain observed data of m historical times and predicted data of m historical times, a range of the m historical times includes a first historical time period and a second historical time period, the m historical times are sequentially marked as t-m, t-m +1, …, t-1 according to a time sequence, where t represents a current time.
In this embodiment, the m historical times t-m, t-m +1, …, t-1 do not represent specific time, but represent sequence points of the data sequence, each sequence point corresponds to a specific historical time, and the earlier the sequence point is, the longer the represented historical time is from the current time; for example, m is 6, t-6 denotes 1/11: 00 in 2021, t-5 denotes 1/12: 00 in 2021, t-4 denotes 1/13: 00 in 2021, t-3 denotes 1/14: 00 in 2021, t-2 denotes 1/15: 00 in 2021, t-1 denotes 1/16: 00 in 2021, t denotes 1/17: 00 in 2021.
The time intervals between two sequence points may be different or the same, preferably the same.
In this embodiment, the training module 22 is specifically configured to train each wind power prediction model ANNiWhen, the historical training data includes [ A ]k，Bk+i，Ck+i]Wherein A iskObservation data representing a first history period, Bk+iPrediction data representing a second history period, Ck+iThe observation data representing the second historical period, and the moments of the first historical period are the t-m moment, the t-m +1 moment, … and the t-1-i moment in sequence; the moments of the second history period are the t-m + i th moment, the t-m +1+ i th moment, … and the t-1 th moment in sequence. Specifically, observed data (i.e., A) at time t-m, time t-m +1, …, time t-1-i of the first history period is obtainedt-m，At-m+1，…，At-1-i) As training input 1; obtaining the predicted data (i.e. B) of the t-m + i th time, the t-m +1+ i th time, …, the t-1 th time of the second history periodt-m+i，Bt-m+1+i，…，Bt-1) As training input 2; the observation data (namely C) at the t-m + i th time, the t-m +1+ i th time, … and the t-1 th time of the second historical period are obtainedt-m+i，Ct-m+1+i，…，Ct-1) As training output, training the neural network model to obtain a wind power prediction model ANNi。
It should be noted that, in training each wind power prediction model ANNiThe number of the historical training data is related to i, namely the number of the historical training data is m-1-i,for example, the wind power prediction model ANN1The historical training data of (1) is m-2, and the wind power prediction model ANN10The historical training data of (a) is m-11.
In this embodiment, for the time t + i to be predicted, the larger i is, the smaller the number of the historical training data of the corresponding wind power prediction model is, that is, the closer the time to be predicted is to the current time, the larger the number of the historical training data of the corresponding wind power prediction model is; and the farther the time to be predicted is from the current time, the smaller the amount of historical training data of the corresponding wind power prediction model is.
In this embodiment, the observation data of the first historical period and the second historical period are both wind speed and wind direction observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value, or the observation data of the first historical period and the second historical period are both wind power observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value.
The training module 22 is specifically configured to train each wind power prediction model ANNiIf the observation data of the first historical period and the second historical period are both the wind speed and wind direction observation values and the prediction data of the second historical period is the wind speed and wind direction prediction value, acquiring the wind speed and wind direction observation values of the t-m moment, the t-m +1 moment, … and the t-1-i moment of the first historical period as training input 1; acquiring a wind speed and direction predicted value at the t-m + i moment, the t-m +1+ i moment, … and the t-1 moment of a second historical period as a training input 2; acquiring the observed values of the wind speed and the wind direction at the t-m + i moment, the t-m +1+ i moment, … moment and the t-1 moment of the second historical period as training output, and training the neural network model to obtain a wind power prediction model ANNi。
If the observation data of the first historical period and the second historical period are both wind power observation values, and the prediction data of the second historical period is a wind speed and wind direction prediction value, acquiring the wind power observation values of the t-m moment, the t-m +1 moment, … and the t-1-i moment of the first historical period as training input 1; acquiring predicted values of wind speed and wind direction at t-m + i, t-m +1+ i, … and t-1 of the second historical moment as predicted valuesTraining input 2; obtaining wind power observation values of the t-m + i moment, the t-m +1+ i moment, … and the t-1 moment of the second historical period as training output, and training the neural network model to obtain a wind power prediction model ANNi。
In the embodiment, the wind speed and direction observation values of a first historical time period and a second historical time period are obtained from a wind speed and direction instrument arranged on a wind tower or a fan inside the wind power plant; acquiring a wind speed and direction predicted value of a second historical time period from a weather forecast database; wind power observations are obtained from the SCADA data for a first historical period and a second historical period.
The embodiment inputs the latest observed data (including data of SCADA and wind power tower) from the wind power plant into the neural network model, and the input of the latest observed data into the neural network model can reduce the prediction error 2-3 hours before prediction. The neural network model requires setting a wind power prediction model for each predicted time step, extracting each time step into the wind power prediction model that refers to it, using the observation data and prediction data time series as inputs, and obtaining a corrected time series by merging the corrected time steps. By setting a wind power prediction model for each time step, the wind power prediction model can automatically mine the relation between the current moment and the future moment, thereby reducing the prediction error caused by uncertainty of the wind power prediction model and the input quantity and also reducing the prediction error of the ultra-short-term wind power.
According to the method, the observation data of each moment in the first historical period and the prediction data of each moment in the second historical period in the obtained historical training data of the wind power plant are used as input, the observation data of each moment in the second historical period are used as output, the neural network model is trained to obtain the wind power prediction model of the moment to be predicted, and the training of the wind power prediction model corresponding to each moment in the future is realized, so that the prediction error caused by uncertainty of the wind power prediction model and the input quantity is reduced, and the accuracy of the wind power prediction model is improved.
Example 3
Fig. 3 is a schematic structural diagram of an electronic device according to embodiment 3 of the present invention. The electronic device comprises a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the model training method of embodiment 1 when executing the program. The electronic device 30 shown in fig. 3 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiment of the present invention.
As shown in fig. 3, the electronic device 30 may be embodied in the form of a general purpose computing device, which may be, for example, a server device. The components of the electronic device 30 may include, but are not limited to: the at least one processor 31, the at least one memory 32, and a bus 33 connecting the various system components (including the memory 32 and the processor 31).
The bus 33 includes a data bus, an address bus, and a control bus.
The memory 32 may include volatile memory, such as Random Access Memory (RAM)321 and/or cache memory 322, and may further include Read Only Memory (ROM) 323.
Memory 32 may also include a program/utility 325 having a set (at least one) of program modules 324, such program modules 324 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.
The processor 31 executes various functional applications and data processing, such as the model training method of embodiment 1 of the present invention, by executing the computer program stored in the memory 32.
The electronic device 30 may also communicate with one or more external devices 34 (e.g., keyboard, pointing device, etc.). Such communication may be through input/output (I/O) interfaces 35. Also, model-generating device 30 may also communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the Internet) via network adapter 36. As shown in FIG. 3, network adapter 36 communicates with the other modules of model-generating device 30 via bus 33. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the model-generating device 30, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems, etc.
It should be noted that although in the above detailed description several units/modules or sub-units/modules of the electronic device are mentioned, such a division is merely exemplary and not mandatory. Indeed, the features and functionality of two or more of the units/modules described above may be embodied in one unit/module according to embodiments of the invention. Conversely, the features and functions of one unit/module described above may be further divided into embodiments by a plurality of units/modules.
Example 4
The present embodiment provides a computer-readable storage medium on which a computer program is stored, which when executed by a processor implements the model training method provided in embodiment 1.
More specific examples, among others, that the readable storage medium may employ may include, but are not limited to: a portable disk, a hard disk, random access memory, read only memory, erasable programmable read only memory, optical storage device, magnetic storage device, or any suitable combination of the foregoing.
In a possible implementation, the invention may also be implemented in the form of a program product comprising program code for causing a terminal device to perform a method for model training as described in embodiment 1 when the program product is run on the terminal device.
Where program code for carrying out the invention is written in any combination of one or more programming languages, the program code may execute entirely on the user device, partly on the user device, as a stand-alone software package, partly on the user device and partly on a remote device or entirely on the remote device.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that this is by way of example only, and that the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention, and these changes and modifications are within the scope of the invention.
Example 5
As shown in fig. 4, the present embodiment provides a wind power prediction method, which includes:
step 201, a time point t + i to be predicted of wind power of a wind power plant is determined, wherein t represents the current time.
In this embodiment, the value of the time point t + i to be predicted may be any time point in the future.
Step 202, obtaining a wind power prediction model ANN corresponding to a to-be-predicted time point t + i of wind power of a wind power planti。
In this embodiment, for each time point t + i to be predicted, a neural network model is trained based on the historical training data corresponding to the time point t + i, so as to correspondingly train a wind power prediction model ANNi。
Step 203, obtaining tjUsing observation data at the moment and prediction data at the t + i moment as a wind power prediction model ANNiWherein t isjThe time is the latest historical time from the current time t.
In this embodiment, t is obtainedjThe observed data at the moment and the predicted data at the t + i moment are used as the input of a wind power prediction model, and specifically t is obtainedjUsing observation data at the moment and prediction data at the t + i moment as a wind power prediction model ANNiWherein t isjThe time is the latest historical time from the current time t.
In this embodiment, the observation data of the historical time closest to the current time t is obtained, so as to obtain the latest observation data of the wind farm.
Step 204, utilizing a wind power prediction model ANNiAnd predicting the wind power of the moment t + i to be predicted.
In this embodiment, the wind power of the time point t + i to be predicted is predicted by using a wind power prediction model, and specifically, the wind power prediction model ANN is usediAnd predicting the wind power of the moment t + i to be predicted. The wind power prediction model is obtained by training by using the training method of embodiment 1.
In an implementation scenario, if the observed data of the first historical period and the second historical period are both wind speed and wind direction observed values, and the predicted data of the second historical period is a wind speed and wind direction predicted value, step 204 specifically includes:
will tjInputting the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i moment into a wind power prediction model ANNi。
And obtaining a wind power predicted value at the t + i moment based on the corresponding relation between the wind speed and the wind direction and the wind power of the wind power plant.
In this embodiment, the wind power predicted value at the time of t + i can be obtained by querying a wind power curve of the fan.
In this embodiment, the wind speed and direction may be a predicted value of wind speed and direction at a specific time, specifically, tjInputting the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i moment into a wind power prediction model ANNiAnd obtaining the wind speed and direction predicted value corrected at the time t + i.
And obtaining the wind power predicted value at the time t + i based on the corresponding relation between the wind speed and wind direction predicted value corrected at the time t + i and the wind power of the wind power plant.
In an implementation scenario, if the observed data of the first historical period and the second historical period are both wind power observed values, and the predicted data of the second historical period is a wind speed and direction predicted value, step 204 specifically includes:
will tjInputting the observed wind power value at the moment and the predicted wind speed and direction value at the t + i moment into a wind power prediction model ANNiAnd obtaining the predicted value of the wind power at the moment of t + i.
The method takes the observation data of each moment in the first historical period and the prediction data of each moment in the second historical period in the obtained historical training data of the wind power plant as input, takes the observation data of each moment in the second historical period as output, trains the neural network model to obtain the wind power prediction model of the moment to be predicted, realizes that the wind power of any future moment is predicted by inputting the latest observation data into the wind power prediction model, reduces the prediction error of the ultra-short-term wind power, and trains the corresponding wind power prediction models for each future moment, thereby reducing the prediction error generated by the uncertainty of the wind power prediction model and the input quantity and improving the accuracy of the wind power prediction model.
Example 6
As shown in fig. 5, the present embodiment provides a wind power prediction system including a determination module 61, a third acquisition module 62, a second acquisition module 63, and a prediction module 64.
The determination module 61 is configured to determine a time point t + i to be predicted of wind power of the wind farm, where t represents a current time.
In this embodiment, the value of the time point t + i to be predicted may be any time point in the future.
The third obtaining module 62 is configured to obtain a prediction model ANN corresponding to a to-be-predicted time point t + i of wind power of the wind farmi。
In this embodiment, for each time point t + i to be predicted, a neural network model is trained based on the historical training data corresponding to the time point t + i, so as to correspondingly train a wind power prediction model ANNi。
The second obtaining module 63 is used for obtaining tjThe observed data at the moment and the predicted data at the t + i moment are used as the input of a wind power prediction model, and specifically t is obtainedjUsing observation data at the moment and prediction data at the t + i moment as a wind power prediction model ANNiWherein t isjThe time is the latest historical time from the current time t.
In this embodiment, the observation data of the historical time closest to the current time t is obtained, so as to obtain the latest observation data of the wind farm.
The prediction module 64 is configured to predict the wind power of the time point t + i to be predicted by using a wind power prediction model.
In this embodiment, the prediction module 64 is specifically configured to utilize a wind power prediction model ANNiAnd predicting the wind power of the moment t + i to be predicted. The wind power prediction model is obtained by training with the training system of embodiment 2.
In an implementation scenario, as shown in fig. 5, if the observed data of the first historical period and the second historical period are both wind speed and wind direction observed values, and the predicted data of the second historical period is a wind speed and wind direction predicted value, the prediction module 64 includes an obtaining unit 641 and a prediction unit 642.
The obtaining unit 641 is used for obtaining tjInputting the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i moment into a wind power prediction model ANNi。
The prediction unit 642 is configured to obtain a wind power prediction value at the time t + i based on a corresponding relationship between a wind speed and a wind direction and a wind power of the wind farm.
In this embodiment, the wind power predicted value at the time of t + i can be obtained by querying a wind power curve of the fan.
In this embodiment, the wind speed and the wind direction may be predicted values of the wind speed and the wind direction at specific time, and specifically, the obtaining unit 641 is configured to use tjInputting the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i moment into a prediction model ANNiAnd obtaining the wind speed and direction predicted value corrected at the time t + i.
The prediction unit 642 is configured to obtain a wind power predicted value at the time t + i based on the corresponding relationship between the wind speed and wind direction predicted value corrected at the time t + i and the wind power of the wind farm.
In an implementation scenario, if the observed data of the first historical period and the second historical period are both wind power observed values, and the predicted data of the second historical period is a wind speed and direction predicted value, the prediction module 64 is specifically configured to predict tjInputting the observed wind power value at the moment and the predicted wind speed and direction value at the t + i moment into a wind power prediction model ANNiTo obtainAnd (4) predicting the wind power at the t + i moment.
The method takes the observation data of each moment in the first historical period and the prediction data of each moment in the second historical period in the obtained historical training data of the wind power plant as input, takes the observation data of each moment in the second historical period as output, trains the neural network model to obtain the wind power prediction model of the moment to be predicted, realizes that the wind power of any future moment is predicted by inputting the latest observation data into the wind power prediction model, reduces the prediction error of the ultra-short-term wind power, and trains the corresponding wind power prediction models for each future moment, thereby reducing the prediction error generated by the uncertainty of the wind power prediction model and the input quantity and improving the accuracy of the wind power prediction model.
Example 7
A schematic structural diagram of an electronic device provided in embodiment 7 of the present invention is the same as the structure in fig. 3. The electronic device comprises a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the wind power prediction method of embodiment 5 when executing the program. The electronic device 30 shown in fig. 3 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiment of the present invention.
As shown in fig. 3, the electronic device 30 may be embodied in the form of a general purpose computing device, which may be, for example, a server device. The components of the electronic device 30 may include, but are not limited to: the at least one processor 31, the at least one memory 32, and a bus 33 connecting the various system components (including the memory 32 and the processor 31).
The bus 33 includes a data bus, an address bus, and a control bus.
The memory 32 may include volatile memory, such as Random Access Memory (RAM)321 and/or cache memory 322, and may further include Read Only Memory (ROM) 323.
Memory 32 may also include a program/utility 325 having a set (at least one) of program modules 324, such program modules 324 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.
The processor 31 executes various functional applications and data processing, such as the wind power prediction method of embodiment 5 of the present invention, by running a computer program stored in the memory 32.
The electronic device 30 may also communicate with one or more external devices 34 (e.g., keyboard, pointing device, etc.). Such communication may be through input/output (I/O) interfaces 35. Also, model-generating device 30 may also communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the Internet) via network adapter 36. As shown in FIG. 3, network adapter 36 communicates with the other modules of model-generating device 30 via bus 33. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the model-generating device 30, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems, etc.
It should be noted that although in the above detailed description several units/modules or sub-units/modules of the electronic device are mentioned, such a division is merely exemplary and not mandatory. Indeed, the features and functionality of two or more of the units/modules described above may be embodied in one unit/module according to embodiments of the invention. Conversely, the features and functions of one unit/module described above may be further divided into embodiments by a plurality of units/modules.
Example 8
The present embodiment provides a computer-readable storage medium on which a computer program is stored, which when executed by a processor implements the wind power prediction method provided in embodiment 5.
More specific examples, among others, that the readable storage medium may employ may include, but are not limited to: a portable disk, a hard disk, random access memory, read only memory, erasable programmable read only memory, optical storage device, magnetic storage device, or any suitable combination of the foregoing.
In a possible implementation, the invention may also be implemented in the form of a program product comprising program code for causing a terminal device to perform a method of implementing a wind power prediction as described in embodiment 5, when the program product is run on the terminal device.
Where program code for carrying out the invention is written in any combination of one or more programming languages, the program code may execute entirely on the user device, partly on the user device, as a stand-alone software package, partly on the user device and partly on a remote device or entirely on the remote device.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that this is by way of example only, and that the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention, and these changes and modifications are within the scope of the invention.

Claims (18)

1. A method of model training, comprising:
acquiring historical training data of a wind power plant, wherein the historical training data comprises observation data of each moment in a first historical period and prediction data of each moment in a second historical period as input, and observation data of each moment in the second historical period as output; the duration of the first historical period is the same as the duration of the second historical period, the earliest moment in the first historical period is earlier than the earliest moment in the second historical period, and the observed data and the predicted data are both related to wind power calculation of the wind farm;
and training a neural network model based on the historical training data to obtain a wind power prediction model at the moment to be predicted.
2. The model training method of claim 1, wherein the step of training a neural network model based on the historical training data to obtain a wind power prediction model for a time to be predicted comprises:
aiming at each time t + i to be predicted, training a neural network model based on the historical training data corresponding to the time t + i to correspondingly train a wind power prediction model ANNi。
3. The model training method of claim 2, wherein the step of obtaining historical training data for a wind farm comprises: acquiring observation data of m historical moments and prediction data of m historical moments, wherein the range of the m historical moments comprises the first historical time period and the second historical time period, and the m historical moments are sequentially marked as t-m, t-m +1, … and t-1 according to the time sequence, wherein t represents the current moment;
aiming at each time t + i to be predicted, training a neural network model based on the historical training data corresponding to the time t + i to correspondingly train a wind power prediction model ANNiComprises the following steps:
training each wind power prediction model ANNiThe historical training data includes [ A ]k，Bk+i，Ck+i]Wherein A iskObservation data representing a first history period, Bk+iPrediction data representing a second history period, Ck+iObservation data representing a second historical period, wherein the moments of the first historical period are the t-m th moment, the t-m +1 th moment, … and the t-1-i th moment in sequence; the moments of the second historical period are the t-m + i th moment, the t-m +1+ i th moment, … and the t-1 th moment in sequence.
4. The model training method of claim 1, wherein the observed data of the first historical period and the second historical period are both wind speed and wind direction observed values, and the predicted data of the second historical period is a wind speed and wind direction predicted value, or the observed data of the first historical period and the second historical period are both wind power observed values, and the predicted data of the second historical period is a wind speed and wind direction predicted value.
5. A model training system is characterized by comprising a first acquisition module and a training module;
the first acquisition module is used for acquiring historical training data of the wind power plant, wherein the historical training data comprises observation data of each moment in a first historical period and prediction data of each moment in a second historical period as input, and observation data of each moment in the second historical period as output; the duration of the first historical period is the same as the duration of the second historical period, the earliest moment in the first historical period is earlier than the earliest moment in the second historical period, and the observed data and the predicted data are both related to wind power calculation of the wind farm;
the training module is used for training a neural network model based on the historical training data to obtain a wind power prediction model at the moment to be predicted.
6. The model training system of claim 5, wherein the training module is specifically configured to, for each time t + i to be predicted, train the neural network model based on the historical training data corresponding to the time t + i to correspondingly train a wind power prediction model ANNi。
7. The model training system of claim 6, wherein the first obtaining module is specifically configured to obtain observed data of m historical moments and predicted data of m historical moments, a range of the m historical moments includes the first historical period and the second historical period, the m historical moments are sequentially marked as t-m, t-m +1, …, t-1 according to a chronological order, where t represents a current moment;
the trainingThe training module is specifically configured to train each wind power prediction model ANNiThe historical training data includes [ A ]k，Bk+i，Ck+i]Wherein A iskObservation data representing a first history period, Bk+iPrediction data representing a second history period, Ck+iObservation data representing a second historical period, wherein the moments of the first historical period are the t-m th moment, the t-m +1 th moment, … and the t-1-i th moment in sequence; the moments of the second historical period are the t-m + i th moment, the t-m +1+ i th moment, … and the t-1 th moment in sequence.
8. The model training system of claim 5, wherein the observed data for the first historical period and the second historical period are both wind speed and direction observed values and the predicted data for the second historical period is a wind speed and direction predicted value, or wherein the observed data for the first historical period and the second historical period are both wind power observed values and the predicted data for the second historical period is a wind speed and direction predicted value.
9. A method of wind power prediction, comprising:
determining a time point t + i to be predicted of wind power of a wind power plant, wherein t represents the current time;
obtaining tjThe observed data at the moment and the predicted data at the t + i moment are used as the input of a wind power prediction model, wherein t isjThe moment is the historical moment nearest to the current moment t;
predicting the wind power of the time point t + i to be predicted by using a wind power prediction model;
wherein the wind power prediction model is trained using the training method of any one of claims 1-4.
10. The wind power prediction method of claim 9, wherein if for each time point t + i to be predicted, training a neural network model based on the historical training data corresponding to the time t + i is performed to correctA wind power prediction model ANN is trainedi(ii) a The wind power prediction method further comprises:
obtaining a wind power prediction model ANN corresponding to a to-be-predicted time point t + i of wind power of the wind power planti；
Using the wind power prediction model ANNiAnd predicting the wind power of the time point t + i to be predicted.
11. The method of claim 10, wherein the wind power prediction model ANN is used if the observed data of the first historical period and the second historical period are both wind speed and wind direction observed values, and the predicted data of the second historical period is a wind speed and wind direction predicted valueiThe step of predicting the wind power of the time point t + i to be predicted specifically includes:
will tjInputting the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i moment into the wind power prediction model ANNi；
And obtaining the wind power predicted value at the t + i moment based on the corresponding relation between the wind speed and the wind direction and the wind power of the wind power plant.
12. The method of claim 10, wherein the wind power prediction model ANN is used if the observed data of the first historical period and the second historical period are both wind power observed values and the predicted data of the second historical period is a wind speed and wind direction predicted valueiThe step of predicting the wind power of the time point t + i to be predicted specifically includes:
will tjInputting the observed wind power value at the moment and the predicted wind speed and direction value at the t + i moment into the wind power prediction model ANNiAnd obtaining the predicted value of the wind power at the t + i moment.
13. A wind power prediction system is characterized by comprising a determining module, a second obtaining module and a prediction module;
the determining module is used for determining a to-be-predicted time point t + i of wind power of the wind power plant, wherein t represents the current time;
the second acquisition module is used for acquiring tjThe observed data at time and the predicted data at time t + i are used as input of a prediction model, wherein t isjThe moment is the historical moment nearest to the current moment t;
the prediction module is used for predicting the wind power of the time point t + i to be predicted by using a wind power prediction model;
wherein the wind power prediction model is trained using a training system according to any one of claims 5-8.
14. The wind power prediction system of claim 13, wherein if for each time t + i to be predicted, training a neural network model based on the historical training data corresponding to the time t + i to correspondingly train a wind power prediction model ANNi(ii) a The wind power prediction system further comprises a third obtaining module;
the third obtaining module is used for obtaining a wind power prediction model ANN corresponding to a to-be-predicted time point t + i of the wind power planti；
The prediction module is specifically configured to utilize the wind power prediction model ANNiAnd predicting the wind power of the time point t + i to be predicted.
15. The wind power prediction system of claim 14 wherein the prediction module comprises an acquisition unit and a prediction unit if the observed data of the first historical period and the second historical period are both wind speed and direction observed values and the predicted data of the second historical period is a wind speed and direction predicted value;
the acquisition unit is used for converting tjInputting the observed value of wind speed and wind direction at the moment and the predicted value of wind speed and wind direction at the t + i moment into the wind power prediction model ANNi；
And the prediction unit is used for obtaining a wind power prediction value at the t + i moment based on the corresponding relation between the wind speed and the wind direction and the wind power of the wind power plant.
16. The wind power prediction system of claim 14, wherein the prediction module is specifically configured to predict t if the observed data of the first historical period and the second historical period are both wind power observed values and the predicted data of the second historical period is a wind speed and direction predicted valuejInputting the observed wind power value at the moment and the predicted wind speed and direction value at the t + i moment into the wind power prediction model ANNiAnd obtaining the predicted value of the wind power at the t + i moment.
17. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the model training method according to any one of claims 1 to 4 or the wind power prediction method according to any one of claims 9 to 12 when executing the computer program.
18. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a model training method according to any one of claims 1-4, or carries out a wind power prediction method according to any one of claims 9-12.
CN202110804988.7A 2021-07-16 2021-07-16 Model training method, wind power prediction method, system, device and medium Pending CN113537595A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110804988.7A CN113537595A (en) 2021-07-16 2021-07-16 Model training method, wind power prediction method, system, device and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110804988.7A CN113537595A (en) 2021-07-16 2021-07-16 Model training method, wind power prediction method, system, device and medium

Publications (1)

Publication Number Publication Date
CN113537595A true CN113537595A (en) 2021-10-22

Family

ID=78128373

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110804988.7A Pending CN113537595A (en) 2021-07-16 2021-07-16 Model training method, wind power prediction method, system, device and medium

Country Status (1)

Country Link
CN (1) CN113537595A (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016095708A1 (en) * 2014-12-16 2016-06-23 高德软件有限公司 Traffic flow prediction method, and prediction model generation method and device
CN111160621A (en) * 2019-12-06 2020-05-15 江苏方天电力技术有限公司 Short-term wind power prediction method integrating multi-source information
CN111429006A (en) * 2020-03-24 2020-07-17 北京明略软件系统有限公司 Financial risk index prediction model construction method and device and risk situation prediction method and device
CN111461444A (en) * 2020-04-07 2020-07-28 上海电气风电集团股份有限公司 Prediction method, system, medium and electronic device for unit power of wind power plant
CN112651560A (en) * 2020-12-28 2021-04-13 华润电力技术研究院有限公司 Ultra-short-term wind power prediction method, device and equipment
CN112735094A (en) * 2020-12-17 2021-04-30 中国地质环境监测院 Geological disaster prediction method and device based on machine learning and electronic equipment
CN112863179A (en) * 2021-01-11 2021-05-28 上海交通大学 Intersection signal lamp control method based on neural network model predictive control

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016095708A1 (en) * 2014-12-16 2016-06-23 高德软件有限公司 Traffic flow prediction method, and prediction model generation method and device
CN111160621A (en) * 2019-12-06 2020-05-15 江苏方天电力技术有限公司 Short-term wind power prediction method integrating multi-source information
CN111429006A (en) * 2020-03-24 2020-07-17 北京明略软件系统有限公司 Financial risk index prediction model construction method and device and risk situation prediction method and device
CN111461444A (en) * 2020-04-07 2020-07-28 上海电气风电集团股份有限公司 Prediction method, system, medium and electronic device for unit power of wind power plant
CN112735094A (en) * 2020-12-17 2021-04-30 中国地质环境监测院 Geological disaster prediction method and device based on machine learning and electronic equipment
CN112651560A (en) * 2020-12-28 2021-04-13 华润电力技术研究院有限公司 Ultra-short-term wind power prediction method, device and equipment
CN112863179A (en) * 2021-01-11 2021-05-28 上海交通大学 Intersection signal lamp control method based on neural network model predictive control

Similar Documents

Publication Publication Date Title
CN111461444B (en) Method, system, medium and electronic equipment for predicting unit power of wind power plant
CN110991666B (en) Fault detection method, training device, training equipment and training equipment for model, and storage medium
CN113937763B (en) Wind power prediction method, device, equipment and storage medium
JP2017187371A (en) Weather prediction device and wind power generation
CN114048896B (en) Method, system, equipment and medium for predicting photovoltaic power generation data
Cui et al. An algorithm for forecasting day-ahead wind power via novel long short-term memory and wind power ramp events
Sengar et al. Ensemble approach for short term load forecasting in wind energy system using hybrid algorithm
CN104036328A (en) Self-adaptive wind power prediction system and prediction method
WO2023115425A1 (en) Ultra-short-time wind speed prediction method and system
CN113822418A (en) Wind power plant power prediction method, system, device and storage medium
CN115034485A (en) Wind power interval prediction method and device based on data space
Mahmoud et al. Advanced method for short‐term wind power prediction with multiple observation points using extreme learning machines
WO2022143903A1 (en) Safety evaluation method and system for vehicle battery, and device and readable storage medium
US20230327440A1 (en) Power generation amount management system and power generation amount management method
CN114330110A (en) Model training method, photovoltaic power generation power prediction method and system
CN114819385A (en) Wind power prediction method and device, electronic equipment and storage medium
CN110188939B (en) Wind power prediction method, system, equipment and storage medium of wind power plant
CN112131753A (en) Method, system and device for evaluating fatigue life of fan and readable medium
CN112052604A (en) Method, system, equipment and readable medium for predicting equivalent fatigue load of fan
CN113537595A (en) Model training method, wind power prediction method, system, device and medium
WO2023178305A2 (en) Systems and methods for forecasting power generated by variable power generation assets
CN116484743A (en) Evaluation method and device for influence of wind power grid connection on reliability of power system
CN112906757A (en) Model training method, optical power prediction method, device, equipment and storage medium
CN114552570A (en) Offshore wind power prediction management system
Liu et al. A Bayesian Deep Learning-Based Probabilistic Risk Assessment and Early-Warning Model for Power Systems Considering Meteorological Conditions

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20211022