CN106251242B - Wind power output interval combination prediction method - Google Patents
Wind power output interval combination prediction method Download PDFInfo
- Publication number
- CN106251242B CN106251242B CN201610642950.3A CN201610642950A CN106251242B CN 106251242 B CN106251242 B CN 106251242B CN 201610642950 A CN201610642950 A CN 201610642950A CN 106251242 B CN106251242 B CN 106251242B
- Authority
- CN
- China
- Prior art keywords
- prediction
- output
- interval
- wind power
- wind
- 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 95
- 230000008859 change Effects 0.000 claims abstract description 84
- 238000005457 optimization Methods 0.000 claims abstract description 15
- 238000013528 artificial neural network Methods 0.000 claims description 18
- 238000013507 mapping Methods 0.000 claims description 8
- 238000010248 power generation Methods 0.000 abstract description 7
- 230000002349 favourable effect Effects 0.000 abstract 1
- 238000004458 analytical method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005183 dynamical system Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Wind Motors (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a wind power output interval combined prediction method, which aims at the problems that the wind power generation is influenced by natural wind speed fluctuation and has larger uncertainty and the prediction precision is reduced due to the fact that the prediction time is advanced, carries out range prediction on wind power output, realizes the combined setting of the wind power output prediction range on the basis of three methods, obtains the wind power output prediction interval range on the basis of three different methods, namely a wind speed change ratio, a predicted value change rate and an actual power optimization value, and then selects the optimal wind power output prediction interval in each time interval according to historical wind power output data. The method integrates different parameters of wind power output such as wind speed change rate, predicted value change rate, actual power and the like, selects an optimal prediction interval from the parameters, and is favorable for realizing more accurate wind power output prediction.
Description
Technical Field
The invention relates to the field of wind power generation output prediction, in particular to wind power output prediction based on wind speed uncertain interval prediction, and a wind power output interval combined prediction model is established based on three different methods.
background
wind power generation is the main way of clean power generation, however, the wind power generation is influenced by the wind speed of natural wind, and great uncertainty exists. Accurate wind power prediction is an effective way for improving the utilization efficiency of wind power and is an important means for ensuring the balance of a power grid under the condition of wind power access. Wind power generation enterprises can reasonably arrange maintenance by using the prediction results, the capacity coefficient of the wind power plant is improved, and the power generation cost is reduced. Wind power output is greatly influenced by factors such as weather and seasons, and main influence factors are as follows: wind speed; the density of the air; the wind direction. The wind speed has the greatest influence on the wind power output, the natural wind speed is not constant, the change in the short time is possibly large, and the prediction on the wind power output is very unfavorable.
At present, the main wind power prediction methods include: a time series prediction method, a regression analysis method, a grey theory method, an expert system method, a support vector machine and the like. The time series prediction method establishes a mathematical model according to historical data and predicts future target data; the regression analysis method is used for analyzing the observation data of the variables by utilizing mathematical statistics and determining the relationship among the variables; the grey theory method is mainly used for modeling data polluted by noise; the expert system method is used for reasoning and making decisions by using stored expert knowledge and experience; the support vector machine is based on statistical learning and structural risk minimization principles. Besides the method, the artificial neural network is widely applied to wind power prediction. The artificial neural network is a nonlinear dynamical system with self-learning, self-organizing and self-adapting capabilities. The method utilizes the past data to continuously correct the weight among the neurons of the neural network, thereby obtaining a more ideal prediction model. A network system formed by a large number of neurons can realize functions of nonlinear mapping, classification recognition, optimization calculation and the like, and is suitable for solving the wind power output prediction problem.
disclosure of Invention
The purpose of the invention is as follows: in order to predict the wind power output more accurately, the invention provides a wind power output interval combination method which can improve the prediction precision of the wind power output interval.
the technical scheme is as follows: in order to achieve the purpose, the wind power output interval combination method comprises the following steps of:
(1) dividing the prediction time interval of the wind power output interval for a certain area at a preset period;
(2) collecting wind power output historical data of each time period and wind speed data of each wind speed collection point of a wind power plant;
(3) For a certain prediction time period, setting a wind power output prediction range by utilizing the collected wind speed data based on a wind speed change rate method and a predicted value change rate method respectively, and optimizing the wind power output prediction range by utilizing the collected wind power output historical data based on an actual power method;
(4) And analyzing the prediction precision of the three methods by using the actual data and the prediction data of the wind power plant output at a time period above the prediction time period, selecting the optimal prediction method of the prediction time period, and taking the corresponding wind power output prediction interval as the wind power output prediction interval of the combined prediction.
specifically, the step (3) of setting the wind power output prediction range based on the wind speed change rate by using the collected wind speed data in a certain prediction period includes the following steps:
1) Calculating the wind speed change ratio between the wind speed data at each moment and the wind speed data at the previous moment in the wind speed sequence at the previous moment of the prediction time interval to form a wind speed change ratio sequence at the previous moment;
2) Based on BP neural network operation, predicting wind speed data at each moment in the prediction time period by using the wind speed sequence in the previous time period, and acquiring a wind speed change rate sequence in the prediction time period;
3) Dividing wind speed change rate intervals according to equal length for the wind speed change rate in the wind speed change rate sequence of the prediction time period, counting the wind speed change rate in each interval to obtain a wind speed change rate interval with the maximum probability, taking the middle value of the interval as a wind speed uncertain interval coefficient, and multiplying the actually measured wind speed at the moment by the uncertain interval coefficient to perform wind speed setting for a certain moment of the prediction time period to obtain the prediction interval of the wind power plant output at the next moment:
Ψ(v×(1-d),t)≤Pt+1≤Ψ(v×(1+d),t)
Psi represents the nonlinear mapping relation between the wind speed at the moment t and the output of the wind power plant at the moment t +1, and the nonlinear mapping relation is obtained through BP neural network operation.
Specifically, the step (3) of setting the wind power output prediction range based on the predicted value change rate by using the collected wind speed data in a certain prediction period includes the following steps:
1) calculating the change rate of the output power at each moment in the last period of the prediction period;
2) Predicting the output power of each moment in the prediction period by using the output power change rate and the output power of each moment in the previous period;
3) Calculating the output power change rate of each moment in the prediction time period, dividing the output power change rate interval according to the equal length, counting the output power change rate in each interval to obtain an interval with the maximum probability, and taking the upper bound of the interval as a power setting parameter delta;
4) Setting the wind power plant output by using the power setting parameter delta to obtain a prediction interval theta of the wind power plant output at a certain moment k +1 in the prediction time period as follows:
Θ＝[(1-δ)·p(k+1)，(1+δ)·p(k+1)]。
specifically, the step (3) of optimizing the wind power output prediction range based on the actual power in a certain prediction time period includes the following steps:
1) training a BP neural network by using historical wind power output data and wind speed data;
2) sequentially acquiring the output power predicted value of each unknown moment in the predicted time period by using the trained BP neural network;
3) For a certain unknown moment of the prediction time interval, calculating a difference value between an actual output power value and a predicted output power value at the previous moment of the unknown moment, taking the difference value as an optimization parameter delta at the moment, optimizing the predicted output power value at the moment by using the optimization parameter delta to obtain a prediction interval of the wind power plant output at the moment, wherein the prediction interval of the wind power plant output at a certain unknown moment k is as follows:
Θk＝[(1-δ)·pk，(1+δ)·pk]。
In the step (4), the actual data of the wind farm output at a time period above the prediction time period and the prediction interval are used for analyzing the prediction accuracy of the three methods and selecting the optimal prediction method of the prediction time period, which specifically comprises the following steps:
and counting the total number of the actual data of the wind power plant output at each moment in the previous period, which respectively fall into the prediction intervals of the three methods, and taking the method corresponding to the maximum total number as the optimal prediction method of the prediction period.
Has the advantages that: the wind power output interval combination method comprises the steps of firstly obtaining a wind power output prediction interval range by utilizing a neural network model and a sequence model based on three different methods of a wind speed change ratio, a predicted value change ratio and an actual power optimization value, then selecting an optimal wind power output prediction interval in each time interval according to historical wind power output data, and selecting an optimal prediction method for prediction of each time interval based on historical data, so that the defect that a certain prediction method is limited when being used independently is overcome, and the accuracy of the wind power output interval prediction is improved; according to the wind power output interval combination method, aiming at wind power output uncertainty, traditional wind power deterministic prediction is replaced by output interval prediction, and the method is more suitable for the accidental characteristics of wind power output; the method can be used for analyzing and researching different geographic environments and wind power plant installation conditions, has strong adaptability, and has a good gain effect on improvement and popularization of the wind power output interval prediction method.
drawings
FIG. 1 is a flow chart of a wind power output interval combination prediction method in the present invention;
FIG. 2 is a flow chart of wind power output prediction range optimization based on actual power;
FIG. 3 is a result of predicting the output of a wind farm in a 95% -105% wind speed interval in a prediction period based on a wind speed variation ratio method in embodiment 1;
Fig. 4 is a prediction result of the wind power output interval in the prediction period based on the predicted value change rate method in embodiment 1;
Fig. 5 is a wind power output prediction range optimization result in a prediction period of the wind power output prediction range optimization setting method based on actual power in embodiment 1.
Detailed Description
The method of the present invention will be further described with reference to the following examples.
as shown in fig. 1, the wind power output interval combination prediction method of the present invention includes the following steps:
(1) and dividing the wind power output interval prediction time interval, and acquiring wind power output historical data of each time interval and wind speed data of each wind speed acquisition point of the wind power plant.
The output of the wind power plant is mainly influenced by the wind speed, the wind speed has large uncertainty, but the wind speed in the same region and the same time period (season) has a certain rule, so the time period for predicting the wind power output interval is divided, the time period for dividing the seasons in different regions is different mainly based on the same season (month) and the same time period (date), and the time period division adopts 2-4 days as a cycle.
for example: for a certain region, 1 to 4 of 8 months are taken as a time period, 5 to 9 are taken as a time period, and the rest is done in sequence by taking 4 days as a cycle.
(2) the method comprises the following steps of utilizing collected wind speed data of each wind speed collection point of the wind power plant to predict the output interval of the wind power plant by adopting three different methods, wherein the previous period data is adopted to predict the next period:
a) Wind power output prediction range setting based on wind speed change ratio
the accuracy of the wind power plant output prediction is closely related to the accuracy of the wind speed prediction, in order to define the wind power plant output range, the wind speed change can be firstly predicted, and the wind power plant output interval prediction is realized based on the wind speed change prediction interval range.
defining a wind speed change rate:
for a wind speed sequence v 1 [ k ] - (v 1, v 2, v 3.. v k } in a certain time period, namely, each item of a next time v 1 [ k ]' - (v 2, v 3.. v k+1 } sequence and a v 1 [ k ] - (v 1, v 2, v 3.. v k } sequence are firstly subjected to difference, and then the proportion of the difference relative to the wind speed at the last time is calculated, so that a wind speed change rate sequence of the time period is obtained:
ew[k]＝{ew(2),ew(3),ew(4),...ew(k+1)}；
Training the BP neural network by using the wind speed sequence of the last time period of a certain prediction time period, predicting the wind speed data of each moment in the prediction time period, and further obtaining the wind speed change rate sequence of the prediction time period;
the wind speed change rate in the wind speed change rate sequence for the prediction period is divided into wind speed change rate intervals according to equal length, for example: dividing the change rate interval into a plurality of equal-length intervals of 0% -5%, 5% -10%, 10% -15% and the like, counting the wind speed change rate falling in each interval to obtain a wind speed change rate interval with the maximum probability of 5% -10%, taking the middle value of the interval as 7.5% of a wind speed uncertain interval coefficient d, and multiplying the actually measured wind speed at the moment by the uncertain interval coefficient d to perform wind speed setting at a certain moment of the prediction time interval to obtain the prediction interval of the wind power plant output at the next moment:
Ψ(v×(1-d),t)≤Pt+1≤Ψ(v×(1+d),t)，
p t+1 represents the output of the wind power plant at the next moment at the moment t, Ψ represents a nonlinear mapping relation between the wind speed at the moment t and the output of the wind power plant at the moment t +1, and the nonlinear mapping relation can be obtained through BP neural network operation.
b) wind power output prediction range setting based on predicted value change rate
because the output power of the wind power plant has larger randomness, sudden change of the wind speed often causes large change of the output power, and the prediction accuracy of points with larger change amplitude is reduced along with the large change. On the basis of the known output power predicted value, the wind power output prediction range can be adjusted by utilizing the change rate of the output power.
defining the output power change rate at a certain time k:
predicting the output power of each moment in a certain prediction time period by using the output power change rate and the output power of each moment in the certain prediction time period;
Calculating the output power change rate of each moment in the prediction period, dividing the output power change rate interval according to the equal length, counting the output power change rate in each interval to obtain the interval with the maximum probability, and taking the upper bound of the interval as a power setting parameter delta, for example: dividing the output power into a plurality of equal-length intervals of 0% -5%, 5% -10%, 10% -15% and the like, and if the probability of the interval falling within the range of 5% -10% is maximum, setting the parameter delta to be 10%;
The power setting parameter delta is used for setting the output of the wind power plant, and the (output power) prediction interval theta of the output of the wind power plant at a certain moment k +1 in the prediction time period is obtained as follows:
Θ＝[(1-δ)·p(k+1)，(1+δ)·p(k+1)]。
c) Wind power output prediction range optimization based on actual power
Both methods are only used for predicting the wind power output in a certain period of time in the future based on historical data, and do not relate to the actual output of the wind power in the prediction period of time. In the prediction process, the wind power prediction interval of the future time node can be corrected by combining the acquired historical data of the actual wind power output.
In the case that the actual output power at the past moment of a certain prediction period is known, the power prediction interval of the BP neural network output is optimized by using a known time period [0, k-1] actual output power sequence p [ k-1] = { r 1, r 2, r 3.. r k-1 }. as shown in fig. 2, the method specifically comprises the following steps:
1) Training a BP neural network by using historical wind power output data and wind speed data to obtain a relation between input wind speed and output wind power plant output;
2) sequentially obtaining the output predicted value of the wind power plant at each unknown moment in the prediction time period by using the trained BP neural network;
3) for a certain unknown moment of the prediction time interval, taking the difference value between the actual wind farm output value and the predicted wind farm output value at the previous moment of the unknown moment as an optimization parameter delta at the moment, optimizing the predicted wind farm output value at the moment by using the optimization parameter delta to obtain a prediction interval of the wind farm output at the moment, wherein the prediction interval of the wind farm output at a certain unknown moment k is as follows:
Θk＝[(1-δ)·pk，(1+δ)·pk]。
for example, for a time k, when the wind farm output actual value sequence p [ k-n, k-1] (p k-n, p k-n+1.. p k-1) and the predicted value sequence r [ k-n, k-1] (rk-n, r k-n+1.. r k-1) of the past n time points are known, the difference between the wind farm output actual value and the predicted value at a certain time k-1 is Δ p k-1 ═ p k-1 -r k-1.
(3) and aiming at different conditions, the interval combination prediction of the wind power field output is carried out on the basis of the three methods. Under the condition that the actual output of the wind power plant at each moment of the prediction time period is unknown, the wind speed change interval can be used for predicting or the predicted value change rate setting range can be used; under the condition that the actual wind power plant output at the past moment of the prediction period is known, the actual power can be used for optimizing the output power prediction at the subsequent unknown moment, and the set wind power plant output range is obtained.
The method comprises the steps of predicting the wind power output of a certain wind power plant in a certain prediction time period, firstly predicting a wind power output interval by adopting the three methods, analyzing the prediction precision of the three methods by using actual data and prediction data of the wind power plant output in the previous time period, selecting the optimal prediction method of the prediction time period, and taking the corresponding wind power output prediction interval as the wind power output prediction interval of combined prediction.
Example 1:
In the embodiment, a Matlab programming simulation platform is adopted, actual data of a certain wind power plant in a certain period are selected for analysis, three methods are respectively adopted for prediction based on actual historical data, actual prediction effects of the three methods in each period are calculated, and the optimal prediction interval in the period is selected as the optimal interval prediction method in the period. And the prediction interval of each time interval is analogized, so that the integral interval prediction is formed.
selecting a certain time period of historical data for analysis, respectively adopting three methods for prediction, comparing the prediction result with the actual data, and selecting the optimal prediction method (interval) of the time period.
(1) the method comprises the steps of firstly carrying out prediction statistics on the change rate of wind speed at each time point by using a wind speed change ratio-based wind power output prediction range setting method, predicting the wind speed change of the existing wind speed at the next moment through a BP (back propagation) neural network, calculating the ratio of the change amplitude by taking the existing wind speed as a reference, carrying out statistics on all the ratios, and counting the maximum possible interval of the wind speed change at the next moment.
The wind speed at the next moment is counted to be in a wind speed change rate interval of 4% -6% when the wind speed proportion (the probability is 48%) at the previous moment, and the wind speed uncertain interval coefficient d is 5%. Thus, the wind speed data can be given upper and lower limit values, and the predicted output range can be obtained. The result is used as the upper and lower wind speed input limits for 95% -105% of the wind speed measured value input by the artificial neural network, and the obtained wind power output range result is shown in fig. 3.
The black solid line is the actual output of the wind power plant, the gray area is the wind power output upper and lower limit values obtained by taking the wind speed upper and lower limits as input, and it can be seen that most of the actual wind power output value is located in the light color area range, which indicates that the mode of predicting the output upper and lower limit values can be adopted when the wind speed greatly fluctuates. The prediction interval of the wind power output obtained by the model still has certain error at the position with severe fluctuation, but can be more suitable for the uncertainty of the wind power output than the method for predicting the wind power output in the prior art, so that the uncertainty in the wind power prediction is balanced by the adopted wind power output upper and lower limit intervals, and a sufficient basis is provided for selecting the standby capacity of the system.
(2) the method for setting the wind power output prediction range based on the predicted value change rate considers the change rate of the wind power plant output prediction range, and reflects the uncertainty of the wind power output by selecting a reasonable mapping relation: when the output change rate of the wind power plant is within the range of 0-5% and the probability is maximum, the output range of the corresponding wind power plant is 95% -105% of the predicted value; when the output change rate of the wind power plant is within the range of 5-10% and the probability is maximum, the corresponding wind power plant output range is 90-110% of the predicted value; by analogy, in this embodiment, when the output change rate of the wind farm is within the range of 15-20% with the highest probability, the maximum wind farm output range is set to 80% -120%, and the obtained output range result is shown in fig. 4.
as can be seen from the results in fig. 4, for the point with relatively large variation of wind power output, the actual value of the wind power plant output can better fall within the setting range; and the effective reduction of the setting interval can be realized at the point with smaller output fluctuation. However, the actual power at some time points still exceeds the setting range.
(3) And when the condition is that the actual wind power plant output at the past moment of the wind power plant is known in the prediction time period, the difference value between the actual value and the predicted value of the wind power plant output at the previous moment is used as input, and a corresponding wind power plant output range interval is obtained. The relationship between the difference range and the setting interval parameter is shown in the following table 1:
TABLE 1 Difference Range and setting Interval parameter Table
From the interval setting result of the wind power output prediction range optimization method based on actual power shown in fig. 5, it can be found that under the condition that the actual wind power plant output of the wind power plant is known, the output range can be better set, and a better wind power plant output prediction interval can be obtained. The output values of the few wind power plants are beyond the prediction range, and most of the output values of the wind power plants fall within the range.
TABLE 2 wind power output prediction interval accuracy statistics of each method
The comparison results of the three were combined from table 2: adopting a wind power output prediction range setting method based on the wind speed change ratio for 5% floating input, wherein 54 data points (144 total data points) exceeding the setting range are provided; 70 data points exceeding the setting range are obtained by adopting a wind power output prediction range setting method based on the predicted value change rate; 32 data points exceeding the setting range are obtained by adopting the wind power output prediction range optimization method based on the actual power. The result shows that a prediction method based on the wind speed prediction change interval can be adopted under the condition that the output of the wind power plant in real time is unknown, and a better prediction result can be obtained by adopting a method of setting the prediction interval by a real-time value compared with the prediction value change rate under the condition that the output of the wind power plant in real time is known. The method integrates three methods to set the output range of the wind power plant, and mainly predicts the setting based on the wind speed change ratio, sets the predicted value change ratio and corrects the setting range based on the real-time value. The wind power output of a certain place can be respectively predicted by adopting the three methods, the accuracy of each method in different time periods is inspected on the basis of historical actual data, the optimal prediction method in each time period is selected for prediction, and a final predicted output interval is formed.
Claims (4)
1. a wind power output interval combination prediction method is characterized by comprising the following steps:
(1) Dividing the prediction time interval of the wind power output interval for a certain area at a preset period;
(2) collecting wind power output historical data of each time period and wind speed data of each wind speed collection point of a wind power plant;
(3) For a certain prediction time period, setting a wind power output prediction range by utilizing the collected wind speed data based on a wind speed change rate method and a predicted value change rate method respectively, and optimizing the wind power output prediction range by utilizing the collected wind power output historical data based on an actual power method;
(4) analyzing the prediction precision of the three methods by using the actual data and the prediction data of the wind power plant output of the last time period of the prediction time period, selecting the optimal prediction method of the prediction time period, and taking the corresponding wind power output prediction range as the wind power output prediction range of the combined prediction;
In the step (3), the wind power output prediction range is optimized based on an actual power method in a certain prediction time period, and the method comprises the following steps:
1) training a BP neural network by using historical wind power output data and wind speed data;
2) The BP neural network sequentially obtains the output power predicted value of each unknown moment in the predicted time period;
3) For a certain unknown moment of the prediction time interval, calculating a difference value between an actual output power value and a predicted output power value at the previous moment of the unknown moment, taking the difference value as an optimization parameter delta at the moment, optimizing the predicted output power value at the moment by using the optimization parameter delta to obtain a prediction interval of the wind power plant output at the moment, wherein the prediction interval of the wind power plant output at a certain unknown moment k is as follows:
Θk＝[(1-δ)·pk,(1+δ)·pk]
Where p k is the actual value of the wind farm output at a certain moment k.
2. the wind power output interval combination prediction method according to claim 1, wherein the wind power output prediction range is set in step (3) by using the collected wind speed data in a prediction time period based on a wind speed change rate method, and the method comprises the following steps:
1) Calculating the wind speed change ratio between the wind speed data at each moment and the wind speed data at the previous moment in the wind speed sequence at the previous moment of the prediction time interval to form a wind speed change ratio sequence at the previous moment;
2) Based on BP neural network operation, predicting wind speed data at each moment in the prediction time period by using the wind speed sequence in the previous time period, and acquiring a wind speed change rate sequence in the prediction time period;
3) Dividing wind speed change rate intervals according to equal length for the wind speed change rate in the wind speed change rate sequence of the prediction time period, counting the wind speed change rate in each interval to obtain a wind speed change rate interval with the maximum probability, taking the middle value of the interval as a wind speed uncertain interval coefficient, and multiplying the actually measured wind speed at the moment by the uncertain interval coefficient to perform wind speed setting for a certain moment of the prediction time period to obtain the prediction interval of the wind power plant output at the next moment:
Ψ(v×(1-d),t)≤Pt+1≤Ψ(v×(1+d),t)
The psi represents a nonlinear mapping relation between the wind speed at the time t and the wind power plant output at the time t +1, the nonlinear mapping relation is obtained through BP neural network operation, v is an actually measured wind speed at a certain time, d is a wind speed uncertain interval coefficient, and P t+1 represents the wind power plant output at the time t +1 next to the time t.
3. the wind power output interval combination prediction method according to claim 1, wherein the wind power output prediction range is set in the step (3) by using the collected wind speed data in a prediction period based on a prediction value change rate method, and the method comprises the following steps:
1) Calculating the output power change rate of each moment in the last period of the prediction period;
2) predicting the output power of each moment in the prediction period by using the output power change rate and the output power of each moment in the previous period;
3) Calculating the output power change rate of each moment in the prediction time period, dividing the output power change rate interval according to the equal length, counting the output power change rate in each interval to obtain an interval with the maximum probability, and taking the upper bound of the interval as a power setting parameter delta;
4) Setting the wind power plant output by using the power setting parameter delta to obtain a prediction interval theta of the wind power plant output at a certain moment k +1 in the prediction time period as follows:
Θ＝[(1-δ)·pk+1,(1+δ)·pk+1]
wherein p k+1 is the wind farm output at a certain time k +1 in the prediction period.
4. the wind power output interval combined prediction method according to claim 1, characterized in that in step (4), the actual data and the prediction data of the wind farm output at the previous time of the prediction time interval are used for analyzing the prediction accuracy of the three methods and selecting the optimal prediction method of the prediction time interval, specifically:
and (4) counting the total number of the actual data of the wind power plant output at each moment in the previous period and respectively predicting the data by using the three methods, and taking the method corresponding to the maximum total number as the optimal prediction method of the prediction period.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610642950.3A CN106251242B (en) | 2016-08-08 | 2016-08-08 | Wind power output interval combination prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610642950.3A CN106251242B (en) | 2016-08-08 | 2016-08-08 | Wind power output interval combination prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106251242A CN106251242A (en) | 2016-12-21 |
| CN106251242B true CN106251242B (en) | 2019-12-10 |
Family
ID=58079181
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610642950.3A Active CN106251242B (en) | 2016-08-08 | 2016-08-08 | Wind power output interval combination prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106251242B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106875033B (en) * | 2016-12-26 | 2020-06-02 | 华中科技大学 | Wind power cluster power prediction method based on dynamic self-adaption |
| CN109960818B (en) * | 2017-12-22 | 2023-06-20 | 北京金风慧能技术有限公司 | Method and device for generating simulated wind speed data of wind power plant |
| CN108446506B (en) * | 2018-03-26 | 2021-06-08 | 东北大学 | Uncertain system modeling method based on interval feedback neural network |
| CN108876060B (en) * | 2018-08-01 | 2021-05-11 | 国网吉林省电力有限公司长春供电公司 | Big data based prediction method for wind power output probability of sample collection |
| CN109449929B (en) * | 2018-11-22 | 2022-03-01 | 南方电网科学研究院有限责任公司 | Method and product for predicting and evaluating influence of distributed power supply on distribution network feeder utilization rate |
| CN112036607B (en) * | 2020-07-30 | 2023-07-25 | 南方电网科学研究院有限责任公司 | Wind power output fluctuation prediction method and device based on output level and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105552970A (en) * | 2016-02-25 | 2016-05-04 | 华北电力科学研究院有限责任公司 | Method and apparatus for improving accuracy of predicting power of wind power station |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8364739B2 (en) * | 2009-09-30 | 2013-01-29 | International Business Machines Corporation | Sparse matrix-vector multiplication on graphics processor units |
-
2016
- 2016-08-08 CN CN201610642950.3A patent/CN106251242B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105552970A (en) * | 2016-02-25 | 2016-05-04 | 华北电力科学研究院有限责任公司 | Method and apparatus for improving accuracy of predicting power of wind power station |
Non-Patent Citations (1)
| Title |
|---|
| 基于卡尔曼滤波的迟滞神经网络风速序列预测;李艳晴等;《北京科技大学学报》;20140831;第36卷(第8期);第1108-1114页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106251242A (en) | 2016-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106251242B (en) | Wind power output interval combination prediction method | |
| CN105574615B (en) | wavelet-BP neural network wind power prediction method based on spatial correlation and GA | |
| CN107194625B (en) | Wind power plant wind curtailment electric quantity evaluation method based on neural network | |
| CN107038292B (en) | Multi-wind-field output correlation modeling method based on self-adaptive multivariate nonparametric kernel density estimation | |
| CN112288164B (en) | Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast | |
| CN111626518A (en) | Urban daily water demand online prediction method based on deep learning neural network | |
| CN110555561A (en) | Medium-and-long-term runoff ensemble forecasting method | |
| CN113919231A (en) | PM2.5 concentration space-time change prediction method and system based on space-time diagram neural network | |
| CN110516840A (en) | Short term prediction method based on the wind light generation power output for improving random forest method | |
| Kolhe et al. | GA-ANN for short-term wind energy prediction | |
| CN110070228B (en) | BP neural network wind speed prediction method for neuron branch evolution | |
| CN110365053B (en) | Short-term wind power prediction method based on delay optimization strategy | |
| CN115661547B (en) | Knowledge-graph-based plant irrigation maintenance method and system | |
| CN112801332B (en) | Short-term wind speed prediction method based on gray level co-occurrence matrix | |
| CN111626473A (en) | Two-stage photovoltaic power prediction method considering error correction | |
| CN116956120A (en) | Prediction method for water quality non-stationary time sequence based on improved TFT model | |
| CN113379142A (en) | Short-term wind power prediction method based on wind speed correction and fusion model | |
| CN105956708A (en) | Grey correlation time sequence based short-term wind speed forecasting method | |
| CN115860797A (en) | Electric quantity demand prediction method suitable for new electricity price reform situation | |
| CN113984198B (en) | Shortwave radiation prediction method and system based on convolutional neural network | |
| CN111159640A (en) | Small rain emptying method, system, electronic equipment and storage medium suitable for grid forecast | |
| CN109388863B (en) | ARIMA model-based distributed photovoltaic output power prediction method | |
| CN107808209B (en) | Wind power plant abnormal data identification method based on weighted kNN distance | |
| CN116070669B (en) | Workshop energy consumption prediction method and management system based on improved deep belief network | |
| CN110765420A (en) | PSO-FI-based ground automatic meteorological station air temperature observation data quality control method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |