CN116306030A - New energy prediction dynamic scene generation method considering prediction error and fluctuation distribution - Google Patents
New energy prediction dynamic scene generation method considering prediction error and fluctuation distribution Download PDFInfo
- Publication number
- CN116306030A CN116306030A CN202310548730.4A CN202310548730A CN116306030A CN 116306030 A CN116306030 A CN 116306030A CN 202310548730 A CN202310548730 A CN 202310548730A CN 116306030 A CN116306030 A CN 116306030A
- Authority
- CN
- China
- Prior art keywords
- value
- prediction
- distribution
- probability distribution
- predicted value
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/06—Wind turbines or wind farms
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The application relates to a new energy prediction dynamic scene generation method considering prediction errors and fluctuation distribution. The method comprises the following steps: obtaining an actual measurement value and a predicted value of the new energy source, and performing per unit operation to obtain a per unit result; constructing a prediction box based on the obtained per unit result, and determining probability distribution of an actual measurement value according to the prediction box; and acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value. The method can better simulate the variation trend with the fluctuation and the randomness of wind power, has better optimization on the generated scene, and reduces the prediction error of the lead time predicted value.
Description
Technical Field
The application relates to the technical field of wind power, in particular to a new energy prediction dynamic scene generation method considering prediction errors and fluctuation distribution.
Background
At present, the academic world mainly has the following methods for generating wind power scenes: 1) Randomly generating a wind speed prediction error scene by using an ARMA model (Autoregressive moving average model, autoregressive moving average), and converting the wind speed scene into a wind power scene; 2) Hierarchical sampling from a probability distribution of wind power prediction errors using Latin hypercube sampling (Latin Hypercube Sampling, LHS) to generate a scene set; 3) Generating a large-scale scene Tree by adopting a scene Tree (Scenario Tree) method; 4) The nonparametric probabilistic prediction of power is converted into a large number of future scenarios.
When the wind power scene is generated by the method for prediction, the obtained prediction data has overestimation or underestimation, so that the method cannot provide more accurate prediction support for the system.
Disclosure of Invention
In view of the foregoing, it is necessary to provide a new energy prediction dynamic scene generation method that can accurately predict wind power fluctuation according to practical conditions, taking prediction errors and fluctuation distribution into consideration.
In a first aspect, the present application provides a new energy prediction dynamic scene generation method considering prediction error and fluctuation distribution. The method comprises the following steps:
obtaining an actual measurement value and a predicted value of the new energy source, and performing per unit operation to obtain a per unit result;
constructing a prediction box based on the obtained per unit result, and determining probability distribution of the actual measurement value according to the prediction box;
and acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
In one embodiment, the measured value is an actual output value, and the predicted value is a predicted value before the day corresponding to the measured value.
In one embodiment, the per-unit result includes a per-unit result corresponding to the predicted value; the constructing a prediction box based on the obtained per unit result, determining probability distribution of the measured value according to the prediction box, includes:
monotonically decreasing and arranging the predicted values, and equally dividing the arranged predicted values into a numerical value interval;
constructing a prediction box based on the per unit result corresponding to the predicted value in the numerical interval;
and carrying out distribution analysis on the actual measurement value according to the constructed prediction box, and estimating probability distribution of the actual measurement value.
In one embodiment, the analyzing the distribution of the measured value according to the constructed prediction box, estimating the probability distribution of the measured value includes:
setting a first random variable in a numerical value interval corresponding to the prediction box, and monotonically increasing and arranging measured values in the numerical value interval;
respectively comparing the first random variable with the arranged measured values to obtain a comparison result of the first random variable;
and taking the average value of the comparison result as the probability distribution of the actual measurement value.
In one embodiment, the samples are multivariate normal random vector samples; the obtaining the lead time predicted value and the sample, and generating the dynamic scene of the new energy according to the probability distribution of the lead time predicted value, the sample and the measured value comprises the following steps:
acquiring covariance key parameters according to a preset objective function and determining a covariance matrix, so as to obtain multi-element normal distribution;
determining a target prediction box in the prediction box based on the lead time predicted value, and obtaining a conditional probability distribution of the lead time predicted value according to a probability distribution of an actual measurement value corresponding to the target prediction box;
generating a plurality of normal random vector samples based on the obtained plurality of normal distributions;
and carrying out inverse transformation sampling operation on the multi-element normal random vector sample based on the conditional probability distribution to obtain a dynamic scene of the new energy.
In one embodiment, the determining a target prediction box in the prediction boxes based on the lead time predicted value, and obtaining the conditional probability distribution of the lead time predicted value according to the probability distribution of the measured value corresponding to the target prediction box includes:
determining a target prediction box corresponding to the lead time predicted value;
generating a second random variable compliant with a standard normal distribution;
sampling the second random variable to obtain an estimated value;
and obtaining a conditional probability distribution of the lead time predicted value based on the estimated value.
In one embodiment, the sampling the second random variable to obtain the estimated value includes:
determining a corresponding standard normal distribution function value according to the second random variable;
determining target cumulative empirical distribution data based on the standard normal distribution function value;
and obtaining an estimated value according to the target accumulated experience distribution data.
In a second aspect, the present application further provides a new energy prediction dynamic scene generating device considering prediction errors and fluctuation distribution. The device comprises:
the data acquisition module is used for acquiring the actual measurement value and the predicted value and carrying out per unit operation to obtain a per unit result;
the prediction box construction module is used for constructing a prediction box based on the obtained per unit result and determining probability distribution of the actual measurement value according to the prediction box;
the dynamic scene generation module is used for acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:
obtaining an actual measurement value and a predicted value of the new energy source, and performing per unit operation to obtain a per unit result;
constructing a prediction box based on the obtained per unit result, and determining probability distribution of the actual measurement value according to the prediction box;
and acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
obtaining an actual measurement value and a predicted value of the new energy source, and performing per unit operation to obtain a per unit result;
constructing a prediction box based on the obtained per unit result, and determining probability distribution of the actual measurement value according to the prediction box;
and acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
According to the new energy prediction dynamic scene generation method considering the prediction error and the fluctuation distribution, the prediction box is constructed through a large number of new energy actual measurement values and the prediction values, the probability distribution of the actual measurement values is determined based on the prediction box, and a large number of inputs are provided for the dynamic scene. On the basis, the wind power photovoltaic possible output of the lead time is estimated by combining the known lead time predicted value (namely the possible output of the wind power in the continuous time period provided by the prediction tool), so that a dynamic scene is generated, the new energy change trend with volatility and randomness is better simulated, the dynamic scene is better optimized, and the prediction error of the lead time predicted value is reduced.
Drawings
FIG. 1 is a flow chart of a method for generating a dynamic scene of a new energy in an embodiment;
FIG. 2 is a flow diagram of the steps for constructing a prediction box in one embodiment;
FIG. 3 (a) is a schematic diagram of a prediction box in one embodiment;
FIG. 3 (b) is a schematic diagram of another prediction box in one embodiment;
FIG. 4 is a flow diagram of determining a cumulative experience distribution in one embodiment;
FIG. 5 is a flow diagram of the steps for generating a dynamic scene in one embodiment;
FIG. 6 is a flow diagram of determining a conditional probability distribution in one embodiment;
FIG. 7 is a flow diagram of an inverse transform step in one embodiment;
FIG. 8 is a schematic diagram of inverse transform sampling in one embodiment;
FIG. 9 is a block diagram of a device for generating a dynamic scene of new energy in one embodiment;
fig. 10 is an internal structural view of a computer device in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
In one embodiment, as shown in fig. 1, a new energy prediction dynamic scene generation method considering prediction error and fluctuation distribution is provided, and this embodiment is illustrated by applying the method to a terminal, it can be understood that the method can also be applied to a server, and can also be applied to a system including the terminal and the server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the steps of:
s102, obtaining an actual measurement value and a predicted value of the new energy source, and carrying out per unit operation to obtain a per unit result.
In the per unit calculation, each physical quantity and parameter are expressed by the ratio of the named value to the reference value. The actual measured value is the actual output value of the new energy source, and the predicted value is the predicted value before the day corresponding to the actual output value.
Specifically, power history data is obtained from a database of the new energy power prediction system, the database comprises measured values and predicted values corresponding to the measured values, each measured value and the predicted value corresponding to the measured value are set as a data set in the database, and the resolution of the measured values and the predicted values is 15min. The data set is [ predicted value, measured value ].
When the database is used, all data sets are subjected to per unit operation to obtain per unit result, wherein the per unit result comprises the per unit result of the measured value and the per unit result of the predicted value, namely the measured value and the predicted value which are all changed between [0,1 ].
S104, constructing a prediction box based on the obtained per unit result, and determining probability distribution of the actual measurement value according to the prediction box.
And dividing the obtained per unit result into intervals, and setting all data sets in each divided interval as a prediction box respectively. And carrying out statistical analysis on the data sets of all the constructed prediction boxes, establishing a frequency histogram about the measured values in the prediction boxes, deducing probability distribution which is met by the measured values according to the orthometric statistical graph, and determining the probability distribution of the corresponding measured values.
S106, acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
Wherein the lead time refers to a time node on a predicted continuous time period, the lead time predicted value refers to a possible wind power output value of the lead time, and the lead time predicted value passesKnown wind power point prediction curve
The dynamic scene is determined to be the dynamic scene before the power day. The samples are samples of a plurality of normal random vectors, and the plurality of normal random vectors are applied to the process of generating dynamic scenes.
Specifically, a prediction tool is used for obtaining a predicted value of the lead time, and a prediction box corresponding to the predicted value of the lead time is judged. After the sample is obtained, carrying out inverse transformation operation on probability distribution of an actual measurement value in a prediction box corresponding to the sample and the advance time predicted value, and generating a dynamic scene of new energy corresponding to the sample.
In the new energy prediction dynamic scene generation method considering the prediction error and the fluctuation distribution, a plurality of prediction boxes are constructed through a data set consisting of a large amount of historical data, each prediction box is analyzed and processed to obtain a probability distribution function of an actual measurement value, and the probability distribution of the actual measurement value can be obtained through the probability distribution function of the actual measurement value. The wind power photovoltaic possible output of the lead time is estimated by combining the known lead time predicted value (namely the possible output of the wind power of the continuous time period provided by a prediction tool) on the basis, so that a dynamic scene is generated, the power fluctuation amplitude and range of the continuous time period can be accurately predicted, and decision support is provided for a circuit system.
In one embodiment, as shown in fig. 2, the per-unit result includes a per-unit result corresponding to the predicted value; constructing a prediction box based on the obtained per unit result, determining probability distribution of the measured value according to the prediction box, and comprising:
s202, the predicted values are arranged in a monotonically decreasing mode, and the arranged predicted values are equally divided into numerical value intervals.
The numerical value interval refers to the data length formed by equally dividing the predicted value, and the corresponding data set (predicted value, actual measured value) is placed in the corresponding numerical value interval according to the size of the predicted value.
Specifically, performing an equal division operation according to the per-unit predicted value to obtain lengths of a plurality of numerical intervals and the number of data groups in the data intervals, and dividing the corresponding data according to the predicted valueThe group divides the collection. In the implementation process, the predictive value in the adopted database can be equally divided into 50 pieces with the length of
Is a numerical interval of (a).
S204, constructing a prediction box based on the per unit result corresponding to the predictive value in the numerical interval.
The per unit result includes a per unit result corresponding to the predicted value.
Specifically, the plurality of numerical intervals divided in S202 are set one by one as the prediction box.
S206, carrying out distribution analysis on the measured value according to the constructed prediction box, and estimating probability distribution of the measured value.
Wherein the probability distribution is a mathematical representation describing the law of the random variable value distribution.
Specifically, the actual measurement value in each prediction box is analyzed in a theoretical distribution manner, and an analysis result is obtained. The analysis result is used for judging that the theoretical distribution cannot be simply inferred, so that the non-parametric empirical distribution is adopted for approximate estimation, namely the empirical distribution is used as the probability distribution of the measured value.
In this embodiment, a method of predicting error distribution by using statistical points of a prediction box is used to process a large amount of historical data. The design of the prediction box enables the prediction value in the prediction box to be very close, but the actual measurement value corresponding to the prediction box has larger error. Therefore, the probability distribution function of the measured value is obtained by analyzing the distribution of the measured value, the probability distribution of the possible wind power fluctuation of the wind power photovoltaic can be accurately estimated by determining the prediction box corresponding to the predicted value, and the error of the wind power fluctuation of the dynamic scene generated later is reduced.
In the actual parsing process, the approximate theoretical distribution of the power prediction error may be different for different prediction means and application objects, and even the application objects cannot be parsed. And carrying out statistical analysis on the measured value in the prediction box, and obtaining an analysis distribution function through measured value data. As shown in fig. 3 (a) and 3 (b), wherein fig. 3 (a) shows a frequency histogram of the measured values in the 10 th prediction box, and a normal distribution curve and a beta distribution curve formed based on the frequency fitting of the measured values; fig. 3 (b) shows a frequency histogram of the measured values in the 40 th prediction box, and a normal distribution curve and a beta distribution curve formed based on the frequency fitting of the measured values. By observing the fitting degree of the fitting curve and the frequency of the measured values in fig. 3 (a) and 3 (b), the measured values in the 10 th prediction box are more compliant with the beta distribution, while the measured values in the 40 th prediction box are more compliant with the normal distribution, so that the theoretical distribution compliant with all the prediction boxes cannot be inferred therefrom.
Since the probability distribution of actual measurement values is considered to provide input for generating dynamic scenes, the method is set
Is a theoretical distribution function of the random variable p +.>
As an empirical distribution function, it is analyzed according to the lattice Li Wenke theorem, which is shown below:
wherein, the liquid crystal display device comprises a liquid crystal display device,
the function is a line-up function>
The function is a joint probability distribution function.
According to the theorem, when the number of samples i is large,
will be very close to->
From this, the description can infer an ensemble from the sample. And as long as the sample values are sufficiently large, the empirical distribution function of the sample will approximate the theoretical distribution of the measured values. Based on the above deduction, non-ginseng is adopted in theoretical distribution analysis of measured valuesThe empirical distribution model of the numbers can approximate the theoretical distribution of the measured values.
In this embodiment, by performing statistical analysis on all the measured values in the prediction boxes and analyzing them, theoretical distribution suitable for all the measured values in the prediction boxes is deduced, so as to reduce prediction error and fluctuation distribution range for the estimation of the subsequent predicted values, and provide more accurate input for generating dynamic scenes.
In one embodiment, as shown in fig. 4, performing distribution analysis on the measured value according to the constructed prediction box, estimating a probability distribution of the measured value includes:
s402, setting a first random variable in a numerical value interval corresponding to the prediction box, and simultaneously, monotonously increasing and arranging actual measurement values in the numerical value interval.
Wherein the random variable p is a power random variable.
Specifically, a random variable p of the measured value is set in the prediction box.
S404, comparing the first random variable with the arranged measured values respectively to obtain a comparison result of the first random variable.
Specifically, samples of the random variable p are arranged according to monotonically increasing, and the random variable p is compared with the samples respectively to obtain comparison results, wherein the specific comparison results are as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,
is the ith sample in the monotonically increasing value interval.
S406, taking the average value of the comparison results as the probability distribution of the measured value.
Specifically, the result of comparing the random variable p with the sample in the prediction box is processed, and the analytical distribution function is determined to be the accumulated empirical distribution, specifically as follows:
where l is the number of samples in the interval of the number.
In this embodiment, based on analysis of the theorem Li Wenke, when the theoretical distribution of the measured value cannot be determined simply, the theoretical distribution of the measured value can be estimated approximately by using a non-parametric empirical distribution model, and the accuracy of probability distribution of the estimated predicted value can be improved by using the measured value in enough historical data.
In one embodiment, as shown in FIG. 5, the samples are multiple normal random vector samples; acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and an actual measurement value, wherein the method comprises the following steps:
s502, acquiring covariance key parameters according to a preset objective function and determining a covariance matrix, so as to obtain multi-element normal distribution.
Specifically, covariance key parameters are obtained through an objective function, and covariance matrixes are determined through the obtained covariance key parameters, so that compliance can be determined
Is the expected value and +.>
For the multivariate normal distribution of covariance matrices
。
S504, determining a target prediction box in the prediction box based on the lead time predicted value, and obtaining the conditional probability distribution of the lead time predicted value according to the probability distribution of the measured value corresponding to the target prediction box.
Specifically, on the basis of providing the lead time predicted value, it is possible to determine which prediction box the lead time predicted value belongs to, and then estimate the probability distribution at the lead time from the cumulative empirical distribution of the measured values in the corresponding prediction boxes.
S506, generating a multi-element normal random vector sample based on the obtained multi-element normal distribution.
The multiple normal random vector samples are the samples.
Specifically, the multivariate normal distribution determined according to step 502
D samples of a plurality of normal random vectors Z are generated by a normal distribution random number generator.
S508, carrying out inverse transformation sampling operation on the multi-element normal random vector samples based on conditional probability distribution to obtain a dynamic scene of the new energy.
Specifically, the inverse transformation operation is performed on the d-Z multi-element normal random vector samples, so that the multi-element normal distribution random numbers can be converted into d power dynamic scenes. Wherein, the power output value (the predicted value of the new energy) of the dynamic scene on any time section obeys the edge prediction error distribution (namely the conditional probability distribution) corresponding to the advanced time predicted value, and the cross-period power output value (the predicted value of the new energy) in the whole prediction time period obeys the joint probability distribution of the power prediction error.
In this embodiment, inverse transformation sampling is performed on each dynamic scene, so as to estimate the power output value (predicted value of new energy) of the lead time corresponding to a large number of normal random vector samples, determine that any time predicted value and cross-period predicted value within the whole predicted time conform to corresponding probability distribution, analyze and process the lead time predicted value of continuous time period and generate the dynamic scene according to the corresponding probability distribution, so that the obtained power possible output value has volatility, improve the prediction accuracy, and provide data support for the power system.
In one embodiment, as shown in fig. 6, determining a target prediction box in the prediction box based on the lead time prediction value, and obtaining a conditional probability distribution of the lead time prediction value according to a probability distribution of an actual measurement value corresponding to the target prediction box includes:
s602, determining a target prediction box corresponding to the lead time predicted value in the prediction box.
Specifically, the per-unit operation is performed on the lead time predicted value, a per-unit result of the lead time predicted value is obtained, and in which numerical interval corresponding to the predicted box the per-unit result of the lead time predicted value is determined, namely, the target predicted box is determined.
S604, generating a second random variable which obeys standard normal distribution.
In particular, in compliance with
Random variable +.>
When sampling, i.e. for the random variable +.>
Random sampling of cumulative distribution function values of (a) to generate a plurality of compliance +.>
Is a random number U of (c). And further, the random number U obeys the standard normal distribution, and then the standard normal distribution function value set of the random number U obeys [0,1]]Evenly distributed between them.
It is possible to obtain a random variable by generating a random variable compliant with a normal distribution of standards
And using the random variable->
The random number is replaced by a standard normal distribution function value. Wherein the random variable->
Is 1 and desirably 0.
S606, sampling the second random variable to obtain an estimated value.
Wherein the samples refer to inverse transformed samples.
Specifically, a random variable is determined
Then the random number of the known random variable is processedInverse transformation sampling operation, thereby obtaining the target prediction in-box and the random variable +.>
The wind power output value corresponding to the standard normal distribution function value, namely the estimated value.
S608, obtaining a conditional probability distribution of the lead time predicted value based on the estimated value.
Specifically, random variables will be utilized
The obtained estimated value is used as a generated scene, and a probability distribution corresponding to the estimated value, that is, a conditional probability distribution on the lead time predicted value is obtained.
In this embodiment, a large number of samples are generated by the random number following the standard normal distribution, and the inverse transformation operation is performed based on the generated samples to obtain an inverse transformation result, so that the random process of the new energy source is more easily represented by the joint probability distribution of the multiple normal distributions. On the other hand, the scene generated by the inverse transformation ensures that the sample values at the leading moment obey
Edge probability distribution of->
The cross-period correlation degree is converted into a joint probability distribution, so that the fluctuation amplitude and range of the lead time predicted value can be set in a reasonable interval, and the accuracy of the dynamic scene in estimating the power output value is improved.
In one embodiment, as shown in FIG. 7, sampling the second random variable to obtain an estimate includes:
s702, determining a corresponding standard normal distribution function value according to the second random variable.
Specifically, by a second random variable
The corresponding standard normal distribution function is calculated, and the specific calculation is as follows:
s704, determining target cumulative empirical distribution data based on the standard normal distribution function value.
In particular, in the random variable obtained
Is a standard normal distribution function value->
Sampling is performed on the basis of the standard normal distribution function value +.>
As the target cumulative empirical distribution value.
S706, an estimated value is obtained according to the target accumulated experience distribution data.
Specifically, the target cumulative empirical distribution value is substituted into the inverse function corresponding to the known cumulative empirical distribution function, and the lead time predicted value is calculated and obtained, specifically as follows:
wherein, the liquid crystal display device comprises a liquid crystal display device,
distribution of experiences for accumulation->
Is an inverse function of (c).
In the embodiment, the randomness of wind power is simulated by using an inverse transformation sampling mode, namely, an initial sample is generated by using random numbers conforming to normal distribution, so that the fluctuation of the wind power is better simulated, the calculated result is better conforming to the edge probability distribution of a predicted value, and the accuracy of prediction is improved.
In one embodiment, in performing the inverse transform sampling, it is assumed that a target is determinedQuasi-normally distributed random numbers
Starting from 1, by the corresponding random variable +.>
The standard normal distribution function of (2) can be calculated to obtain the standard normal distribution function +.>
0.8413, and->
The corresponding power output value is +.>
. The specific conversion process can be as shown in FIG. 8 when determining the random number +.>
At this time, the right graph in FIG. 8 corresponds to the determination of the standard normal distribution function +.>
The standard normal distribution function in the right graph is then +.>
Transfer to left graph, standard normal distribution function +.>
The function value of (2) is used as the accumulated empirical distribution value to further determine the power output value. Wherein the arrow direction in the figure is the sampling direction.
In the embodiment, a plurality of dynamic scenes can be generated through inverse transformation sampling, so that single scene precision errors are reduced, and the power output is better close to real power output.
It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a new energy prediction dynamic scene generation device for realizing the generation method of the new energy dynamic scene. The implementation scheme of the device for solving the problem is similar to the implementation scheme recorded in the method, so the specific limitation in the embodiment of the device for generating the new energy prediction dynamic scene taking the prediction error and the fluctuation distribution into consideration can be referred to the limitation of the method for generating the new energy dynamic scene, and the description is omitted herein.
In one embodiment, as shown in fig. 9, there is provided a new energy prediction dynamic scene generation apparatus considering prediction error and fluctuation distribution, including: a data acquisition module 902, a prediction box construction module 904, and a dynamic scene generation module 906, wherein:
the data obtaining module 902 is configured to obtain an actual measurement value and a predicted value of the new energy, and perform per unit operation to obtain a per unit result.
The prediction box construction module 904 is configured to construct a prediction box based on the obtained per unit result, and determine a probability distribution of the measured value according to the prediction box.
The dynamic scene generation module 906 is configured to obtain the lead time predicted value and the sample, and generate a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
In one embodiment, the prediction box construction module 904 further includes a probability distribution estimation module, configured to monotonically arrange the predicted values and equally divide the arranged predicted values into a number interval; constructing a prediction box based on per unit result corresponding to the predicted value in the numerical interval; and carrying out distribution analysis on the measured value according to the constructed prediction box, and estimating probability distribution of the measured value.
In one embodiment, the probability distribution estimation function module further includes a function analysis module, configured to set a first random variable in a numerical interval corresponding to the prediction box, and meanwhile, monotonically increasing and arranging actual measurement values in the numerical interval; respectively comparing the first random variable with the arranged measured values to obtain a comparison result of the first random variable; the average value of the comparison results is taken as the probability distribution of the measured value.
In one embodiment, the dynamic scene generating module 906 further includes a random vector generating module, configured to obtain a covariance key parameter according to a preset objective function and determine a covariance matrix, so as to obtain a multivariate normal distribution; determining a target prediction box in the prediction box based on the lead time predicted value, and obtaining a conditional probability distribution of the lead time predicted value according to a probability distribution of an actual measurement value corresponding to the target prediction box; generating a multi-element normal random vector sample based on the obtained multi-element normal distribution; and carrying out inverse transformation sampling operation on the multi-element normal random vector samples based on conditional probability distribution to obtain a dynamic scene of the new energy.
In one embodiment, the random vector generation module further includes a probability estimation module, configured to determine a target prediction box corresponding to the lead time prediction value in the prediction box; generating a second random variable compliant with a standard normal distribution; sampling the second random variable to obtain an estimated value; a conditional probability distribution of lead time predictors is obtained based on the estimates.
In one embodiment, the probability estimation module further comprises an inverse transformation module for determining a corresponding standard normal distribution function value from the second random variable; determining target cumulative empirical distribution data based on the standard normal distribution function values; an estimated value is obtained from the target cumulative empirical distribution data.
The above-described respective modules in the new energy prediction dynamic scene generation device considering the prediction error and the fluctuation distribution may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 10. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used for storing measured value and predicted value data. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements a new energy prediction dynamic scene generation method that takes into account prediction errors and fluctuation distribution.
It will be appreciated by those skilled in the art that the structure shown in fig. 10 is merely a block diagram of some of the structures associated with the present application and is not limiting of the computer device to which the present application may be applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
In an embodiment, there is also provided a computer device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the method embodiments described above when the computer program is executed.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, carries out the steps of the method embodiments described above.
It should be noted that, the user information (including, but not limited to, user equipment information, user personal information, etc.) and the data (including, but not limited to, data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data are required to comply with the related laws and regulations and standards of the related countries and regions.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can take many forms, such as static Random access memory (Static Random Access Memory, SRAM) or Dynamic Random access memory (Dynamic Random AccessMemory, DRAM), among others. The databases referred to in the various embodiments provided herein may include at least one of relational databases and non-relational databases. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.
Claims (10)
1. A new energy prediction dynamic scene generation method considering prediction error and fluctuation distribution, characterized in that the method comprises:
obtaining an actual measurement value and a predicted value of the new energy source, and performing per unit operation to obtain a per unit result;
constructing a prediction box based on the obtained per unit result, and determining probability distribution of the actual measurement value according to the prediction box;
and acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
2. The method of claim 1, wherein the measured value is an actual output value and the predicted value is a predicted value before day corresponding to the measured value.
3. The method according to claim 1, wherein the per-unit result includes a per-unit result corresponding to the predicted value; the constructing a prediction box based on the obtained per unit result, determining probability distribution of the measured value according to the prediction box, includes:
monotonically decreasing and arranging the predicted values, and equally dividing the arranged predicted values into a numerical value interval;
constructing a prediction box based on the per unit result corresponding to the predicted value in the numerical interval;
and carrying out distribution analysis on the actual measurement value according to the constructed prediction box, and estimating probability distribution of the actual measurement value.
4. A method according to claim 3, wherein said analyzing the distribution of the measured values from the constructed prediction box, estimating the probability distribution of the measured values, comprises:
setting a first random variable in a numerical value interval corresponding to the prediction box, and monotonically increasing and arranging measured values in the numerical value interval;
respectively comparing the first random variable with the arranged measured values to obtain a comparison result of the first random variable;
and taking the average value of the comparison result as the probability distribution of the actual measurement value.
5. The method of claim 1, wherein the samples are multivariate normal random vector samples; the obtaining the lead time predicted value and the sample, and generating the dynamic scene of the new energy according to the probability distribution of the lead time predicted value, the sample and the measured value comprises the following steps:
acquiring covariance key parameters according to a preset objective function and determining a covariance matrix, so as to obtain multi-element normal distribution;
determining a target prediction box in the prediction box based on the lead time predicted value, and obtaining a conditional probability distribution of the lead time predicted value according to a probability distribution of an actual measurement value corresponding to the target prediction box;
generating a plurality of normal random vector samples based on the obtained plurality of normal distributions;
and carrying out inverse transformation sampling operation on the multi-element normal random vector sample based on the conditional probability distribution to obtain a dynamic scene of the new energy.
6. The method of claim 5, wherein the determining a target prediction box in the prediction boxes based on the lead time prediction values, and obtaining the conditional probability distribution of the lead time prediction values from the probability distribution of the measured values corresponding to the target prediction box, comprises:
determining a target prediction box corresponding to the lead time predicted value in the prediction box;
generating a second random variable compliant with a standard normal distribution;
sampling the second random variable to obtain an estimated value;
and obtaining a conditional probability distribution of the lead time predicted value based on the estimated value.
7. The method of claim 6, wherein said sampling the second random variable to obtain an estimate comprises:
determining a corresponding standard normal distribution function value according to the second random variable;
determining target cumulative empirical distribution data based on the standard normal distribution function value;
and obtaining an estimated value according to the target accumulated experience distribution data.
8. A new energy prediction dynamic scene generation device considering prediction error and fluctuation distribution, characterized in that the device comprises:
the data acquisition module is used for acquiring the actual measurement value and the predicted value and carrying out per unit operation to obtain a per unit result;
the prediction box construction module is used for constructing a prediction box based on the obtained per unit result and determining probability distribution of the actual measurement value according to the prediction box;
the dynamic scene generation module is used for acquiring a lead time predicted value and a sample, and generating a dynamic scene of the new energy according to probability distribution of the lead time predicted value, the sample and the measured value.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310548730.4A CN116306030A (en) | 2023-05-16 | 2023-05-16 | New energy prediction dynamic scene generation method considering prediction error and fluctuation distribution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310548730.4A CN116306030A (en) | 2023-05-16 | 2023-05-16 | New energy prediction dynamic scene generation method considering prediction error and fluctuation distribution |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116306030A true CN116306030A (en) | 2023-06-23 |
Family
ID=86799882
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310548730.4A Pending CN116306030A (en) | 2023-05-16 | 2023-05-16 | New energy prediction dynamic scene generation method considering prediction error and fluctuation distribution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116306030A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116976532A (en) * | 2023-09-25 | 2023-10-31 | 南方电网数字电网研究院有限公司 | Non-parameterized new energy scene-probability-interval integrated prediction method |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110854932A (en) * | 2019-11-21 | 2020-02-28 | 国网山东省电力公司青岛供电公司 | Multi-time scale optimization scheduling method and system for AC/DC power distribution network |
| CN111738487A (en) * | 2020-05-14 | 2020-10-02 | 中国电力科学研究院有限公司 | Wind power prediction scene generation method and system |
| CN112653198A (en) * | 2020-12-17 | 2021-04-13 | 山东大学 | Wind power output scene generation method and system based on prediction box |
| CN113837444A (en) * | 2021-08-26 | 2021-12-24 | 国网河南省电力公司经济技术研究院 | Trading clearing optimization method for virtual power plant participating in multiple markets |
| CN114154328A (en) * | 2021-12-03 | 2022-03-08 | 国网新疆电力有限公司电力科学研究院 | Flexibility-improved two-stage random optimization scheduling method for electric heating comprehensive energy system |
| CN114513012A (en) * | 2020-11-16 | 2022-05-17 | 中国电力科学研究院有限公司 | New energy output scene construction method and system considering time-space correlation |
| CN115169089A (en) * | 2022-06-21 | 2022-10-11 | 广东电网有限责任公司电力调度控制中心 | Wind power probability prediction method and device based on kernel density estimation and copula |
| CN115392101A (en) * | 2021-05-24 | 2022-11-25 | 中国电力科学研究院有限公司 | Method and system for generating comprehensive energy random scene |
| CN116014749A (en) * | 2022-12-25 | 2023-04-25 | 武汉大学 | Power distribution network peak shaving random optimization method considering three-phase unbalance |
-
2023
- 2023-05-16 CN CN202310548730.4A patent/CN116306030A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110854932A (en) * | 2019-11-21 | 2020-02-28 | 国网山东省电力公司青岛供电公司 | Multi-time scale optimization scheduling method and system for AC/DC power distribution network |
| CN111738487A (en) * | 2020-05-14 | 2020-10-02 | 中国电力科学研究院有限公司 | Wind power prediction scene generation method and system |
| CN114513012A (en) * | 2020-11-16 | 2022-05-17 | 中国电力科学研究院有限公司 | New energy output scene construction method and system considering time-space correlation |
| CN112653198A (en) * | 2020-12-17 | 2021-04-13 | 山东大学 | Wind power output scene generation method and system based on prediction box |
| CN115392101A (en) * | 2021-05-24 | 2022-11-25 | 中国电力科学研究院有限公司 | Method and system for generating comprehensive energy random scene |
| CN113837444A (en) * | 2021-08-26 | 2021-12-24 | 国网河南省电力公司经济技术研究院 | Trading clearing optimization method for virtual power plant participating in multiple markets |
| CN114154328A (en) * | 2021-12-03 | 2022-03-08 | 国网新疆电力有限公司电力科学研究院 | Flexibility-improved two-stage random optimization scheduling method for electric heating comprehensive energy system |
| CN115169089A (en) * | 2022-06-21 | 2022-10-11 | 广东电网有限责任公司电力调度控制中心 | Wind power probability prediction method and device based on kernel density estimation and copula |
| CN116014749A (en) * | 2022-12-25 | 2023-04-25 | 武汉大学 | Power distribution network peak shaving random optimization method considering three-phase unbalance |
Non-Patent Citations (1)
| Title |
|---|
| 王小飞;胡志坚;仉梦林;胡美玉;: "短期风功率场景质量的评估框架构建研究", 电网技术, no. 05, pages 1583 - 1587 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116976532A (en) * | 2023-09-25 | 2023-10-31 | 南方电网数字电网研究院有限公司 | Non-parameterized new energy scene-probability-interval integrated prediction method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Castruccio et al. | High-order composite likelihood inference for max-stable distributions and processes | |
| CN111709533A (en) | Distributed training method and device of machine learning model and computer equipment | |
| WO2017076154A1 (en) | Method and apparatus for predicting network event and establishing network event prediction model | |
| CN112068003B (en) | Method and device for predicting service life of cadmium-nickel storage battery based on linear wiener process | |
| US20180189655A1 (en) | Data meta-scaling apparatus and method for continuous learning | |
| CN110858062B (en) | Target optimization parameter obtaining method and model training method and device | |
| Kachour et al. | First‐order rounded integer‐valued autoregressive (RINAR (1)) process | |
| CN116306030A (en) | New energy prediction dynamic scene generation method considering prediction error and fluctuation distribution | |
| Hong et al. | Pathwise estimation of probability sensitivities through terminating or steady-state simulations | |
| CN110955685A (en) | Big data base estimation method, system, server and storage medium | |
| CN110807508B (en) | Bus peak load prediction method considering complex weather influence | |
| US20140324760A1 (en) | Synthetic time series data generation | |
| CN116154768B (en) | Power interval prediction method adopting point prediction error empirical distribution inverse transformation | |
| CN116450486B (en) | Modeling method, device, equipment and medium for nodes in multi-element heterogeneous computing system | |
| CN116191398A (en) | Load prediction method, load prediction device, computer equipment and storage medium | |
| CN115564152A (en) | Carbon emission prediction method and device based on STIRPAT model | |
| CN111476408B (en) | Power communication equipment state prediction method and system | |
| CN109766181A (en) | A kind of RMS schedulability determination method and device based on deep learning | |
| CN114117349A (en) | Power system random variable probability analysis method, system, equipment and storage medium | |
| CN114266414A (en) | Loan amount prediction method, loan amount prediction device, loan amount prediction electronic device, and loan amount prediction medium | |
| Pan et al. | Hierarchical resampling algorithm and architecture for distributed particle filters | |
| CN112085926A (en) | River water pollution early warning method and system | |
| Wang et al. | Bayesian analysis for NHPP-based software fault detection and correction processes | |
| JP2015001954A (en) | Information processing system and information processing method | |
| CN116976532A (en) | Non-parameterized new energy scene-probability-interval integrated prediction method |
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 |