CN114662393A - Photovoltaic power prediction error compensation method based on weather conditions - Google Patents

Photovoltaic power prediction error compensation method based on weather conditions Download PDF

Info

Publication number
CN114662393A
CN114662393A CN202210307054.7A CN202210307054A CN114662393A CN 114662393 A CN114662393 A CN 114662393A CN 202210307054 A CN202210307054 A CN 202210307054A CN 114662393 A CN114662393 A CN 114662393A
Authority
CN
China
Prior art keywords
photovoltaic
output power
prediction error
longicorn
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.)
Granted
Application number
CN202210307054.7A
Other languages
Chinese (zh)
Other versions
CN114662393B (en
Inventor
赵振兴
张普
宁勇
刘增
朱积嘉
彭子舜
张福家
陈颖
戴瑜兴
杨亚超
王俊
潘文武
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hunan Institute of Engineering
Original Assignee
Hunan Institute of Engineering
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hunan Institute of Engineering filed Critical Hunan Institute of Engineering
Priority to CN202210307054.7A priority Critical patent/CN114662393B/en
Publication of CN114662393A publication Critical patent/CN114662393A/en
Application granted granted Critical
Publication of CN114662393B publication Critical patent/CN114662393B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Water Supply & Treatment (AREA)
  • Tourism & Hospitality (AREA)
  • Public Health (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Artificial Intelligence (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Photovoltaic Devices (AREA)

Abstract

The invention provides a weather condition-based photovoltaic power prediction error compensation method, and relates to the technical field of photovoltaic power generation and energy storage control. The method comprises the steps of classifying photovoltaic data in preset time according to different weather conditions, obtaining photovoltaic prediction error data under various weather conditions by using an LSSVM prediction algorithm, establishing prediction error probability models under various weather conditions by using Gaussian Coipla functions, taking the output of a super capacitor as a target function, taking the prediction error as a random variable, solving the models by using a Tianniu-firefly algorithm to obtain the optimal value of the output power of the super capacitor, and compensating the actual output power of the photovoltaic. The invention realizes the compensation of the photovoltaic prediction error, reduces the difference between the predicted value and the actual value of the photovoltaic output power, and leads the actual output power to be closer to the predicted output power.

Description

Photovoltaic power prediction error compensation method based on weather conditions
Technical Field
The invention relates to the technical field of photovoltaic power generation and energy storage control, in particular to a photovoltaic power prediction error compensation method based on weather conditions.
Background
In recent years, with the large-scale use of new energy, the number of new photovoltaic installation machines is increased dramatically year by year, the permeability of photovoltaic power generation in the whole power grid system is higher and higher, and a photovoltaic prediction algorithm is also generally applied to a photovoltaic power generation system. The national energy agency northwest regulatory agency and power plant grid-connected operation management implementation fine rules stipulate that the single-point deviation (the actual prediction error allowable range) of the photovoltaic power daily prediction curve in 1min time scale prediction does not exceed +/-20% of the deviation, and the power exceeding the prediction error and the power fluctuation range is not only penalized, but also affects the safe and stable operation of the power grid.
Because the output power of photovoltaic power generation is greatly influenced by temperature and cloud layers, an accurate prediction model is difficult to obtain by the existing prediction algorithm, the obtained photovoltaic planned output is an inaccurate output model, and when the photovoltaic planned output is connected with power generation systems with different energies together, the power shortage of a user end can be caused, even under the condition of large error, the power grid is broken down, and great damage is caused to some equipment of the power grid system. Therefore, an accurate photovoltaic plan output is obtained, and the method is very beneficial to realizing the economic dispatching of the power grid.
At present, a photovoltaic prediction error compensation method mainly includes the steps of predicting meteorological conditions such as cloud layers, temperature and irradiance, establishing a meteorological prediction model, and correcting the photovoltaic power prediction model through the established meteorological prediction model.
Chinese patent application document CN 110598896a discloses a photovoltaic power prediction method based on prediction error correction, first using irradiance as a similar variable, and screening to obtain a similar day; and then calculating a weighting coefficient of each similar day by using the similar sunlight photovoltaic power generation prediction error data, and correcting a predicted value by using the photovoltaic power prediction error expectation of the adjacent day.
Chinese patent application document CN 111626473a discloses a two-stage photovoltaic power prediction method considering error correction, which first adopts a principal component analysis method to perform non-correlation processing on weather factor data after standardization processing; then, a regression analysis method is adopted to construct a differential preliminary prediction model of different meteorological types in each time period; and finally, establishing error correction models corresponding to different meteorological types according to the distribution characteristics of the preliminary prediction errors, and realizing error correction of the preliminary prediction results.
Chinese patent application document CN 106372749A discloses an ultra-short-term photovoltaic power prediction method based on cloud change analysis, which predicts the meteorological conditions above a photovoltaic power station, and applies the prediction result as a correction parameter to a photovoltaic power generation model to predict the photovoltaic output power.
The methods proposed above all have the following disadvantages:
(1) accurate cloud layer and irradiance data are required to be obtained, otherwise errors are brought again;
(2) the hardware cost required for obtaining the cloud layer and irradiance is high.
Disclosure of Invention
Aiming at the problem of the difference between the predicted value and the actual value of the existing photovoltaic power, the invention provides a photovoltaic power prediction error compensation method based on weather conditions, which establishes a photovoltaic prediction error probability model and provides a method for compensating the photovoltaic actual output value by using a super capacitor, so that the photovoltaic actual output value is close to the photovoltaic planned output value, and the difference between the actual value and the predicted value is reduced. According to the photovoltaic power prediction method, accurate data of the cloud layer and the irradiance are not required to be obtained, photovoltaic power prediction error compensation is achieved, a power grid can obtain more accurate and timely photovoltaic power prediction information, and management can be carried out in a more effective mode.
The invention provides a photovoltaic power prediction error compensation method based on weather conditions, which comprises the following steps:
s100: collecting photovoltaic output power data at equal intervals from historical data in a preset time period of a photovoltaic power station as sample data;
s200: removing nighttime data from the sample data, dividing the data according to weather conditions, and carrying out normalization processing on the data;
s300: constructing an LSSVM prediction model, predicting photovoltaic output power under different weather conditions, and comparing the photovoltaic output power with actual output power to obtain an output power prediction error under the weather conditions;
s400: respectively establishing a prediction error model for the photovoltaic output power prediction error under each weather condition by using a Gaussian Coipla function;
s500: predicting the photovoltaic output power at the next moment by using the constructed LSSVM prediction model to obtain a photovoltaic output power predicted value PprePredicting the photovoltaic output power PpreSubstituting the prediction error model into the established prediction error model to obtain a predicted value P of the known photovoltaic output powerpreConditional probability distribution of photovoltaic output power prediction errors under conditions;
s600: compensating a difference value between the photovoltaic predicted value and an actual output value of the photovoltaic output power by using the super capacitor to construct a super capacitor constraint condition;
s700: carrying out deterministic transformation on the opportunity constraint and constructing an opportunity constraint condition;
s800: output power P of super capacitorcAs an objective function, the prediction error delta is used as a random variable, a longicorn-firefly search algorithm is adopted to solve a prediction error model, and when the super-capacitor constraint condition and the opportunity constraint condition are met at the same time, the optimal value of the output power of the super-capacitor is obtained;
s900: and compensating the photovoltaic predicted output power by adopting the optimal value of the output power of the super capacitor to obtain the compensated photovoltaic actual output power.
Preferably, the step S400 includes the steps of:
s401: recording the historical output actual value as alphai.tAnd historical predicted values are recorded as betai.tAnd historical output actual value alpha of weather conditioni.tAnd historical predicted value betai.tRespectively counting to obtain the edge distribution function F of the actual photovoltaic outputɑtAnd edge distribution function F of photovoltaic prediction outputβt；
Wherein, the first and the second end of the pipe are connected with each other,
i belongs to [ 1.,. n ] represents the weather condition, and n is the type number of the weather condition;
t belongs to [1, T ] to represent the period;
t is the total time interval of the historical data;
s402: solving historical output actual value alphai.tAnd historical predicted value betai.tThe Kendall correlation coefficient of (1);
s403: and respectively establishing prediction error models under different weather conditions by adopting Gaussian Coipla functions according to the Kendall correlation coefficients.
According to Kendall correlation coefficients, historical output actual values alpha under different weather conditions can be knowni.tAnd historical predicted value betai.tThe difference in (b) is significant, so when building a prediction error model using the Coupla theory, the influence of different weather on the prediction error distribution must be considered.
Preferably, the conditional probability density function of the photovoltaic output power prediction error in S500 is:
Figure BDA0003565902880000031
wherein the content of the first and second substances,
delta is the prediction error;
fXY(x,Ppre) Is a joint probability density function;
fY(Ppre) To predict point PpreThe edge distribution probability density of (1);
c(FX(δ+Ppre),FX(Ppre) Is a coupli density function;
fX(δ+Ppre) The edge distribution probability density of the actual output power value;
x is the actual output value of the photovoltaic output power;
Ppreis a predicted value of the photovoltaic output power.
As can be seen from the formula, the conditional probability density function of the prediction error can be converted into an edge distribution probability density function f of the actual output power valueX(δ+Ppre) And the coupul density function are modeled separately.
Preferably, constructing the opportunity constraint in S700 includes:
output power P of super capacitorcAs an objective function, the prediction error delta is used as a random variable, and an opportunity constraint condition is constructed as follows:
P{f(Pc,δ)≤σ,t＝1,2,...}≥θ
Figure BDA0003565902880000041
wherein the content of the first and second substances,
sigma is a given allowable range of error;
θ is the given confidence;
Figure BDA0003565902880000042
output power for the super capacitor at t;
δtis the prediction error at t;
Figure BDA0003565902880000043
is the predicted value of the photovoltaic output power at t.
Preferably, step S800 includes the steps of:
s801: adopting a longicorn-firefly search algorithm, changing the number of the longicorn in the BAS algorithm from one longicorn to a plurality of longicorn, and adopting Cubic mapping to replace pseudo-random number generation in the original initialization to initialize the position of the longicorn;
s802: adding odor factors to the longicorn, and setting the mutual attraction degree beta (r), wherein the objective function is the maximum odor source;
s803: adopting an optimization objective function F as the fitness of each longicorn in the population;
s804: updating the position of the population individual;
s805: the longicorn with the largest smell updates the position of the longicorn according to the following formula;
Figure BDA0003565902880000044
s806: obtaining an optimal solution of the objective function to obtain an optimal value of the output power of the super capacitor;
wherein the content of the first and second substances,
Figure BDA0003565902880000045
Figure BDA0003565902880000046
is the initial direction of the longicorn;
s is a spatial dimension;
rands () is a random function;
t is the current moment, and t +1 is the next moment;
Figure BDA0003565902880000047
the fitness of the right tendrils of the longicorn at the time t;
Figure BDA0003565902880000048
the adaptability of the left tendril of the longicorn at the time t;
Figure BDA0003565902880000051
the position of the ith longicorn at the time t in the s-dimensional space;
αtis the step size at the iteration number t.
Preferably, in step S802, the mutual attraction β (r) between longicorn in the longicorn-firefly search algorithm is as follows:
Figure BDA0003565902880000052
wherein the content of the first and second substances,
β0is the initial attraction degree;
r is the distance between two longicorn;
gamma is the odor absorption coefficient; the odor gradually diminishes with increasing distance.
Preferably, in step S803, an optimization objective function F is used as the fitness of each longicorn in the population, and the formula is as follows:
Fit(xis)＝F
wherein the content of the first and second substances,
Fit(xis) Representing the fitness value of the ith longicorn in the current spatial position;
xisrepresenting the current spatial position of the ith longicorn in the s-dimensional space;
i represents the ith longicorn.
Preferably, in step S804, when the t-th iteration is performed, the position of the ith longicorn in the S-dimensional space is updated according to the following formula:
Figure BDA0003565902880000053
wherein the content of the first and second substances,
Figure BDA0003565902880000054
in the form of an incremental function of the function,
Figure BDA0003565902880000055
the expression is as follows:
Figure BDA0003565902880000056
wherein Xij＝Xi-Xj，XiAnd XjRepresenting the spatial positions of two longicorn i and j;
sign () is a sign function;
Figure BDA0003565902880000057
and
Figure BDA0003565902880000058
the fitness of the left and right longicorn whiskers is respectively.
Preferably, S806 obtains an optimal solution of the objective function to obtain the output power P of the super capacitorcThe optimal values include:
the accuracy of the prediction result measured by the cosine value of the included angle between the predicted output power value and the actual output power value is shown as the formula:
Figure BDA0003565902880000061
wherein the content of the first and second substances,
Figure BDA0003565902880000062
has a value range of [0,1]](ii) a When in use
Figure BDA0003565902880000063
Is shown byThe predicted power output curve is most similar to the actual power output curve;
Figure BDA0003565902880000064
predicting an n-dimensional vector consisting of output power values for the photovoltaic cells;
Figure BDA0003565902880000065
and the vector is an n-dimensional vector formed by photovoltaic actual output power values.
Firstly, the super capacitor is output with power PcTaking an upper limit value M, when M meets the super capacitor constraint condition and the opportunity constraint condition, taking M as the output power of the super capacitor at the time t, and when M does not meet the super capacitor constraint condition and the opportunity constraint condition, taking M as the output power P of the super capacitorcAnd (6) re-valuing. When in use
Figure BDA0003565902880000066
At maximum, the super capacitor outputs power PcThe value of (d) is the optimal output.
Therefore, in the invention, the change of the actual photovoltaic output power value is realized through the charging and discharging of the super capacitor, thereby changing
Figure BDA0003565902880000067
And the photovoltaic actual output power value is close to the predicted output power value.
Preferably, the weather conditions include sunny days, cloudy days, light rain and heavy rain.
Preferably, the data is linearly transformed using min-max normalization in S200, with the resulting values mapped between [0,1], and the transfer function is as follows:
Figure BDA0003565902880000068
wherein the content of the first and second substances,
xminis the minimum value of sample data;
xmaxis sample dataMaximum value of (d);
x*is the normalized value.
Preferably, the super-capacitor constraint condition constructed in S600 is:
SOCmin≤SOC≤SOCmax
when SOC is reachedmin≤SOC≤SOCmaxIn time, the super capacitor can be charged and also can be discharged;
when SOC is less than or equal to SOCminIn time, the capacitor can only be charged;
when SOC is more than or equal to SOCmaxIn time, the capacitor can only discharge;
wherein the content of the first and second substances,
SOCminand SOCmaxRespectively the minimum value and the maximum value of the super capacitor SOC.
Preferably, in step S900, the compensated actual photovoltaic output power is Po＝Ph+Pc；
Wherein the content of the first and second substances,
Pothe compensated photovoltaic actual output power is obtained;
Phthe actual output power of the uncompensated photovoltaic is obtained;
Pcand outputting power for the super capacitor.
Compared with the prior art, the invention has the following beneficial effects:
(1) the invention establishes a model for compensating photovoltaic output power prediction errors by using a super capacitor, wherein the model is divided into two parts, namely a prediction error probability distribution model and a super capacitor compensation error model; the model makes full use of the characteristics of energy storage, effectively reduces the difference between the predicted value and the actual value of the photovoltaic power, and meets the requirement on the accuracy of the photovoltaic design output power value.
(2) The adopted longicorn-firefly search algorithm improves the BAS algorithm, changes a single longicorn into a plurality of longicorn, overcomes the problem that a single optimizing individual is easy to fall into a local optimal solution, simultaneously introduces odor factors, accelerates the iteration process, can quickly obtain the optimal solution, and meets the requirements of the prediction error model solution on the rapidity and the accuracy of the algorithm.
(3) The adopted longicorn-firefly search algorithm adds an optimization mechanism of the firefly algorithm, accelerates the iterative process, and realizes the requirements of the prediction error model solution on the rapidity and the accuracy of the algorithm.
Drawings
FIG. 1 is a flow chart of a photovoltaic power prediction error compensation method of one embodiment of the present invention;
FIG. 2 is a flow chart of a longicorn-firefly algorithm according to an embodiment of the present invention;
FIG. 3 is a graph comparing a predicted photovoltaic output power value, an actual uncompensated photovoltaic output power value, and an actual compensated photovoltaic output power value, in accordance with an embodiment of the present invention.
Detailed Description
The following describes in detail embodiments of the present invention.
The invention provides a photovoltaic power prediction error compensation method based on weather conditions, which comprises the following steps:
s100: collecting photovoltaic output power data at equal intervals from historical data in a preset time period of a photovoltaic power station as sample data;
s200: removing nighttime data from the sample data, dividing the data according to weather conditions, and carrying out normalization processing on the data; the photovoltaic power generation has the characteristic of day and night alternation, the power generation work can be carried out only under enough illumination intensity, and the photovoltaic power station can not carry out the power generation work at night.
S300: constructing an LSSVM prediction model, predicting photovoltaic output power under different weather conditions, and comparing the photovoltaic output power with actual output power to obtain an output power prediction error under the weather conditions;
s400: respectively establishing a prediction error model for the photovoltaic output power prediction error under each weather condition by using a Gaussian Coipla function;
s500: predicting the photovoltaic output power at the next moment by using the constructed LSSVM prediction model to obtain a photovoltaic output power predicted value PprePredicting the photovoltaic output power PpreSubstituting the prediction error model into the established prediction error model to obtain a predicted value P of the known photovoltaic output powerpreConditional probability distribution of photovoltaic output power prediction errors under conditions;
s600: compensating a difference value between the photovoltaic predicted value and an actual output value of the photovoltaic output power by using the super capacitor to construct a super capacitor constraint condition;
s700: carrying out deterministic transformation on the opportunity constraint and constructing an opportunity constraint condition;
s800: output power P of super capacitorcAs a target function, taking the prediction error delta as a random variable, solving a prediction error model by adopting a longicorn-firefly search algorithm, and obtaining an output power optimal value of the super capacitor when a super capacitor constraint condition and an opportunity constraint condition are simultaneously met;
s900: and compensating the photovoltaic predicted output power by adopting the optimal value of the output power of the super capacitor to obtain the compensated photovoltaic actual output power.
According to a specific embodiment of the present invention, the step S400 includes the steps of:
s401: recording the historical output actual value as alphai.tAnd historical predicted values are recorded as betai.tAnd outputting the actual value alpha to the history of the weather conditioni.tAnd historical predicted value betai.tRespectively counting to obtain the edge distribution function F of the actual photovoltaic outputɑtAnd edge distribution function F of photovoltaic prediction outputβt；
Wherein, the first and the second end of the pipe are connected with each other,
i belongs to [ 1.,. n ] represents the weather condition, and n is the type number of the weather condition;
t belongs to [1, T ] to represent the period;
t is the total time interval of the historical data;
s402: solving historical output actual value alphai.tAnd historical predicted value betai.tThe Kendall correlation coefficient of (1);
s403: and respectively establishing prediction error models under different weather conditions by adopting Gaussian Coipla functions according to the Kendall correlation coefficients.
The Kendall correlation coefficient can be used to obtainAnd historical output actual value alpha under different weather conditionsi.tAnd historical predicted value betai.tThe difference in (b) is significant, so when building a prediction error model using the Coupla theory, the influence of different weather on the prediction error distribution must be considered.
According to a specific embodiment of the present invention, the two-dimensional probability density function of Gaussian Copula is:
Figure BDA0003565902880000091
in the formula:
(k, w) is the independent variable of the two-dimensional Copula probability density function, and the domain is [0,1]]2；
Figure BDA0003565902880000092
An inverse function representing a standard normal distribution function;
ρnand the correlation coefficient parameter of the Copula is expressed, and can be directly estimated through Kendall rank correlation coefficients between the photovoltaic output power actual value sequence and the predicted value sequence.
The Kendall rank correlation coefficient τ is a measure of the consistency of random variables and describes the ordered correlation information between two random variables. Since the linear correlation coefficient is easily affected by the edge distribution of the variables, in contrast, analyzing the consistency between the variables (i.e. replacing the correlation coefficient with the rank correlation coefficient) can describe the correlation between the variables more exactly.
After the Kendall rank correlation coefficient τ is obtained, the correlation coefficient parameter ρ of Gaussian Copula can be obtained by the following formulanAnd (3) estimating:
ρn＝sin(πτ/2)
will rhonAnd substituting the probability density function formula of Gaussian Copula into the probability density function formula of Gaussian Copula to obtain the conditional probability density functions of the prediction errors under different weathers.
According to a specific embodiment of the present invention, the conditional probability density function of the photovoltaic output power prediction error in S500 is:
Figure BDA0003565902880000093
wherein the content of the first and second substances,
delta is the prediction error;
fXY(x,Ppre) Is a joint probability density function;
fY(Ppre) To predict point PpreThe edge distribution probability density of (1);
c(FX(δ+Ppre),FX(Ppre) Is a coupli density function;
fX(δ+Ppre) The edge distribution probability density of the actual output power value;
x is the actual output value of the photovoltaic output power;
Ppreis a predicted value of the photovoltaic output power.
As can be seen from the formula, the conditional probability density function of the prediction error can be converted into an edge distribution probability density function f of the actual output power valueX(δ+Ppre) And the coupul density function are modeled separately.
According to a specific embodiment of the present invention, constructing the opportunity constraint in S700 includes:
output power P to the super capacitorcAs an objective function, the prediction error delta is used as a random variable, and an opportunity constraint condition is constructed as follows:
P{f(Pc,δ)≤σ,t＝1,2,...}≥θ
Figure BDA0003565902880000101
wherein, the first and the second end of the pipe are connected with each other,
sigma is a given allowable range of error;
θ is the given confidence;
Figure BDA0003565902880000102
output power for the super capacitor at t;
δtto predict the error at t;
Figure BDA0003565902880000103
is the predicted value of the photovoltaic output power at t.
According to a specific embodiment of the present invention, step S800 includes the steps of:
s801: adopting a longicorn-firefly search algorithm, changing the number of the longicorn in the BAS algorithm from one longicorn to a plurality of longicorn, and adopting Cubic mapping to replace pseudo-random number generation in the original initialization to initialize the position of the longicorn;
s802: adding odor factors to the longicorn mutually, and setting a mutual attraction degree beta (r), wherein an objective function is the maximum odor source;
s803: adopting an optimization objective function F as the fitness of each longicorn in the population;
s804: updating the position of the population individual;
s805: the longicorn with the largest smell updates the position of the longicorn according to the following formula;
Figure BDA0003565902880000104
s806: obtaining an optimal solution of the objective function to obtain an optimal value of the output power of the super capacitor;
wherein the content of the first and second substances,
Figure BDA0003565902880000105
Figure BDA0003565902880000111
is the initial direction of the longicorn;
s is a spatial dimension;
rands () is a random function;
t is the current moment, and t +1 is the next moment;
Figure BDA0003565902880000112
the fitness of the right tendrils of the longicorn at the time t;
Figure BDA0003565902880000113
the adaptability of the left beard of the longicorn at the time t;
Figure BDA0003565902880000114
the position of the ith longicorn at the time t in the s-dimensional space;
αtis the step size at the iteration number t.
According to an embodiment of the present invention, in step S802, the mutual attraction β (r) between longicorn in the longicorn-firefly search algorithm is as follows:
Figure BDA0003565902880000115
wherein, the first and the second end of the pipe are connected with each other,
β0is the initial attraction degree;
r is the distance between two longicorn;
gamma is the odor absorption coefficient; the odor gradually diminishes with increasing distance.
According to a specific embodiment of the present invention, in step S803, an optimization objective function F is used as the fitness of each longicorn in the population, and the formula is as follows:
Fit(xis)＝F
wherein, the first and the second end of the pipe are connected with each other,
Fit(xis) Representing the fitness value of the ith longicorn in the current spatial position;
xisrepresenting the current spatial position of the ith longicorn in the s-dimensional space;
i represents the ith longicorn.
According to an embodiment of the present invention, in step S804, when the t-th iteration is performed, the position of the ith longicorn in the S-dimensional space is updated according to the following formula:
Figure BDA0003565902880000116
wherein the content of the first and second substances,
Figure BDA0003565902880000121
in the form of an incremental function of the function,
Figure BDA0003565902880000122
the expression is as follows:
Figure BDA0003565902880000123
wherein Xij＝Xi-Xj，XiAnd XjRepresenting the spatial positions of two longicorn i and j;
sign () is a sign function;
Figure BDA0003565902880000124
and
Figure BDA0003565902880000125
the fitness of the left and right longicorn whiskers is respectively.
According to an embodiment of the present invention, S806 obtains an optimal solution of the objective function to obtain the output power P of the super capacitorcThe optimal values include:
the accuracy of the prediction result measured by the cosine value of the included angle between the predicted output power value and the actual output power value is shown as the formula:
Figure BDA0003565902880000126
wherein the content of the first and second substances,
Figure BDA0003565902880000127
has a value range of [0,1]](ii) a When in use
Figure BDA0003565902880000128
When the predicted power output curve is most similar to the actual power output curve, the predicted power output curve is represented;
Figure BDA0003565902880000129
predicting an n-dimensional vector consisting of output power values for the photovoltaic cells;
Figure BDA00035659028800001210
and the vector is an n-dimensional vector formed by photovoltaic actual output power values.
Firstly, the super capacitor is output with power PcTaking an upper limit value M, when M meets the super capacitor constraint condition and the opportunity constraint condition, taking M as the output power of the super capacitor at the time t, and when M does not meet the super capacitor constraint condition and the opportunity constraint condition, taking M as the output power P of the super capacitorcAnd (6) re-valuing. When in use
Figure BDA00035659028800001211
At maximum, the super capacitor outputs power PcThe value of (d) is the optimal output.
Therefore, in the invention, the change of the actual photovoltaic output power value is realized through the charging and discharging of the super capacitor, thereby changing
Figure BDA00035659028800001212
And the photovoltaic actual output power value is close to the predicted output power value.
According to a particular embodiment of the invention, the weather conditions include sunny, cloudy, light rain and heavy rain.
According to a specific embodiment of the present invention, the data is linearly transformed using min-max normalization in S200, so that the resulting values are mapped between [0,1], and the transfer function is as follows:
Figure BDA00035659028800001213
wherein the content of the first and second substances,
xminis the minimum value of the sample data;
xmaxis the maximum value of the sample data;
x*is the normalized value.
According to a specific embodiment of the present invention, the super capacitor constraint condition constructed in S600 is:
SOCmin≤SOC≤SOCmax
when SOC is reachedmin≤SOC≤SOCmaxIn time, the super capacitor can be charged and also can be discharged;
when SOC is less than or equal to SOCminWhile, the capacitor can only be charged;
when SOC is more than or equal to SOCmaxIn time, the capacitor can only discharge;
wherein the content of the first and second substances,
SOCminand SOCmaxRespectively the minimum value and the maximum value of the super capacitor SOC.
According to an embodiment of the present invention, in step S900, the compensated actual photovoltaic output power is Po＝Ph+Pc。
Wherein the content of the first and second substances,
Pothe compensated photovoltaic actual output power is obtained;
Phthe actual output power of the uncompensated photovoltaic is obtained;
Pcand outputting power for the super capacitor.
Example 1
The method for compensating the photovoltaic power prediction error based on the weather condition according to the embodiment of the invention is described in detail below.
The invention provides a photovoltaic power prediction error compensation method based on weather conditions, which comprises the following steps:
s100: collecting photovoltaic output power data at equal intervals from historical data in a preset time period of a photovoltaic power station as sample data;
s200: removing nighttime data from the sample data, dividing the data according to weather conditions, and carrying out normalization processing on the data;
s300: constructing an LSSVM prediction model, predicting photovoltaic output power under different weather conditions, and comparing the photovoltaic output power with actual output power to obtain an output power prediction error under the weather conditions;
s400: respectively establishing a prediction error model for the photovoltaic output power prediction error under each weather condition by using a Gaussian Coiplar function;
s500: predicting the photovoltaic output power at the next moment by using the constructed LSSVM prediction model to obtain a photovoltaic output power predicted value PprePredicting the photovoltaic output power PpreSubstituting the prediction error model into the established prediction error model to obtain a predicted value P of the known photovoltaic output powerpreConditional probability distribution of photovoltaic output power prediction errors under conditions;
s600: compensating a difference value between the photovoltaic predicted value and an actual output value of the photovoltaic output power by using the super capacitor to construct a super capacitor constraint condition;
s700: carrying out deterministic transformation on the opportunity constraint and constructing an opportunity constraint condition;
s800: output power P of super capacitorcAs an objective function, the prediction error delta is used as a random variable, a longicorn-firefly search algorithm is adopted to solve a prediction error model, and when the super-capacitor constraint condition and the opportunity constraint condition are met at the same time, the optimal value of the output power of the super-capacitor is obtained;
s900: and compensating the photovoltaic predicted output power by adopting the optimal value of the output power of the super capacitor to obtain the compensated photovoltaic actual output power.
Example 2
The method for compensating the photovoltaic power prediction error based on the weather condition according to the embodiment of the invention is described in detail below.
The invention provides a photovoltaic power prediction error compensation method based on weather conditions, which comprises the following steps of:
s100: collecting photovoltaic output power data at equal intervals from historical data in a preset time period of a photovoltaic power station as sample data;
s200: removing nighttime data from the sample data, dividing the data according to weather conditions, and carrying out normalization processing on the data; the data is linearly transformed using min-max normalization, mapping the resulting values between [0,1], with the transfer function as follows:
Figure BDA0003565902880000141
wherein the content of the first and second substances,
xminis the minimum value of sample data;
xmaxis the maximum value of the sample data;
x*is the normalized value;
s300: constructing an LSSVM prediction model, predicting photovoltaic output power under different weather conditions, and comparing the photovoltaic output power with actual output power to obtain an output power prediction error under the weather conditions;
s400: respectively establishing a prediction error model for the photovoltaic output power prediction error under each weather condition by using a Gaussian Coipla function;
s500: predicting the photovoltaic output power at the next moment by using the constructed LSSVM prediction model to obtain a photovoltaic output power predicted value PprePredicting the photovoltaic output power PpreSubstituting the prediction error model into the established prediction error model to obtain a predicted value P of the known photovoltaic output powerpreConditional probability distribution of photovoltaic output power prediction errors under conditions;
s600: compensating a difference value between the photovoltaic predicted value and an actual output value of the photovoltaic output power by using the super capacitor to construct a super capacitor constraint condition; the super-capacitor constraint conditions constructed in the step S600 are as follows:
SOCmin≤SOC≤SOCmax
when SOC is reachedmin≤SOC≤SOCmaxIn time, the super capacitor can be charged and also can be discharged;
when SOC is less than or equal to SOCminWhile, the capacitor can only be charged;
when SOC is more than or equal to SOCmaxIn time, the capacitor can only discharge;
wherein the content of the first and second substances,
SOCminand SOCmaxRespectively the minimum value and the maximum value of the super capacitor SOC;
s700: carrying out deterministic transformation on the opportunity constraint and constructing an opportunity constraint condition; constructing opportunity constraints in S700 includes:
output power P to the super capacitorcAs an objective function, the prediction error delta is used as a random variable, and an opportunity constraint condition is constructed as follows:
P{f(Pc,δ)≤σ,t＝1,2,...}≥θ
Figure BDA0003565902880000151
wherein the content of the first and second substances,
sigma is a given allowable range of error;
θ is the given confidence;
Figure BDA0003565902880000152
output power for the super capacitor at t;
δtis the prediction error at t;
Figure BDA0003565902880000153
is the photovoltaic output power predicted value at t;
s800: output power P of super capacitorcAs a target function, taking the prediction error delta as a random variable, solving a prediction error model by adopting a longicorn-firefly search algorithm, and obtaining an output power optimal value of the super capacitor when a super capacitor constraint condition and an opportunity constraint condition are simultaneously met;
s900: and compensating the photovoltaic predicted output power by adopting the optimal value of the output power of the super capacitor to obtain the compensated photovoltaic actual output power.
Example 3
The method for compensating the photovoltaic power prediction error based on the weather conditions according to the embodiment of the present invention is described in detail below.
The invention provides a photovoltaic power prediction error compensation method based on weather conditions, which comprises the following steps:
s100: collecting photovoltaic output power data at equal intervals from historical data in a preset time period of a photovoltaic power station as sample data;
s200: removing nighttime data from the sample data, dividing the data according to weather conditions, and carrying out normalization processing on the data; weather conditions include sunny, cloudy, light rain, and heavy rain; the data is linearly transformed using min-max normalization, mapping the resulting values between [0,1], with the following transfer function:
Figure BDA0003565902880000161
wherein the content of the first and second substances,
xminis the minimum value of the sample data;
xmaxis the maximum value of the sample data;
x*is the normalized value;
s300: constructing an LSSVM prediction model, predicting photovoltaic output power under different weather conditions, and comparing the photovoltaic output power with actual output power to obtain an output power prediction error under the weather conditions;
s400: respectively establishing a prediction error model for the photovoltaic output power prediction error under each weather condition by using a Gaussian Coipla function; step S400 includes the steps of:
s401: recording the historical output actual value as alphai.tAnd historical predicted values are recorded as betai.tAnd outputting the actual value alpha to the history of the weather conditioni.tHezhou calendarHistorical predicted value betai.tRespectively counting to obtain the edge distribution function F of the actual photovoltaic outputɑtAnd edge distribution function F of photovoltaic prediction outputβt；
Wherein the content of the first and second substances,
i belongs to [ 1.,. n ] represents the weather condition, and n is the type number of the weather condition;
t epsilon [1, T ] represents the period of time,
t is the total time interval of the historical data;
s402: solving historical output actual value alphai.tAnd historical predicted value betai.tThe Kendall correlation coefficient of (1);
s403: respectively establishing prediction error models under different weather conditions by adopting Gaussian Coipla functions according to Kendall correlation coefficients;
s500: predicting the photovoltaic output power at the next moment by using the constructed LSSVM prediction model to obtain a photovoltaic output power predicted value PprePredicting the photovoltaic output power PpreSubstituting the prediction error model into the established prediction error model to obtain a predicted value P of the known photovoltaic output powerpreConditional probability distribution of photovoltaic output power prediction errors under conditions;
the conditional probability density function of the photovoltaic output power prediction error in S500 is:
Figure BDA0003565902880000171
wherein the content of the first and second substances,
delta is the prediction error;
fXY(x,Ppre) Is a joint probability density function;
fY(Ppre) To predict point PpreThe edge distribution probability density of (1);
c(FX(δ+Ppre),FX(Ppre) Is a coupli density function;
fX(δ+Ppre) The edge distribution probability density of the actual output power value;
x is the actual output value of the photovoltaic output power;
Ppreis a predicted value of the photovoltaic output power.
S600: compensating a difference value between the photovoltaic predicted value and an actual output value of the photovoltaic output power by using the super capacitor to construct a super capacitor constraint condition; the super-capacitor constraint conditions constructed in the step S600 are as follows:
SOCmin≤SOC≤SOCmax
when SOC is reachedmin≤SOC≤SOCmaxIn time, the super capacitor can be charged and also can be discharged;
when SOC is less than or equal to SOCminIn time, the capacitor can only be charged;
when SOC is more than or equal to SOCmaxIn time, the capacitor can only discharge;
wherein the content of the first and second substances,
SOCminand SOCmaxRespectively the minimum value and the maximum value of the super capacitor SOC;
s700: carrying out deterministic transformation on the opportunity constraint and constructing an opportunity constraint condition; constructing opportunity constraints in S700 includes:
output power P of super capacitorcAs an objective function, the prediction error delta is used as a random variable, and an opportunity constraint condition is constructed as follows:
P{f(Pc,δ)≤σ,t＝1,2,...}≥θ
Figure BDA0003565902880000181
wherein the content of the first and second substances,
sigma is a given allowable range of error;
θ is the given confidence;
Figure BDA0003565902880000182
output power for the super capacitor at t;
δtis the prediction error at t;
Figure BDA0003565902880000183
is the photovoltaic output power predicted value at t;
s800: output power P of super capacitorcAs an objective function, the prediction error delta is used as a random variable, a longicorn-firefly search algorithm is adopted to solve a prediction error model, and when the super-capacitor constraint condition and the opportunity constraint condition are met at the same time, the optimal value of the output power of the super-capacitor is obtained; step S800 includes the following steps:
s801: adopting a longicorn-firefly search algorithm, changing the number of the longicorn in the BAS algorithm from one longicorn to a plurality of longicorn, adopting Cubic mapping to replace pseudo-random number generation in the original initialization, and initializing the position of the longicorn;
s802: adding odor factors to the longicorn mutually, and setting a mutual attraction degree beta (r), wherein an objective function is the maximum odor source; in step S802, the mutual attraction β (r) between longicorn in the longicorn-firefly search algorithm is as follows:
Figure BDA0003565902880000184
wherein the content of the first and second substances,
β0is the initial attraction degree;
r is the distance between two longicorn;
gamma is the odor absorption coefficient; the odor gradually weakens with the increase of the distance;
s803: adopting an optimization objective function F as the fitness of each longicorn in the population; in step S803, an optimization objective function F is used as the fitness of each longicorn in the population, and the formula is as follows:
Fit(xis)＝F
wherein the content of the first and second substances,
Fit(xis) Representing the fitness value of the ith longicorn in the current spatial position;
xisrepresenting the current spatial position of the ith longicorn in the s-dimensional space;
i represents the ith longicorn;
s804: updating the position of the population individual; in step S804, when the t-th iteration is performed, the position update formula of the ith longicorn in the S-dimensional space is as follows:
Figure BDA0003565902880000191
wherein, the first and the second end of the pipe are connected with each other,
Figure BDA0003565902880000192
in the form of an incremental function of the function,
Figure BDA0003565902880000193
the expression is as follows:
Figure BDA0003565902880000194
wherein Xij＝Xi-Xj，XiAnd XjRepresenting the spatial positions of two longicorn i and j;
sign () is a sign function;
Figure BDA0003565902880000195
and
Figure BDA0003565902880000196
the fitness of the left and right longicorn whiskers is respectively;
αtthe step length is the step length of the iteration time t;
s805: the longicorn with the largest smell updates the position of the longicorn according to the following formula;
Figure BDA0003565902880000197
wherein the content of the first and second substances,
Figure BDA0003565902880000198
Figure BDA0003565902880000199
is the initial direction of the longicorn;
s is a spatial dimension;
rands () is a random function;
t is the current moment, and t +1 is the next moment;
Figure BDA00035659028800001910
the fitness of the right tendrils of the longicorn at the time t;
Figure BDA00035659028800001911
the adaptability of the left tendril of the longicorn at the time t;
Figure BDA0003565902880000201
the position of the ith longicorn at the time t in the s-dimensional space;
αtthe step length is the iteration time t;
s806: obtaining an optimal solution of the objective function to obtain an optimal value of the output power of the super capacitor; s806 includes the steps of:
the accuracy of the prediction result measured by the cosine value of the included angle between the predicted output power value and the actual output power value is shown as the formula:
Figure BDA0003565902880000202
wherein the content of the first and second substances,
Figure BDA0003565902880000203
has a value range of [0,1]](ii) a When in use
Figure BDA0003565902880000204
Time, represents the predicted power output curve versus the actual power output curveMost similar;
Figure BDA0003565902880000205
predicting an n-dimensional vector consisting of output power values for the photovoltaic cells;
Figure BDA0003565902880000206
and the vector is an n-dimensional vector formed by photovoltaic actual output power values.
S900: compensating the photovoltaic predicted output power by adopting the optimal value of the output power of the super capacitor to obtain the compensated photovoltaic actual output power;
in step S900, the compensated actual photovoltaic output power is Po＝Ph+Pc；
Wherein the content of the first and second substances,
Pothe compensated photovoltaic actual output power is obtained;
Phactual output power for uncompensated photovoltaic;
Pcand outputting power for the super capacitor.
Example 4
The prediction error compensation effect of the present invention will be described in detail below according to an embodiment of the present invention.
Table 1 below compares the normal distribution model with the prediction error distribution model of the present invention, demonstrating the accuracy of the method of the present invention.
TABLE 1 prediction error probability model evaluation
Figure BDA0003565902880000207
TABLE 2 Kendall correlation coefficients in four different weathers
In sunny days Light rain Heavy rain Cloudy day
0.9034 0.7057 0.6512 0.7942
Fig. 3 shows the compensated photovoltaic power actual output curve and the photovoltaic power predicted output curve obtained based on the method proposed by the present invention, and it can be seen that the compensated photovoltaic power actual output curve is closer to the photovoltaic power predicted output curve.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims (10)

1. A photovoltaic power prediction error compensation method based on weather conditions is characterized by comprising the following steps:
s100: collecting photovoltaic output power data at equal intervals from historical data in a preset time period of a photovoltaic power station as sample data;
s200: removing nighttime data from the sample data, dividing the data according to weather conditions, and carrying out normalization processing on the data;
s300: constructing an LSSVM prediction model, predicting photovoltaic output power under different weather conditions, and comparing the photovoltaic output power with actual output power to obtain an output power prediction error under the weather conditions;
s400: respectively establishing a prediction error model for the photovoltaic output power prediction error under each weather condition by using a Gaussian Coipla function;
s500: predicting the photovoltaic output power at the next moment by using the constructed LSSVM prediction model to obtain a photovoltaic output power predicted value PprePredicting the photovoltaic output power PpreSubstituting the prediction error model into the established prediction error model to obtain a predicted value P of the known photovoltaic output powerpreConditional probability distribution of photovoltaic output power prediction errors under conditions;
s600: compensating a difference value between the photovoltaic predicted value and an actual output value of the photovoltaic output power by using the super capacitor to construct a super capacitor constraint condition;
s700: carrying out deterministic transformation on the opportunity constraint and constructing an opportunity constraint condition;
s800: output power P of super capacitorcAs an objective function, the prediction error delta is used as a random variable, a longicorn-firefly search algorithm is adopted to solve a prediction error model, and when the super-capacitor constraint condition and the opportunity constraint condition are met at the same time, the optimal value of the output power of the super-capacitor is obtained;
s900: and compensating the photovoltaic predicted output power by adopting the optimal value of the output power of the super capacitor to obtain the compensated photovoltaic actual output power.
2. The weather condition-based photovoltaic power prediction error compensation method of claim 1, wherein the step S400 comprises the steps of:
s401: recording the historical output actual value as alphai.tAnd historical predicted values are recorded as betai.tAnd outputting the actual value alpha to the history of the weather conditioni.tAnd historical predicted value betai.tRespectively counting to obtain the edge distribution function F of the actual photovoltaic outputɑtAnd edge distribution function F of photovoltaic prediction outputβt；
Wherein the content of the first and second substances,
i belongs to [ 1.,. n ] represents the weather condition, and n is the type number of the weather condition;
t belongs to [1, T ] to represent the period;
t is the total time interval of the historical data;
s402: solving historical output actual value alphai.tAnd historical predicted value betai.tThe Kendall correlation coefficient of (1);
s403: and respectively establishing prediction error models under different weather conditions by adopting Gaussian Coipla functions according to the Kendall correlation coefficients.
3. The weather condition-based photovoltaic power prediction error compensation method of claim 2, wherein the conditional probability density function of the photovoltaic output power prediction error in S500 is as follows:
Figure FDA0003565902870000021
wherein the content of the first and second substances,
delta is the prediction error;
fXY(x,Ppre) Is a joint probability density function;
fY(Ppre) To predict point PpreThe edge distribution probability density of (1);
c(FX(δ+Ppre),FX(Ppre) Is a coupli density function;
fX(δ+Ppre) The edge distribution probability density of the actual output power value;
x is the actual output value of the photovoltaic output power;
Ppreis a predicted value of the photovoltaic output power.
4. The weather condition-based photovoltaic power prediction error compensation method of claim 3, wherein constructing the opportunity constraint condition in S700 comprises:
output power P of super capacitorcAs an objective function, the prediction error delta is used as a random variable, and an opportunity constraint condition is constructed as follows:
P{f(Pc,δ)≤σ,t＝1,2,...}≥θ
Figure FDA0003565902870000022
wherein the content of the first and second substances,
sigma is a given allowable range of error;
θ is the given confidence;
Figure FDA0003565902870000023
output power for the super capacitor at t;
δtis the prediction error at t;
Figure FDA0003565902870000024
is the predicted value of the photovoltaic output power at t.
5. The weather condition-based photovoltaic power prediction error compensation method of claim 4, wherein the step S800 comprises the steps of:
s801: adopting a longicorn-firefly search algorithm, changing the number of the longicorn in the BAS algorithm from one longicorn to a plurality of longicorn, adopting Cubic mapping to replace pseudo-random number generation in the original initialization, and initializing the position of the longicorn;
s802: adding odor factors to the longicorn, and setting the mutual attraction degree beta (r), wherein the objective function is the maximum odor source;
s803: adopting an optimization objective function F as the fitness of each longicorn in the population;
s804: updating the position of the population individual;
s805: the longicorn with the largest smell updates the position of the longicorn according to the following formula;
Figure FDA0003565902870000031
s806: obtaining an optimal solution of the objective function to obtain an optimal value of the output power of the super capacitor;
wherein the content of the first and second substances,
Figure FDA0003565902870000032
Figure FDA0003565902870000033
is the initial direction of the longicorn;
s is a spatial dimension;
rands () is a random function;
t is the current moment, and t +1 is the next moment;
Figure FDA0003565902870000034
the fitness of the right tendrils of the longicorn at the time t;
Figure FDA0003565902870000035
the adaptability of the left beard of the longicorn at the time t;
Figure FDA0003565902870000036
the position of the ith longicorn at the time t in the s-dimensional space;
αtis the step size at the iteration number t.
6. The weather condition-based photovoltaic power prediction error compensation method of claim 5, wherein in step S802, the mutual attraction degree β (r) between the longicorn in the longicorn-firefly search algorithm is as follows:
Figure FDA0003565902870000037
wherein the content of the first and second substances,
β0is the initial attraction degree;
r is the distance between two longicorn;
gamma is the odor absorption coefficient.
7. The weather condition-based photovoltaic power prediction error compensation method according to claim 6, wherein in step S803, an optimization objective function F is adopted as the fitness of each longicorn in the population, and the formula is as follows:
Fit(xis)＝F
wherein, the first and the second end of the pipe are connected with each other,
Fit(xis) Representing the fitness value of the ith longicorn in the current spatial position;
xisrepresenting the current spatial position of the ith longicorn in the s-dimensional space;
i represents the ith longicorn.
8. The weather-condition-based photovoltaic power prediction error compensation method of claim 7, wherein in step S804, when the t-th iteration is performed, a position update formula of the ith longicorn in the S-dimensional space is as follows:
Figure FDA0003565902870000041
wherein the content of the first and second substances,
Figure FDA0003565902870000042
in the form of an incremental function of the function,
Figure FDA0003565902870000043
the expression is as follows:
Figure FDA0003565902870000044
wherein Xij＝Xi-Xj，XiAnd XjRepresenting the spatial positions of two longicorn i and j;
sign () is a sign function;
Figure FDA0003565902870000045
and
Figure FDA0003565902870000046
the fitness of the left and right longicorn whiskers is respectively.
9. The weather condition-based photovoltaic power prediction error compensation method of claim 1, wherein S806 obtains an optimal solution of an objective function to obtain the output power P of the super capacitorcThe optimal values include:
the accuracy of the prediction result is measured by the cosine value of the included angle between the predicted output power value and the actual output power value, as shown in the formula:
Figure FDA0003565902870000047
wherein the content of the first and second substances,
Figure FDA0003565902870000048
has a value range of [0,1]](ii) a When in use
Figure FDA0003565902870000049
When the predicted power output curve is most similar to the actual power output curve, the predicted power output curve is represented;
Figure FDA00035659028700000410
an n-dimensional vector composed of photovoltaic prediction output power values;
Figure FDA0003565902870000051
and the vector is an n-dimensional vector formed by photovoltaic actual output power values.
10. The weather-condition-based photovoltaic power prediction error compensation method of claim 1, wherein the weather conditions include sunny days, cloudy days, light rain, and heavy rain.
CN202210307054.7A 2022-03-25 2022-03-25 Photovoltaic power prediction error compensation method based on weather conditions Active CN114662393B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210307054.7A CN114662393B (en) 2022-03-25 2022-03-25 Photovoltaic power prediction error compensation method based on weather conditions

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210307054.7A CN114662393B (en) 2022-03-25 2022-03-25 Photovoltaic power prediction error compensation method based on weather conditions

Publications (2)

Publication Number Publication Date
CN114662393A true CN114662393A (en) 2022-06-24
CN114662393B CN114662393B (en) 2023-06-30

Family

ID=82032560

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210307054.7A Active CN114662393B (en) 2022-03-25 2022-03-25 Photovoltaic power prediction error compensation method based on weather conditions

Country Status (1)

Country Link
CN (1) CN114662393B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1632667A1 (en) * 2004-09-07 2006-03-08 Renault s.a.s. Method of regeneration of a particle filter and apparatus for implementing such a method
CN103762616A (en) * 2014-01-03 2014-04-30 国家电网公司 Hybrid energy storage photovoltaic power generation system based on power prediction method
CN108205713A (en) * 2016-12-16 2018-06-26 中国电力科学研究院 A kind of region wind power prediction error distribution determination method and device
CN112507613A (en) * 2020-12-01 2021-03-16 湖南工程学院 Second-level ultra-short-term photovoltaic power prediction method
US20220029581A1 (en) * 2020-07-24 2022-01-27 Tianjin University Photovoltaic power generation virtual inertia compensation system and method based on super capacitor energy storage

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1632667A1 (en) * 2004-09-07 2006-03-08 Renault s.a.s. Method of regeneration of a particle filter and apparatus for implementing such a method
CN103762616A (en) * 2014-01-03 2014-04-30 国家电网公司 Hybrid energy storage photovoltaic power generation system based on power prediction method
CN108205713A (en) * 2016-12-16 2018-06-26 中国电力科学研究院 A kind of region wind power prediction error distribution determination method and device
US20220029581A1 (en) * 2020-07-24 2022-01-27 Tianjin University Photovoltaic power generation virtual inertia compensation system and method based on super capacitor energy storage
CN112507613A (en) * 2020-12-01 2021-03-16 湖南工程学院 Second-level ultra-short-term photovoltaic power prediction method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HAO HE; PRAMOD K. VARSHNEY: "A Coalitional Game for Distributed Inference in Sensor Networks With Dependent Observations", IEEE *
田庆锋;杨清;张银银;: "核电资产证券化:运营模式、成本以及价值评估", 企业经济, no. 02 *
聂齐齐;张建成;王宁;: "独立光伏供电系统中多储能单元协调控制策略的研究", 可再生能源, no. 03 *

Also Published As

Publication number Publication date
CN114662393B (en) 2023-06-30

Similar Documents

Publication Publication Date Title
Gao et al. Day-ahead power forecasting in a large-scale photovoltaic plant based on weather classification using LSTM
Li et al. Photovoltaic power forecasting with a hybrid deep learning approach
CN110909912B (en) Park electric power system net load combination prediction method based on self-adaptive error feedback
Heydari et al. A novel composite neural network based method for wind and solar power forecasting in microgrids
Ramsami et al. A hybrid method for forecasting the energy output of photovoltaic systems
CN111260136A (en) Building short-term load prediction method based on ARIMA-LSTM combined model
CN111915092B (en) Ultra-short-term wind power prediction method based on long-short-term memory neural network
CN111144644B (en) Short-term wind speed prediction method based on variation variance Gaussian process regression
CN108694470B (en) Data prediction method and device based on artificial intelligence
CN112288164B (en) Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast
Zhang et al. Wind speed prediction research considering wind speed ramp and residual distribution
Zhang et al. A new solar power output prediction based on hybrid forecast engine and decomposition model
CN111815027A (en) Photovoltaic station generated power prediction method and system
CN112329990A (en) User power load prediction method based on LSTM-BP neural network
CN115759415A (en) Power consumption demand prediction method based on LSTM-SVR
CN113537582A (en) Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction
CN115186923A (en) Photovoltaic power generation power prediction method and device and electronic equipment
CN115600729A (en) Grid load prediction method considering multiple attributes
CN115545333A (en) Method for predicting load curve of multi-load daily-type power distribution network
CN113991711B (en) Capacity configuration method for energy storage system of photovoltaic power station
CN115759389A (en) Day-ahead photovoltaic power prediction method based on weather type similar day combination strategy
CN114117852B (en) Regional heat load rolling prediction method based on finite difference working domain division
CN110795841A (en) Mathematical modeling method for uncertainty of intermittent energy output
Meng et al. A new PV generation power prediction model based on GA-BP neural network with artificial classification of history day
CN114091647A (en) Solar 10.7 cm radio flow forecasting method based on BP neural network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant