CN110610262B - Long-time scale photovoltaic time sequence output generation method considering weather elements - Google Patents

Abstract

The utility model discloses a long-time scale photovoltaic time sequence output generation method considering weather elements, which comprises the following steps of: s1, generating solar radiation data with constant atmospheric transparency; s2, correcting the atmospheric transparency of the photovoltaic installation site by using the recorded data of the NASA; s3, generating time sequence weather change data by using a Monte Carlo method and correcting the solar radiation intensity; s4, generating a photovoltaic time sequence active output curve according to the output characteristic curve of the photovoltaic panel; s5, adjusting buckling coefficients according to the number of annual utilization hours; the method can enable the final power supply configuration scheme to be more in line with the actual situation and meet the requirement of power supply reliability when the power supply planning research of the photovoltaic equipment is carried out; the power supply configuration scheme has a good improvement effect on the power supply reliability.

Description

A long-term scale photovoltaic time-series output generation method considering weather elements

technical field

The invention is applicable to the field of power supply optimization planning including photovoltaic equipment, and in particular relates to a power supply optimization method for evaluating power supply reliability.

Background technique

In order to deal with environmental problems, global warming and achieve sustainable development, it is necessary to vigorously develop clean energy, among which photovoltaic power generation has achieved considerable development. When the cumulative installed capacity of photovoltaics reaches a relatively high level, it is necessary to effectively take into account the uncertain output of photovoltaics when carrying out power optimization planning including photovoltaic equipment, especially when evaluating the power supply reliability of power supply schemes, it is often necessary to use long-term Time-scale PV output data. However, due to the volatility of photovoltaic power generation, it is affected by uncontrollable factors such as day and night alternation, weather changes, and seasonal changes. .

There are many methods for predicting the intensity of solar radiation, and the existing methods can be roughly divided into three categories: 1) physical methods, that is, calculations based on the relative position of the sun and the earth and the atmospheric transparency of the earth's atmosphere; 2) sampling methods, generally considered that the solar radiation obeys Beta distribution, which can be realized through random number sampling according to Beta distribution; 3) Forecasting methods, using methods such as support vector machines, neural networks, and fuzzy theory. The above methods are mainly aimed at short-term solar radiation intensity prediction. When generating long-term photovoltaic time-series output, the existing methods often have the following shortcomings: 1) The solar radiation intensity data obtained by the sampling method are independent and cannot reflect The weather continues to change, and the irregularity of the sampling data is quite mild compared to the long-term continuous weak radiation that exists due to rainy weather in reality, which makes the solar radiation intensity data based on the sampling method a good choice for power planning. There is a large error between the requirements for peak shaving capacity and the actual situation; 2) Forecasting methods generally require data of solar radiation intensity, ambient temperature, atmospheric humidity, etc. However, these data are difficult to obtain, and it is also difficult to judge the error range of the time-series solar radiation intensity data obtained by the prediction method.

Contents of the invention

Aiming at the technical problems existing in the prior art, the present invention proposes a long-term scale photovoltaic time-series output generation method that takes weather elements into account. This method can make the final power supply configuration plan more realistic when carrying out research on power supply planning with photovoltaic equipment The situation meets the reliability requirements of power supply; it has a very good effect on improving the power supply configuration scheme considering the reliability of power supply.

In order to solve the prior art problems, the present invention adopts the following technical solutions:

A long-term scale photovoltaic time-series output generation method considering weather elements, comprising the following steps:

S1. Collect statistical data such as latitude and longitude of the location of photovoltaic equipment, monthly average radiation, historical weather statistics, and annual utilization hours of photovoltaic equipment;

S2. According to the latitude data of the location of the photovoltaic equipment, calculate the direct solar radiation and scattered radiation when the atmospheric transparency is constant, and finally obtain the total solar radiation data;

S4、从S2中提取太阳辐射强度月均值并做标幺处理得到/>

按照设定步长修正/>

直至第last次修正后/>

与/>

相同，最终得到/>

S3. Extract the historical data of solar radiation intensity statistics by month from the NASA database, and obtain it after per-unit processing

S4. Extract the monthly mean value of solar radiation intensity from S2 and perform per-unit processing to obtain />

Correct according to the set step size />

until after revision last />

with />

same, end up with />

并得到修正后的太阳辐射向量I′₁；S5. Use the Monte Carlo sampling method to generate time-series weather change data, convert the weather change data into a discount coefficient that the weather is transparent to the atmosphere, and correct

And get the corrected solar radiation vector I′ ₁ ;

S6. Generate a photovoltaic time-series output vector according to the output characteristic curve of the photovoltaic panel;

S7. Calculate the annual utilization hours C ₁ corresponding to the photovoltaic time-series output vector according to the convergence criterion of the Monte Carlo method;

S8. Taking the statistical value C _std of the annual utilization hours of photovoltaic equipment at the photovoltaic installation site as the target, correct WF according to the set step size until the difference between the annual utilization hours C _final and C _std of the photovoltaic equipment at the final time is less than the set value The value ε', at this time, the target photovoltaic output vector is obtained.

The S5 step adjusts the transparency of the atmosphere according to the discount coefficient corresponding to the weather conditions:

5.1. According to the recorded data of the global weather network, all weather conditions are discretized into seven types: sunny, cloudy, cloudy, rain, snow, sand and dust, and the number of days they appear is recorded as a vector W, W=[w ₁ ,w ₂ ,…,w ₇ ], where the subscripts correspond to the above seven weather types in turn;

5.2. Assuming that the durations of the seven weathers all obey the logarithmic normal distribution, and the mean and standard deviation of the seven weather durations are respectively recorded as vectors M and D, then M=[m ₁ ,m ₂ ,..., m ₇ ], D = [d ₁ , d ₂ , . . . , d ₇ ]. The occurrence probability of seven kinds of weather is recorded as vector P, P=[p _w,1 ,p _w,2 ,…,p _w,7 ]; the probability distribution vector corresponding to seven kinds of weather is recorded as vector P′, P′=[ p′ _w,1 ,p′ _w,2 ,…,p′ _w,7 ], where

由SW和/>

可以得到时序太阳辐射强度数据I′₁，I′₁＝[I′_1,1,I′_1,2,…,I′_1,8760]，其中5.3. The influence coefficient of seven kinds of weather on the solar radiation intensity at the photovoltaic installation site is recorded as a vector WF, WF=[wf ₁ ,wf ₂ ,…,wf ₇ ], and 0<wf _i <1, which means that the solar radiation intensity reaches the solar radiation under this kind of weather. The discount factor for the solar radiation intensity of the panel. Based on the Markov process, use M, D and P′ to generate time-series weather change data at intervals of hours within a year, and then obtain the time-series data of the influence of seven weathers on solar radiation intensity in a year according to WF, which is recorded as vector SW, then SW={sw ₁ ,sw ₂ ,...,sw ₈₇₆₀ }, sw _i ∈ {wf ₁ ,wf ₂ ,...,wf ₇ },

by SW and />

The time series solar radiation intensity data I′ ₁ can be obtained, I′ ₁ =[I′ _1,1 ,I′ _1,2 ,…,I′ _1,8760 ], where

Beneficial effect

First, the present invention takes the long-term time-series output vector of photovoltaic units as the research object. When using this invention, it only needs to input three types of statistical data that are easy to obtain: latitude and longitude data of photovoltaic installation sites, NASA’s monthly average light intensity of photovoltaic installation sites The statistical data from the global weather network and the historical weather data from the global weather network can obtain long-term scale photovoltaic output data taking into account diurnal changes, seasonal changes and weather elements, and then apply time-series photovoltaic output data to optimization problems such as power planning. Due to full consideration of periodic changes with sunrise and sunset, overall trend changes with months or seasons, and high and low changes that last for several days with weather changes, the generated time series output curve has high accuracy. The reliability of the power supply planning scheme is greatly improved, and it has good engineering application value.

Second, the present invention highlights irregular changes in photovoltaic output through effective processing of atmospheric transparency parameters, and can simultaneously take into account the periodic changes in photovoltaic output with sunrise and sunset, and the overall trend that occurs with months or seasons Changes, high and low changes that last for several days with weather changes, and the annual utilization hours of photovoltaic equipment can be adjusted according to local lighting statistics. Overall, under the premise of slightly increasing the difficulty and complexity of calculation, the effective consideration of weather elements is realized, and the obtained long-term scale photovoltaic output data is closer to the actual situation.

Description of drawings

Fig. 1 is a flow chart of a method for generating a long-term scale photovoltaic time-series output in consideration of weather elements in the present invention;

Fig. 2 is the step-by-step generation process of the long-term scale photovoltaic time-series output time-series curve of the present invention;

Fig. 3 is the output characteristic curve of the photovoltaic equipment of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

The long-term scale photovoltaic output in the present invention refers to the photovoltaic output with a length of at least 1 year. The implementation process of the patent of the present invention will be further described in detail below in conjunction with the accompanying drawings.

As shown in Figure 1, the present invention provides a method for optimizing the power supply of photovoltaic equipment based on weather requirements, including the following steps:

S1. Collect statistical data such as latitude and longitude of the location of photovoltaic equipment, monthly average radiation, historical weather statistics, and annual utilization hours of photovoltaic equipment;

S2, (101, 201) respectively calculate the solar direct radiation and diffuse radiation when the atmospheric transparency is constant according to the latitude data of the photovoltaic equipment location, and finally obtain the total solar radiation data;

This step generates solar radiation data with constant atmospheric transparency, including the following:

Set the transparency of the atmosphere as a constant value p ₀ (such as p ₀ =1), and take hours (or shorter time, such as 1 minute) as time intervals to calculate the time-series solar direct radiation, diffuse radiation and total radiation for a period of one year Intensity, the longitude and latitude information of the photovoltaic equipment installation site needs to be input during the calculation process.

sinα _s = sinφ*sinδ+cosφ*cosδ*cosω (1-2)

ω＝(k-12)*15 (1-3)

I _b ＝r*I _se *p ^m *sinα _s (1-4)

I _h =I _b +I _d (1-8)

Among them, δ represents the solar declination angle, -23.44°≤δ≤23.44°; n represents the nth day of the year, assuming that n=1 represents January 1st; α _s represents the solar altitude angle; φ represents the simulated location ω represents the solar hour angle, 0° at noon, 1h every 15°, negative in the morning and positive in the afternoon; I _b represents the direct solar radiation intensity on the horizontal plane, W/m ² ; I _se represents the upper limit of the atmosphere The solar radiation intensity of , usually taken as 1367W/m ² ; p represents atmospheric transparency, dimensionless; m represents atmospheric mass, dimensionless; r represents the sun-terrestrial correction coefficient; I _d represents the horizontal plane solar radiation intensity, W/m ² ; M ₁ and M ₂ are empirical values corresponding to atmospheric transparency p ₀ , dimensionless; I _h represents the total solar radiation on the horizontal plane, W/m ² .

如图2中图a所示大气透明度为常数的太阳辐射强度时序曲线，该部分内容不属于本文的核心创新工作，具体可参考文献[1]。该步骤可得到如图2.a所示的太阳辐射向量。When taking hours as the time interval, the above formula can be used to obtain the time-series total solar radiation vector of 8760h throughout the year

Figure a in Figure 2 shows the time-series curve of solar radiation intensity with constant atmospheric transparency. This part of the content does not belong to the core innovation work of this paper. For details, please refer to [1]. This step can get the solar radiation vector shown in Figure 2.a.

S3, (301) extract the historical data of solar radiation intensity statistics by month in the NASA database, and obtain after doing per-unit processing

This step extracts the recorded data of NASA and performs per-unit processing, including the following contents:

可得到12个月份每月的辐射总量，记为/>

修正/>

直至与I_sd的标幺值相同。The historical average value of the solar radiation intensity of the photovoltaic installation site according to the month is recorded as a vector I _sd (can be obtained through the NASA website), I _sd = [I _sd,1 ,I _sd ,2,...,I _sd,12 ]; according to

The total amount of radiation per month for 12 months can be obtained, denoted as />

Amend />

Up to the same per unit value as I _sd .

按照设定步长修正/>

直至第last次修正后/>

与/>

相同，最终得到/>

S4, (401) Extract the monthly average value of solar radiation intensity from S2 and perform per unit processing to obtain

Correct according to the set step size />

until after revision last />

with />

same, end up with />

This step is based on NASA's monthly historical statistical data to adjust the solar radiation intensity, including the following:

为目标调整每月的平均大气透明度和每天的平均大气透明度，使得/>

与/>

相吻合，为简化模拟过程，本文假定每天的大气透明度保持不变。每月第15日的大气透明度记为p_i,15，i∈{1,2,…,12}。搜索/>

与/>

差别最大的元素，对应的序号记为k，k∈{1,2,…,12}：by

The monthly average atmospheric transparency and the daily average atmospheric transparency are adjusted for the target such that />

with />

Coincidentally, in order to simplify the simulation process, this paper assumes that the atmospheric transparency remains constant every day. The atmospheric transparency on the 15th day of each month is recorded as p _i,15 , i∈{1,2,…,12}. search />

with />

The element with the largest difference, the corresponding serial number is recorded as k, k∈{1,2,…,12}:

p _k,15 =p _k,15 +Δp (1-13)

根据12个月份第15日的大气透明度p_i,15，利用线性插值方法可以计算出每一天的大气透明度，此处不再展开。Among them, Δp represents the corrected step size for p _i,15 , formula (1-12) and formula (1-13) guarantee

According to the atmospheric transparency p _i,15 on the 15th day of the 12 months, the linear interpolation method can be used to calculate the atmospheric transparency of each day, which will not be expanded here.

对应的月辐射总量变为/>

然后再搜索/>

与/>

差别最大的元素并重复上述修正过程，直到第last次修正后得到/>

和/>

且/>

与/>

所有元素的差别都小于某设定值ε(比如ε＝0.01)。该步骤可得到如图2.b所示的太阳辐射向量。Using the corrected atmospheric transparency to recalculate the total solar radiation time series vector becomes

The corresponding total monthly radiation becomes />

then search for />

with />

The element with the largest difference and repeat the above correction process until the last correction is obtained />

and />

and/>

with />

The differences of all elements are smaller than a certain set value ε (such as ε=0.01). This step can get the solar radiation vector shown in Figure 2.b.

并得到修正后的太阳辐射向量I′₁；S5, (501,601) use Monte Carlo sampling method to generate time-series weather change data, convert the weather change data into a discount coefficient that the weather is transparent to the atmosphere, and correct

And get the corrected solar radiation vector I′ ₁ ;

In this step, the Monte Carlo method is used to generate time-series weather change data, and the solar radiation intensity is corrected according to the atmospheric transparency discount coefficient corresponding to the weather data. The specific content is as follows:

According to the recorded data of the Global Weather Network, all weather conditions are discretized into seven types: sunny, cloudy, cloudy, rain, snow, sand and dust, and the number of days they appear is recorded as a vector W, W=[w ₁ ,w ₂ ,…,w ₇ ], where the subscripts correspond to the above seven weather types in turn.

Assuming that the durations of the seven weathers all obey the logarithmic normal distribution, and the mean and standard deviation of the durations of the seven weathers are recorded as vectors M and D respectively, then M=[m ₁ ,m ₂ ,…,m ₇ ], D=[d ₁ ,d ₂ ,...,d ₇ ]. The occurrence probability of seven kinds of weather is recorded as vector P, P=[p _w,1 ,p _w,2 ,…,p _w,7 ]; the probability distribution vector corresponding to seven kinds of weather is recorded as vector P′, P′=[ p′ _w,1 ,p′ _w,2 ,…,p′ _w,7 ], where

由SW和I_tlast可以得到时序太阳辐射强度数据I′₁，I′₁＝[I′_1,1,I′_1,2,…,I′_1,8760]，其中The influence coefficient of seven kinds of weather at the photovoltaic installation site on the solar radiation intensity is recorded as vector WF, WF=[wf ₁ ,wf ₂ ,…,wf ₇ ], and 0<wf _i <1, which means Discount factor for solar radiation intensity. Based on the Markov process, use M, D and P′ to generate time-series weather change data at intervals of hours within a year, and then obtain the time-series data of the influence of seven weathers on solar radiation intensity in a year according to WF, which is recorded as vector SW, then SW={sw ₁ ,sw ₂ ,...,sw ₈₇₆₀ }, sw _i ∈ {wf ₁ ,wf ₂ ,...,wf ₇ },

The time series solar radiation intensity data I′ ₁ can be obtained from SW and _Itlast , I′ ₁ =[I′ _1,1 ,I′ _1,2 ,…,I′ _1,8760 ], where

This step can get the solar radiation vector shown in Figure 2.c.

S6. (701) Generate a photovoltaic time-series output vector according to the output characteristic curve of the photovoltaic panel;

The real-time output of photovoltaics mainly depends on the light intensity. The relationship between the output of the photovoltaic array and the light intensity under this model is shown in Figure 3, which consists of three parts: nonlinear region, linear region and constant region.

S7. (801) Calculate the annual utilization hours C ₁ corresponding to the photovoltaic sequence output vector according to the convergence criterion of the Monte Carlo method;

This step is to calculate the annual utilization hours corresponding to the initially generated long-term scale photovoltaic output data. The specific content is as follows:

光伏的年利用小时数是N年数据的统计值，第i(1≤i≤N)年的光伏利用小时数为：According to the above steps, the time-series photovoltaic output vector of N (N can be taken as 10000) years is generated, which is denoted as

The annual utilization hours of photovoltaics is the statistical value of N-year data, and the utilization hours of photovoltaics in the i-th year (1≤i≤N) is:

Then the average annual utilization hours are:

S8. (901) Taking the statistical value C _std of the annual utilization hours of the photovoltaic equipment at the photovoltaic installation site as the target, and correcting WF according to the set step size, until the difference between the annual utilization hours C _final and C _std of the photovoltaic equipment at the final time If it is less than the set value ε', the target photovoltaic output vector is obtained at this time.

This step is to correct the long-term scale photovoltaic output vector according to the statistical value of the annual utilization hours. The specific content is as follows:

Since WF is a set amount, there is a difference between C ₁ and the actual photovoltaic panel utilization hours C _std of the simulated site. Next, adjust WF with C _std as the standard, and finally make the annual utilization hours the same as C _std . According to differences in research objectives or differences in regions, the same or different step sizes can be used to adjust wf _i sequentially during the adjustment process. The present invention uses the same adjustment step size:

变为/>

对应的年利用小时数平均值由C₁变为C₂；然后重复上述调整过程直到第final次调整后C_final与C_std的差距小于某设定值ε′(比如ε′＝3)。记录第final次调整后的WF，得到如图2.d所示的光伏电池时序功率曲线。After the first adjustment, the timing output vector of photovoltaic panels is given by

becomes />

The corresponding average annual utilization hours change from C ₁ to C ₂ ; then repeat the above adjustment process until the difference between C _final and C _std after the final adjustment is less than a certain set value ε' (eg ε'=3). Record the WF after the final adjustment, and obtain the time series power curve of the photovoltaic cell as shown in Figure 2.d.

It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the scope of protection of the utility model patent should be based on the appended claims.

Claims (1)

1. A long-term scale photovoltaic time-series output generation method considering weather elements, characterized in that it comprises the following steps: S1. Collect statistical data such as latitude and longitude of the location of photovoltaic equipment, monthly average radiation, historical weather statistics, and annual utilization hours of photovoltaic equipment; S2. According to the latitude data of the location of the photovoltaic equipment, calculate the direct solar radiation and scattered radiation when the atmospheric transparency is constant, and finally obtain the total solar radiation data; S3. Extract the historical data of solar radiation intensity statistics by month from the NASA database, and obtain it after per-unit processing

S4. Extract the monthly mean value of solar radiation intensity from S2 and perform per unit processing to obtain

Correct according to the set step size />

until after revision last />

with />

same, end up with />

S5. Use the Monte Carlo sampling method to generate time-series weather change data, convert the weather change data into a discount coefficient that the weather is transparent to the atmosphere, and correct

And obtain the corrected solar radiation vector I′ ₁ ; the S5 step adjusts the transparency of the atmosphere according to the discount coefficient corresponding to the weather conditions: 5.1. According to the recorded data of the global weather network, all weather conditions are discretized into seven types: sunny, cloudy, cloudy, rain, snow, sand and dust, and the number of days they appear is recorded as a vector W, W=[w ₁ ,w ₂ ,…,w ₇ ], where the subscripts correspond to the above seven weather types in turn; 5.2. Assuming that the durations of the seven weathers all obey the lognormal distribution, and the mean and standard deviation of the seven weather durations are respectively recorded as vectors M and D, then M=[m ₁ ,m ₂ ,…,m ₇ ], D=[d ₁ ,d ₂ ,...,d ₇ ]. The occurrence probability of seven kinds of weather is recorded as vector P, P=[p _w,1 ,p _w,2 ,…,p _w,7 ]; the probability distribution vector corresponding to seven kinds of weather is recorded as vector P′, P′=[ p′ _w,1 ,p′ _w,2 ,…,p′ _w,7 ], where

5.3. The influence coefficient of seven kinds of weather on the solar radiation intensity at the photovoltaic installation site is recorded as vector WF, WF=[wf ₁ ,wf ₂ ,…,wf ₇ ], and 0<wf _i <1, which means that the solar radiation intensity reaches the solar radiation under this kind of weather. The discount factor for the solar radiation intensity of the panel; Based on the Markov process, M, D and P′ are used to generate time-series weather change data at intervals of hours within a year, and then, according to WF, the time-series data of the influence of seven kinds of weather on solar radiation intensity in a year is obtained, which is recorded as vector SW, Then SW={sw ₁ ,sw ₂ ,…,sw ₈₇₆₀ }, sw _i ∈{wf ₁ ,wf ₂ ,…,wf ₇ },

by SW and />

The time series solar radiation intensity data I′ ₁ can be obtained, I′ ₁ =[I′ _1,1 ,I′ _1,2 ,…,I′ _1,8760 ], where:

S6. Generate a photovoltaic time-series output vector according to the output characteristic curve of the photovoltaic panel; S7. Calculate the annual utilization hours C ₁ corresponding to the photovoltaic time-series output vector according to the convergence criterion of the Monte Carlo method; S8. Taking the statistical value C _std of the annual utilization hours of photovoltaic equipment at the photovoltaic installation site as the target, correct WF according to the set step size until the difference between the annual utilization hours C _final and C _std of the photovoltaic equipment at the final time is less than the set value value ε′, at this time, the target photovoltaic output vector is obtained, where: Since WF is a set value, there is a difference between C ₁ and the actual utilization hours C _std of photovoltaic panels at the simulated site. Next, adjust WF with C _std as the standard, and finally make the annual utilization hours the same as C _std ; According to differences in research objectives or differences in regions, the same or different step sizes can be used to adjust wf _i sequentially during the adjustment process. The present invention uses the same adjustment step size:

The adjusted photovoltaic panel timing output vector is given by

becomes />

The corresponding average annual utilization hours change from C ₁ to C ₂ ; repeat the above adjustment process until the difference between C _final and C _std after the final adjustment is less than a certain set value ε′.

Images