A regional photovoltaic power station power generation power ultra-short-term prediction method

Abstract

The present application relates to a kind of regional photovoltaic power station power generation power ultra-short term prediction method.(1), certain demand area can be divided into several unit areas S1, S2, S3 …, weather type in unit area is regarded as consistent at any time, and photovoltaic power station power generation power in unit area is regarded as same under the influence of weather type;(2), determine the duration T of complex weather in a unit area;(3), determine the photovoltaic power station power generation power Pt under ideal weather condition in a unit area T period at a time t;(4), confirm the weather type Wt at t time in T period;(5), set the number of times of Monte Carlo cycle simulation as M;(6), determine each type of parameter influenced by weather type Wt;(7), determine the influence Mt of planned grid connection and routine maintenance on photovoltaic power station output;(8), repeat step (5) to step (7).

Description

TECHNICAL FIELD

[0001] The regional photovoltaic power station power ultra-short-term prediction method relates to a power prediction method in the photovoltaic power generation technical field, in particular to a photovoltaic power prediction method based on a Monte Carlo algorithm for the large fluctuation of photovoltaic power station power caused by the complex weather change in a future time period. BACKGROUND

[0002] In order to solve various problems caused by traditional energy, solar photovoltaic power generation technology develops rapidly; the randomness and intermittence of photovoltaic power generation bring significant influence on the stable and reliable operation of the power grid, and the accurate short-term photovoltaic and ultra-short-term photovoltaic power prediction is increasingly important, and there are many related research results at present.

[0003] The photovoltaic power station power prediction can be divided into indirect prediction and direct prediction according to the prediction process; among them, the direct prediction models the historical power generation and weather data, has the advantages of simple modeling process and no need to measure the solar radiation intensity; the meta-heuristic learning algorithm mainly based on neural network is widely used in the field of direct photovoltaic power prediction due to the advantages of strong nonlinear expression capacity; the direct prediction of photovoltaic power station power is based on historical power generation and weather forecast data, and due to the complex atmospheric environment, the change of weather type becomes one of the main reasons for the decrease of the prediction accuracy of photovoltaic power; when the original data sample is small and the weather forecast type is scattered, the learning efficiency of neural network decreases, and the complex law will be covered; when a relatively complex weather type is encountered, the direct prediction result of neural network is often greatly different from the actual photovoltaic power station power result, and cannot reflect the future photovoltaic power change trend.

[0004] Therefore, it is necessary to design a new algorithm to meet the actual prediction demand. SUMMARY

[0005] Therefore, the regional photovoltaic power station power ultra-short-term prediction method is provided, which is used for the complex weather change in a future time period in a region, and the Monte Carlo algorithm is used to predict the expected power generation of photovoltaic power station under sunny weather, so that the prediction accuracy can be improved, and the error caused by the law covering phenomenon can be reduced.

[0006] The steps of the regional photovoltaic power station power ultra-short-term prediction method are as follows:

[0007] (1) determining that a certain demand region can be divided into a plurality of unit regions S1, S2, S3, …, the weather type in the unit region is considered to be consistent at any time, and the photovoltaic power station power in the unit region is considered to be the same under the influence of the weather type;

[0008] (2) determining the complex weather duration T in a unit region;

[0009] (3), determine the photovoltaic power station power Pt at time t in the unit area T period under ideal weather conditions;

[0010] (4), confirm the weather type Wt at time t in the T period;

[0011] (5), set the number of Monte Carlo cycle simulation M;

[0012] (6), determine the parameters affected by the weather type Wt;

[0013] (7), determine the influence Mt of planned grid connection and routine maintenance on photovoltaic power station output;

[0014] (8), repeat steps (5)-(7) to obtain the photovoltaic power station power Pt at time t in the unit area complex weather duration T;

[0015] (9), let t = t + 1, repeat steps (2)-(8), superimpose the unit area photovoltaic power station power prediction curve with a total number of T to obtain the photovoltaic power station power prediction curve in the unit area complex weather duration T.

[0016] (6.1), the parameters affected by the weather type Wt in step (6) include: determining the influence St of the weather type on the total amount of solar radiation on the ground.

[0017] (6.2), the parameters affected by the weather type Wt in step (6) include: determining the influence E1t of the weather type at time t on the output of the solar photovoltaic cell assembly of the photovoltaic power station.

[0018] (6.3), the parameters affected by the weather type Wt in step (6) include: determining the power temperature coefficient change E2t of the weather type at time t on the characteristics of the solar photovoltaic cell assembly of the photovoltaic power station.

[0019] (6.4), the parameters affected by the weather type Wt in step (6) include: determining the influence E3t of the humidity change caused by the weather type at time t on the power generation of the photovoltaic power station.

[0020] (6.5), the parameters affected by the weather type Wt in step (6) include: determining the influence E4t of the destructive factors caused by the weather type at time t on the power attenuation of the solar photovoltaic cell assembly of the photovoltaic power station.

[0021] By using the technical scheme disclosed in the present application, the expected power generation of a photovoltaic power station under sunny weather is taken as the basis for prediction in view of the complex weather changes in a region in a future period of time, the prediction accuracy can be improved, and the error caused by regular shading phenomenon can be reduced, thus having good use and promotion effects. BRIEF DESCRIPTION OF DRAWINGS

[0022] The present application will be further described below with reference to the accompanying Figures 1-6 The present application will be further described below with reference to the accompanying

[0023] Figure 1 is a photovoltaic power station power generation curve diagram under sunny weather of a unit region.

[0024] Figure 2 is a photovoltaic power station power generation prediction curve diagram under complex weather (strong wind, temperature drop and rainfall) of the unit region, and the prediction method is a neural network algorithm.

[0025] Figure 3 is a comparison diagram of the neural network algorithm prediction curve and the actual photovoltaic power station power generation curve.

[0026] Figure 4 is a comparison diagram of the Monte Carlo algorithm prediction curve and the actual photovoltaic power station power generation curve.

[0027] Figure 5 is a comparison diagram of the neural network algorithm accuracy and the Monte Carlo algorithm accuracy.

[0028] Figure 6 is a curve diagram of the influence of five types of variables St, E1t, E2t, E3t, E4t and the sum Ut of the five types of variables related to the weather type Wt at t time on the photovoltaic power station output. DETAILED DESCRIPTION

[0029] The present application will be further described below with reference to the accompanying Figures 1-6 A regional photovoltaic power station power generation ultra-short-term prediction method, the steps of the research method are as follows:

[0030] (1) Determine that a certain demand region can be divided into a plurality of unit regions S1, S2, S3……, the weather type in the unit region is considered to be consistent at any time, and the photovoltaic power station power generation in the unit region is considered to be the same under the influence of the weather type;

[0031] (2) Determine the complex weather duration T in a unit region;

[0032] (3) Determine the photovoltaic power station power generation Pt under ideal weather conditions at a time t in the T period of the unit region;

[0033] (4) Confirm the weather type Wt at t time in the T period;

[0034] (5) Set the number of Monte Carlo cycle simulation M;

[0035] (6) Determine the parameters affected by the weather type Wt;

[0036] (7) Determine the impact Mt of the planned grid connection and routine maintenance on the output of the photovoltaic power station;

[0037] (8) Repeat steps (5) to (7) to obtain the photovoltaic power station power Pt at a certain time t in the unit area complex weather duration T;

[0038] (9) Let t = t + 1, repeat steps (2) to (8), superimpose the unit area photovoltaic power station power prediction curve with a total number of T, and obtain the photovoltaic power station power prediction curve graph in the unit area complex weather duration T.

[0039] (6.1) The parameters affected by the weather type Wt determined in step (6) include: determining the impact St of the weather type on the total amount of solar radiation on the ground.

[0040] (6.2) The parameters affected by the weather type Wt determined in step (6) include: determining the impact E1t of the weather type at time t on the shading of the photovoltaic power station solar photovoltaic cell assembly on the output of the solar cell assembly.

[0041] (6.3) The parameters affected by the weather type Wt determined in step (6) include: determining the power temperature coefficient change E2t caused by the weather type at time t on the characteristics of the photovoltaic power station solar photovoltaic cell assembly.

[0042] (6.4) The parameters affected by the weather type Wt determined in step (6) include: determining the impact E3t of the humidity change caused by the weather type at time t on the power generation of the photovoltaic power station.

[0043] (6.5) The parameters affected by the weather type Wt determined in step (6) include: determining the impact E4t of the destructive factors caused by the weather type at time t on the power attenuation of the photovoltaic power station solar photovoltaic cell assembly.

[0044] Embodiment

[0045] (1), sunny weather, photovoltaic power generation power curve similar to the bell-shaped curve, photovoltaic power generation power highest time often appears at noon, the solar elevation angle (solar light and the horizon angle) maximum, the atmospheric path of equal solar radiation is the shortest, the least affected by the atmosphere, the solar radiation reaching the ground is the most. At the same time, due to the maximum solar elevation angle, the area of equal solar radiation is the smallest, and the solar radiation per unit area on the ground is the most, so the noon solar radiation is the strongest. The start and end time of the photovoltaic power generation power curve is related to the sunrise time and sunset time, and the sunrise time and sunset time change with the season.

[0046] (2), in practical application, the photovoltaic power generation power curve is not an analog signal. The analog signal is continuous in time and continuous in amplitude. The photovoltaic power generation power curve is often a digital signal that is discrete in time and discrete in amplitude due to the intermittent collection time and the intermittent transmission and reception signal time. In this prediction research method, since the weather type Wt needs to be determined according to the weather forecast for t time in the complex weather duration T, the time interval between t and t+1 is set to 15 minutes.

[0047] (3), under sunny weather, the meta-heuristic learning algorithm based on neural network is used for photovoltaic power prediction; under complex weather, the regional photovoltaic power generation power ultra-short-term prediction research method based on Monte Carlo algorithm is switched to for photovoltaic power prediction. Referring to Table 1, the definition of complex weather is defined, that is, when and how to start the Monte Carlo algorithm for photovoltaic power prediction. When the weather forecast in the prediction area reaches the level of moderate rain and above, such as heavy rain, snow, short-time thunderstorm gale and other strong convective weather, the Monte Carlo algorithm can be started.

[0048] Table 1: Definition of weather type of meteorological system

[0049]

[0050] (4), according to the weather forecast, the initial time of extreme weather in the unit area is determined, the photovoltaic power generation power prediction power model in the unit area is established, and the complex weather duration T in the unit area is determined; the photovoltaic power generation power Pt in the unit area under ideal weather conditions at a certain time t in the T period is determined, which can be obtained from the bell-shaped curve of photovoltaic power generation power under sunny weather.

[0051] (5), confirm the weather type Wt at time t in the T period in the unit area, confirm the influence St of the weather type Wt at time t on the total solar radiation on the ground, and according to the weather forecast of the unit area every 15 minutes, the change of solar radiation intensity in the unit area can be well grasped, mainly the low-altitude thick cloud will have a more obvious influence on the solar radiation intensity, the rapid movement of this cloud system will cause the photovoltaic power generation power curve to fluctuate greatly, such asFigure 3 The number of Monte Carlo cycle simulations is set to M, as shown by the substantial fluctuation of real-time power.

[0052] (6) Determine the impact of weather type at time t on the shading of solar photovoltaic cell components of the photovoltaic power station on the output of the solar cell components E1t, mainly precipitation and snow shading, according to the weather type Wt, E1t is directly proportional to Pt.

[0053] (7) Determine the power temperature coefficient change E2t caused by the weather type at time t on the characteristics of the solar photovoltaic cell components of the photovoltaic power station, according to the weather type Wt, in the range of 20-100℃, the voltage of each cell decreases by 2mV for every 1℃ increase; while the current increases slightly with temperature, in general, the power of the solar cell decreases with temperature, the typical power temperature coefficient is -0.35% / ℃, that is, the power decreases by 0.35% for every 1℃ increase in cell temperature.

[0054] (8) Determine the impact of humidity changes caused by the weather type at time t on the power generation of the photovoltaic power station E3t, according to the weather type Wt, E3t is directly negatively related to Pt.

[0055] (9) Determine the impact of destructive factors caused by the weather type at time t on the power attenuation of the solar photovoltaic cell components of the photovoltaic power station E4t, according to the weather type Wt, E4t is directly proportional to Pt.

[0056] (10) Determine the impact of planned grid connection and routine maintenance on the output of the photovoltaic power station Mt, which is directly subtracted from Pt.

[0057] (11) Repeat steps (5) to (10) to obtain the photovoltaic power station power Pt at time t in the unit area complex weather duration T;

[0058] (12) Let t = t + 1, repeat steps (4) to (11), superimpose the unit area photovoltaic power station power prediction curve for a total of T, and obtain the photovoltaic power station power prediction curve graph in the unit area complex weather duration T.

[0059] (13), the superposition of the power prediction curves of all unit areas S1, S2, S3… in the region can obtain the power prediction curve of the photovoltaic power station in the region in a certain complex weather duration T. Since the region contains multiple unit areas, when some unit areas are in complex weather and other unit areas are in sunny weather, the Monte Carlo algorithm is still applicable, that is, the weather type has no effect on the shading of the solar photovoltaic cell components of the photovoltaic power station, and the weather type has no destructive effect on the power attenuation of the solar photovoltaic cell components of the photovoltaic power station.

[0060] (14), the unit area is in complex weather (strong wind, cooling and rain) all day, so the complex weather duration T covers the whole day; since the Monte Carlo algorithm is based on 15-minute interval unit area weather forecast, it can accurately grasp the fluctuation of low-altitude thick clouds on solar radiation intensity in the short term, so it can more accurately predict the effect of weather type Wt on the total amount of solar radiation on the ground St in the short term; in the medium and long term, it is impossible to predict the change of low-altitude thick clouds in the unit area according to the existing meteorological system technical conditions, so the medium and long term St cannot be accurately predicted, that is, the more accurate power fluctuation curve of the photovoltaic power station cannot be predicted, E1t, E2t, E3t and E4t can be more accurately predicted according to the 15-minute interval weather forecast; the accuracy calculation formula is (YCt is the predicted power, SJt is the actual power, and PT is the installed capacity of the photovoltaic power station), and the accuracy of the neural network algorithm and the Monte Carlo algorithm can be obtained.

[0061] As Figure 5 , the accuracy of each t time is calculated, that is It can be seen that the accuracy of the neural network algorithm and the Monte Carlo algorithm at different times is compared, and the advantages and disadvantages of the two different algorithms in short-term, medium-term and long-term power prediction are compared. The Monte Carlo algorithm can accurately predict the power change curve fluctuation of the photovoltaic power station affected by the weather type Wt in the short term. In the medium and long term, since it is impossible to predict the change of low-altitude thick clouds in the unit area of the meteorological system, the Monte Carlo algorithm cannot make the power change curve of the photovoltaic power station, but it can make the minute shape curve diagram according to the accurate unit area short time interval weather forecast.

[0062] (15), refer to the attached Figure 6, under the Monte Carlo algorithm, the photovoltaic power station power Pt in the T period at time t is a fixed value under ideal weather conditions, the influence Mt of the planned grid connection and routine maintenance on the output of the photovoltaic power station is also a fixed value, the weather type Wt has an influence St on the total amount of solar radiation on the ground within the range (i, j), that is, the variable St(i, j). Similarly, E1t, E2t, E3t and E4t are also range values, that is, the variable E1t(i, j), the variable E2t(i, j), the variable E3t(i, j) and the variable E4t(i, j) are also range values. When the Monte Carlo cycle M is 1, a set of data St(i1, j1), E1t(i1, j1), E2t(i1, j1), E3t(i1, j1) and E4t(i1, j1) is obtained; when the Monte Carlo cycle M is m, a set of data St(im, jm), E1t(im, jm), E2t(im, jm), E3t(im, jm) and E4t(im, jm) is obtained; until the data St(iM, jM), E1t(iM, jM), E2t(iM, jM), E3t(iM, jM) and E4t(iM, jM) are obtained. The sum of St(i, j), E1t(i, j), E2t(i, j), E3t(i, j) and E4t(i, j) is called Ut(i, j), that is, Ut(i, j) = St(i, j) + E1t(i, j) + E2t(i, j) + E3t(i, j) + E4t(i, j), all Ut(i, j) are counted, and the highest probability U t is selected as the value calculated by the Monte Carlo algorithm, that is, the sum of the five variables of the weather type Wt at time t, the influence St of the total amount of solar radiation on the ground, the influence E1t of the shading of the photovoltaic solar cell module of the photovoltaic power station on the output of the solar cell module, the power temperature coefficient change E2t caused by the characteristics of the photovoltaic solar cell module of the photovoltaic power station, the influence E3t of humidity change on the power generation of the photovoltaic power station, and the attenuation influence E4t of the destructive factors on the power of the photovoltaic solar cell module of the photovoltaic power station. The larger M is, the more accurate Ut is. The Monte Carlo algorithm obtains the unit area photovoltaic power station power at time t as Pt-Ut-Mt.

Claims

1. A method for ultra-short-term prediction of power generation of regional photovoltaic power plants based on the Monte Carlo algorithm, characterized in that... , The steps of this method are as follows: (1) A certain demand area can be divided into several unit areas S1, S2, S3... The weather type in the unit area is considered to be consistent at any time, and the power generation of the photovoltaic power station in the unit area is considered to be the same due to the influence of the weather type. (2) Determine the duration T of complex weather within a certain unit area; (3) Determine the power generation Pt of the photovoltaic power station under ideal weather conditions at a certain time t within the time period T of the unit area; (4) Confirm the weather type Wt at time t within time period T; (5) Set the number of Monte Carlo simulation cycles to M; (6) Determine the various parameters that affect weather type Wt; (7) Determine the impact of planned grid connection and routine maintenance on the output Mt of the photovoltaic power station; (8) Repeat steps (5)-(7) to obtain the photovoltaic power generation Pt at a certain time t within the duration of complex weather in the unit area; (9) Let t = t + 1, repeat steps (2) - (8), and superimpose to obtain the photovoltaic power generation prediction curve of the unit area with a total number of T, and obtain the photovoltaic power generation prediction curve of the unit area during a certain complex weather duration T. (6.1) The parameters for determining the impact of weather type Wt in step (6) include: determining the impact of weather type on the total solar radiation on the Earth's surface St; (6.2) The parameters for determining the influence of weather type Wt in step (6) include: determining the influence of weather type at time t on the output of solar photovoltaic modules of photovoltaic power station due to shading E1t; (6.3) The parameters for determining the influence of weather type Wt in step (6) include: determining the change in power temperature coefficient E2t caused by the weather type at time t on the characteristics of solar photovoltaic cell modules of the photovoltaic power station; (6.4) The parameters for determining the influence of weather type Wt in step (6) include: determining the influence of humidity change caused by weather type at time t on the power generation of photovoltaic power station E3t; (6.5) The parameters for determining the influence of weather type Wt in step (6) include: determining the influence of the destructive factors caused by the weather type at time t on the power attenuation of the solar photovoltaic cell module of the photovoltaic power station, E4t.

Images