Long-term net load prediction method, device and equipment considering distributed photovoltaic, and storage medium

By acquiring and calculating the deterministic output and occlusion factor sequence of distributed photovoltaics, combining weather and electricity load, the problem of failure to consider distributed photovoltaic output in the existing technology is solved, and the accuracy of long-term net load prediction is improved.

CN118944069APending Publication Date: 2024-11-12GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410989841.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing long-term net load prediction fails to effectively consider the output characteristics of distributed photovoltaics, resulting in the inability to accurately predict the new load patterns after a large number of long-term distributed photovoltaics are installed.

Method used

By obtaining the hourly average load, weather forecast data and the predicted installed capacity of distributed photovoltaics in the area to be measured, the deterministic output sequence and occlusion factor sequence of distributed photovoltaics are calculated using the preset electricity load and occlusion factor prediction model, and combined with the prediction use Electric load calculates the net load.

Benefits of technology

The impact of distributed photovoltaic output is fully considered in long-term net load prediction, and the accuracy and reliability of the prediction are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118944069A_ABST
    Figure CN118944069A_ABST
Patent Text Reader

Abstract

The invention discloses a long-term net load prediction method, device and equipment considering distributed photovoltaic, and a storage medium. The method comprises the following steps: obtaining an hour average load of a to-be-measured year of a to-be-measured region; obtaining weather prediction data of a to-be-predicted year, and obtaining a predicted electrical load of the to-be-predicted year according to the weather prediction data, the hourly average load and a preset electrical load prediction model; obtaining the predicted installed capacity of the distributed photovoltaic in the to-be-tested year, and calculating a distributed photovoltaic deterministic output sequence in the to-be-tested year according to the predicted installed capacity and a preset photovoltaic output physical model; inputting the weather prediction data into a preset shielding factor prediction model to obtain a shielding factor prediction sequence of the year to be measured; and according to the shielding factor prediction sequence, the distributed photovoltaic deterministic output sequence and the predicted electrical load, obtaining the net load of the to-be-tested year of the to-be-tested region. According to the invention, the distributed photovoltaic output can be considered when the long-term net load is predicted.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of long-term net load prediction, and in particular to a long-term net load prediction method, device, equipment and storage medium considering distributed photovoltaics. Background Art

[0002] With the development of social economy and technology, electric energy has become one of the indispensable energy sources in people's production and life. Therefore, ensuring a stable supply of electric energy is crucial to the power system. Long-term power demand forecasting plays an important role in power system planning. By predicting the power demand in the next few years to decades, the planning department can determine how many power plants need to be built and their types (such as coal-fired power, hydropower, nuclear power or renewable energy) to ensure the reliability and economy of power supply. Secondly, by analyzing the growth and changing trends of future power demand in various regions, more reasonable and efficient regional and inter-regional transmission networks can be planned to reduce network losses and improve the overall efficiency of the power system.

[0003] With the promotion of distributed photovoltaics in China, the future load pattern will be very different from the current one. At present, China's distributed photovoltaic installations are still relatively small, and the impact of distributed photovoltaics on the load in the regional historical load is also relatively small. Data-driven methods are difficult to capture the output characteristics of distributed photovoltaics and cannot predict the new load pattern after a large number of distributed photovoltaics are installed in the long term. Therefore, the existing long-term net load forecast does not take into account the output of distributed photovoltaics.

[0004] Therefore, taking distributed PV output into account when forecasting long-term net load is an urgent issue that needs to be addressed. Summary of the invention

[0005] The present invention provides a method, device, equipment and storage medium for predicting long-term net load considering distributed photovoltaics, which can take distributed photovoltaic output into consideration when predicting long-term net load.

[0006] An embodiment of the present invention provides a long-term net load prediction method considering distributed photovoltaics, including:

[0007] Get the hourly average load of the region to be tested in the year to be tested;

[0008] Obtaining weather forecast data for the year to be tested, and obtaining a predicted power load for the year to be tested based on the weather forecast data, the hourly average load, and a preset power load prediction model;

[0009] Obtain the predicted installed capacity of distributed photovoltaics in the above-mentioned year to be tested, and calculate the distributed photovoltaic deterministic output sequence in the above-mentioned year to be tested based on the above-mentioned predicted installed capacity and a preset photovoltaic output physical model;

[0010] Input the above weather forecast data into a preset shading factor prediction model to obtain a shading factor prediction sequence for the above test year;

[0011] Based on the above-mentioned shading factor prediction sequence, the above-mentioned distributed photovoltaic deterministic output sequence and the predicted electricity load, the net load of the test area in the test year is calculated.

[0012] Furthermore, the above-mentioned obtaining of the hourly average load of the region to be tested in the year to be tested includes:

[0013] Obtain the predicted annual electricity consumption of the area to be tested in the year to be tested and the number of hours in the year in which the predicted annual electricity consumption occurs;

[0014] Based on the quotient of the above predicted annual electricity consumption and the above number of hours, the hourly average load of the above year to be tested is obtained.

[0015] Furthermore, the weather forecast data of the year to be tested is obtained to obtain the predicted power load of the year to be tested according to the weather forecast data, the hourly average load and the preset power load prediction model, including:

[0016] Input the above weather forecast data into the above preset power load forecast model to obtain the normalized forecast power load for the above test year;

[0017] The predicted electricity load for the above-mentioned year to be tested is calculated based on the product of the above-mentioned hourly average load and the normalized predicted electricity load;

[0018] The training of the preset power load forecasting model includes:

[0019] Acquire a plurality of first sample data with first true values ​​within a period of historical time in the area to be tested; wherein the first sample data includes: calendar time information and first historical weather data corresponding to the calendar time information; the first true value is used to represent the normalized hourly actual power load within the period of historical time in the area to be tested;

[0020] For each first sample data, the first sample data is input into the power load prediction model to be trained to obtain the normalized predicted power load corresponding to each first sample data, and the value of the first loss function is calculated based on the normalized predicted power load and the first true value;

[0021] After each calculation to obtain a value of the first loss function, it is determined whether the first loss function converges. If so, it is determined that the training of the power load prediction model to be trained is completed. If not, the training of the power load prediction model to be trained continues.

[0022] Furthermore, the acquisition of the normalized hourly actual power load includes:

[0023] Obtaining second sample data within a period of historical time in the area to be tested; wherein the second sample data includes: net load data, the first historical weather data, and the installed capacity of distributed photovoltaics;

[0024] Inputting the first historical weather data into a preset shielding factor prediction model to obtain a shielding factor corresponding to each of the second sample data within the historical period;

[0025] According to the above-mentioned shielding factor, the above-mentioned installed capacity and the preset photovoltaic output physical model, the distributed photovoltaic output corresponding to each sample data in the above-mentioned historical period is obtained;

[0026] According to the above net load data and the sum of the above distributed photovoltaic output, the hourly actual power load corresponding to each of the above sample data in the above historical period is calculated;

[0027] Divide the above hourly actual power load by year, and add up the number of above hourly actual power loads corresponding to each year to obtain the annual power consumption of each year in the above historical period;

[0028] According to the quotient of the annual power consumption and the number of hours in the year, the hourly average load of the year is obtained, and then the hourly average load corresponding to each year is obtained;

[0029] According to the quotient of the actual power load corresponding to each year and the hourly average load corresponding to each year, the normalized hourly actual power load is obtained;

[0030] The training of the preset occlusion factor prediction model includes:

[0031] Acquire a plurality of second historical weather data with second real values ​​in a historical period of time in any region; wherein the second real values ​​are used to represent the shielding factors corresponding to the second historical weather data;

[0032] For each second historical weather data, input the second historical weather data into the power load prediction model to be trained to obtain the shielding factor corresponding to each second historical weather data, and calculate the value of the second loss function according to the shielding factor and the second true value;

[0033] After each calculation to obtain a value of the second loss function, it is determined whether the second loss function converges. If so, it is determined that the training of the occlusion factor prediction model to be trained is completed. If not, the training of the occlusion factor prediction model to be trained continues.

[0034] Furthermore, the above-mentioned preset photovoltaic output physical model includes:

[0035] The actual photovoltaic output is constructed based on the solar radiation intensity and temperature under standard conditions, the photovoltaic panel output under standard conditions, the power temperature coefficient of the photovoltaic panel, and the actual solar radiation intensity and temperature at a certain moment;

[0036] Based on the solar radiation intensity without considering shading, the solar radiation intensity under standard conditions, and the photovoltaic panel output under standard conditions, the output of the photovoltaic power station without considering shading and temperature influence is constructed;

[0037] Based on the solar constant and the annual sequence of days, the radiation intensity of the sun directly shining on the earth's atmosphere is constructed;

[0038] Construct the ground solar radiation intensity based on the sum of direct solar radiation and scattered solar radiation;

[0039] Construct a photovoltaic shading factor according to the power temperature coefficient of the photovoltaic panel, the actual temperature at time t, the temperature under the standard conditions, the solar radiation intensity without considering shading, the actual solar radiation intensity at time t, the actual photovoltaic output, and the output of the photovoltaic power station without considering shading and temperature influence.

[0040] The above-mentioned preset photovoltaic output physical model is obtained based on the above-mentioned actual photovoltaic output, the output of the photovoltaic power station without considering the influence of shielding and temperature, the radiation intensity of the sun directly shining on the earth's atmosphere, the ground solar radiation intensity and the photovoltaic shielding factor.

[0041] Furthermore, the net load of the area to be tested in the year to be tested is calculated based on the above-mentioned shading factor prediction sequence, the above-mentioned distributed photovoltaic deterministic output sequence and the predicted power load, including:

[0042] The predicted distributed photovoltaic output is calculated based on the product of the above-mentioned shading factor prediction sequence and the above-mentioned distributed photovoltaic deterministic output sequence;

[0043] The above net load is calculated based on the difference between the above predicted electricity load and the above predicted distributed photovoltaic output.

[0044] Furthermore, after obtaining the above net load, it also includes:

[0045] Obtain the preset net load of the above-mentioned area to be tested in the year to be tested;

[0046] The above-mentioned preset net load is compared with the above-mentioned net load, and it is determined whether the above-mentioned net load is within the range of the above-mentioned preset net load. If not, the minimum net load difference when the above-mentioned net load meets the above-mentioned preset net load range is calculated, and the predicted installed capacity of the above-mentioned distributed photovoltaic is adjusted according to the above-mentioned minimum net load difference.

[0047] Based on the above method embodiment, the present invention provides a corresponding device embodiment;

[0048] The present invention provides a long-term net load prediction device considering distributed photovoltaics, comprising:

[0049] Hourly average load acquisition module, power load prediction module, distributed photovoltaic deterministic output sequence calculation module, shading factor prediction sequence calculation module and net load calculation module;

[0050] The above-mentioned hourly average load acquisition module is used to obtain the hourly average load of the area to be tested in the year to be tested;

[0051] The power load prediction module is used to obtain the weather forecast data of the year to be tested, so as to obtain the predicted power load of the year to be tested according to the weather forecast data, the hourly average load and the preset power load prediction model;

[0052] The above-mentioned distributed photovoltaic deterministic output sequence calculation module is used to obtain the predicted installed capacity of the distributed photovoltaic in the above-mentioned year to be tested, so as to calculate the distributed photovoltaic deterministic output sequence in the above-mentioned year to be tested based on the above-mentioned predicted installed capacity and the preset photovoltaic output physical model;

[0053] The above-mentioned shading factor prediction sequence calculation module is used to input the above-mentioned weather forecast data into a preset shading factor prediction model to obtain the shading factor prediction sequence of the above-mentioned year to be tested;

[0054] The net load calculation module is used to calculate the net load of the area to be tested in the year to be tested based on the above-mentioned shading factor prediction sequence, the above-mentioned distributed photovoltaic deterministic output sequence and the predicted power load.

[0055] Based on the above method embodiment, the present invention provides a terminal device embodiment;

[0056] The present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, a long-term net load prediction method considering distributed photovoltaics described in any embodiment of the present invention is implemented.

[0057] Based on the above method embodiment, the present invention provides a storage medium embodiment;

[0058] The present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute a long-term net load prediction method considering distributed photovoltaics as described in any embodiment of the present invention.

[0059] The embodiments of the present invention have the following beneficial effects:

[0060] The present invention provides a long-term net load prediction method, device, equipment and storage medium considering distributed photovoltaics. The method first obtains the hourly average load of the above-mentioned area to be tested in the year to be tested; then obtains the weather forecast data of the above-mentioned year to be tested, and obtains the predicted electricity load of the above-mentioned year to be tested based on the above-mentioned weather forecast data, the above-mentioned hourly average load and a preset electricity load prediction model; then obtains the predicted installed capacity of distributed photovoltaics in the above-mentioned year to be tested, and calculates the distributed photovoltaic deterministic output sequence of the above-mentioned year to be tested based on the above-mentioned predicted installed capacity and a preset photovoltaic output physical model; then inputs the above-mentioned weather forecast data into a preset shading factor prediction model to obtain the shading factor prediction sequence of the above-mentioned year to be tested; finally, according to the above-mentioned shading factor prediction sequence, the above-mentioned distributed photovoltaic deterministic output sequence and the predicted electricity load, the net load of the area to be tested in the year to be tested is calculated. Therefore, the present invention utilizes the predicted installed capacity and trained photovoltaic output physical model to obtain the distributed photovoltaic deterministic output sequence, and then predicts the long-term net load based on the above distributed photovoltaic deterministic output sequence and the shading factor prediction sequence. This scheme fully considers the predicted impact of distributed photovoltaic output on future long-term net load. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a flow chart of a method for long-term net load prediction considering distributed photovoltaics provided by an embodiment of the present invention.

[0062] Figure 2 It is a schematic diagram of the structure of a device for predicting long-term net load of distributed photovoltaics provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] like Figure 1As shown, an embodiment of the present invention provides a method for predicting long-term net load considering distributed photovoltaics, including:

[0065] Step S101: obtaining the hourly average load of the region to be tested in the year to be tested;

[0066] In a preferred embodiment, the above-mentioned obtaining the hourly average load of the region to be tested in the year to be tested includes:

[0067] Obtain the predicted annual electricity consumption of the area to be tested in the year to be tested and the number of hours in the year in which the predicted annual electricity consumption occurs;

[0068] Based on the quotient of the above predicted annual electricity consumption and the above number of hours, the hourly average load of the above year to be tested is obtained.

[0069] Specifically, the predicted annual electricity consumption can be obtained from the existing annual electricity consumption prediction results for the area to be tested. If the year to be tested is a normal year, the number of hours is 8760, and if the year to be tested is a leap year, the number of hours is 8784.

[0070] In this preferred embodiment, the predicted annual electricity consumption and hours of the above-mentioned area in the year to be tested are obtained from existing data, and the hourly average load of the above-mentioned year to be tested is obtained based on the quotient of the above-mentioned predicted annual electricity consumption and the above-mentioned hours.

[0071] Step S102: obtaining weather forecast data for the year to be tested, and obtaining a predicted power load for the year to be tested based on the weather forecast data, the hourly average load, and a preset power load prediction model;

[0072] Specifically, there is a lot of research and open source data on long-term weather forecasts. For example, a study by the National Aeronautics and Space Administration (NASA) provides global weather forecast data until 2100, so the above weather forecast data can be obtained from it.

[0073] In a preferred embodiment, the weather forecast data of the year to be tested is obtained to obtain the predicted power load of the year to be tested according to the weather forecast data, the hourly average load and a preset power load prediction model, including:

[0074] Input the above weather forecast data into the above preset power load forecast model to obtain the normalized forecast power load for the above test year;

[0075] The predicted electricity load for the above-mentioned year to be tested is calculated based on the product of the above-mentioned hourly average load and the normalized predicted electricity load;

[0076] The training of the preset power load forecasting model includes:

[0077] Acquire a plurality of first sample data with first true values ​​within a period of historical time in the area to be tested; wherein the first sample data includes: calendar time information and first historical weather data corresponding to the calendar time information; the first true value is used to represent the normalized hourly actual power load within the period of historical time in the area to be tested;

[0078] For each first sample data, the first sample data is input into the power load prediction model to be trained to obtain the normalized predicted power load corresponding to each first sample data, and the value of the first loss function is calculated based on the normalized predicted power load and the first true value;

[0079] After each calculation to obtain a value of the first loss function, it is determined whether the first loss function converges. If so, it is determined that the training of the power load prediction model to be trained is completed. If not, the training of the power load prediction model to be trained continues.

[0080] Specifically, the data set of the first sample data is D1={w t ,s t ,r t}, where w t represents the first historical weather data, which is a vector, each item of which represents a weather index, including temperature, humidity, wind speed, cloud conditions and rainfall; s t Represents the above calendar time information, which is a vector, each item of which represents a calendar information, including: month, day of the week (Monday to Sunday) and hour; r t Represents the first real value mentioned above, which is a scalar.

[0081] Specifically, the first sample data set is divided into a training set and a validation set in a ratio of 7:3. After the power load prediction model is trained, the validation set is used to evaluate its accuracy.

[0082] Preferably, the normalized hourly actual power load is used as the first true value, which can remove the dimension of the hourly actual power load and facilitate subsequent calculation and analysis. The above power load forecasting model can use any statistical or machine learning model as a basic forecasting model.

[0083] In this preferred embodiment, the electricity load prediction model is trained to obtain a trained electricity load prediction model, and then the weather forecast data is input into the above-mentioned electricity load prediction model to obtain the normalized predicted electricity load for the year to be tested, and then the product of the hourly average load and the normalized predicted electricity load is calculated to obtain the predicted electricity load for the year to be tested.

[0084] In another preferred embodiment, the acquisition of the normalized hourly actual power load includes:

[0085] Obtaining second sample data within a period of historical time in the area to be tested; wherein the second sample data includes: net load data, the first historical weather data, and the installed capacity of distributed photovoltaics;

[0086] Specifically, the data set of the second sample data is h t {t,w′ t ,l t ,C t}, where t represents the time of the second sample data; w′ t represents the first historical weather data at time t, the first historical weather data including: temperature, humidity, wind speed, cloud conditions and rainfall; l t represents the above net load data at time t; C t It represents the installed capacity of the above distributed photovoltaic at time t, which can be obtained through statistics or estimation. In the photovoltaic output physical model, the installed capacity of distributed photovoltaic corresponds to the photovoltaic panel output under standard conditions.

[0087] Inputting the first historical weather data into a preset shielding factor prediction model to obtain a shielding factor corresponding to each of the second sample data within the historical period;

[0088] Preferably, the above-mentioned shielding factor describes the ratio of the actual photovoltaic output to the ideal clear sky output under the local weather conditions at that time. Existing relevant studies have shown that this ratio has nothing to do with the geographical location (latitude and longitude) and time, but only with the weather conditions (temperature and humidity, clouds, rainfall, etc.).

[0089] According to the above-mentioned shielding factor, the above-mentioned installed capacity and the preset photovoltaic output physical model, the distributed photovoltaic output corresponding to each sample data in the above-mentioned historical period is obtained;

[0090] According to the above net load data and the sum of the above distributed photovoltaic output, the hourly actual power load corresponding to each of the above sample data in the above historical period is calculated;

[0091] Divide the above hourly actual power load by year, and add up the number of above hourly actual power loads corresponding to each year to obtain the annual power consumption of each year in the above historical period;

[0092] For example, if there are M years of actual hourly power load L real , then after dividing by year, we can get L real,1 , L real,2 …Lreal,M For each year, the annual electricity consumption E of the region is calculated based on the hourly load of the year. i , E i Divide by the total number of hours in the year n (8760 in a normal year or 8784 in a leap year), we can get the hourly average load l for the year. avg,i Then, the actual hourly electricity load L of each year is calculated. real,i Divide by the hourly average load of the year l avg,i , and obtain the first true value R={r t}, t∈T.

[0093] Specifically, since the unit of electricity consumption is kilowatt-hour, the numerical value can be obtained by directly adding up the hourly loads of the year.

[0094] According to the quotient of the annual power consumption and the number of hours in the year, the hourly average load of the year is obtained, and then the hourly average load corresponding to each year is obtained;

[0095] According to the quotient of the actual power load corresponding to each year and the hourly average load corresponding to each year, the normalized hourly actual power load is obtained;

[0096] The training of the preset occlusion factor prediction model includes:

[0097] Acquire a plurality of second historical weather data with second real values ​​in a historical period of time in any region; wherein the second real values ​​are used to represent the shielding factors corresponding to the second historical weather data;

[0098] Specifically, the above-mentioned second historical weather data with second true values ​​can be regarded as a third sample data set D2{w″ t ,η t}, where w″ t represents the second historical weather data, which is a scalar, each of which is a weather indicator; η t Represents the above occlusion factor, which is a scalar and the true value of the occlusion factor prediction model training process. D2 is divided into training set and validation set in a ratio of 7:3.

[0099] For each second historical weather data, input the second historical weather data into the power load prediction model to be trained to obtain the shielding factor corresponding to each second historical weather data, and calculate the value of the second loss function according to the shielding factor and the second true value;

[0100] After each calculation to obtain a value of the second loss function, it is determined whether the second loss function converges. If so, it is determined that the training of the occlusion factor prediction model to be trained is completed. If not, the training of the occlusion factor prediction model to be trained continues.

[0101] Specifically, after the occlusion factor prediction model is trained, the accuracy of the trained model is evaluated and verified using a validation set. The occlusion factor prediction model can use any statistical or machine learning model as a basic prediction model, such as linear regression, XGBoost or MLP.

[0102] In this preferred embodiment, the second sample data is first obtained, and the first historical weather data in the second sample data is input into a preset shading factor prediction model to obtain the shading factor corresponding to each second sample data in the historical period. Subsequently, the installed capacity, net load data and the obtained shading factor in the second sample data are used to calculate with a preset photovoltaic output physical model to obtain the hourly actual power load. Then, the hourly actual power load is divided by year, and several hourly actual power loads corresponding to each year are added together to calculate the annual power consumption of each year in the historical period. Then, the hourly average load of the year is obtained based on the quotient of the annual power consumption and the number of hours in the year, and then the hourly average load corresponding to each year is obtained. Finally, the normalized hourly actual power load is obtained based on the quotient of the actual power load corresponding to each year and the hourly average load corresponding to each year. Among them, the training of the shading factor prediction model utilizes the characteristic that the shading factor is only related to the weather conditions, obtains the second historical weather data and the shading factor corresponding to each second historical weather data as the true value of the model, inputs the above second historical weather data into the electricity load prediction model to be trained, trains the model, and finally obtains the trained electricity load prediction model.

[0103] In another preferred embodiment, the above-mentioned preset photovoltaic output physical model includes:

[0104] The actual photovoltaic output is constructed based on the solar radiation intensity and temperature under standard conditions, the photovoltaic panel output under standard conditions, the power temperature coefficient of the photovoltaic panel, and the actual solar radiation intensity and temperature at a certain moment;

[0105] Specifically, the actual photovoltaic output is constructed according to the following formula:

[0106]

[0107] Where P t represents the actual photovoltaic output, P stc It represents the output of photovoltaic panels under the above standard conditions and is generally used to describe installed capacity.r,t represents the actual solar radiation intensity at time t, T t represents the actual temperature at time t, α T It represents the power temperature coefficient of the photovoltaic panel, which is an inherent property of the photovoltaic panel and its typical value is α T =-0.004 / ℃, I stc represents the solar radiation intensity under the above standard conditions, and I stc =1000W / m 2 , T stc represents the temperature under the above standard conditions, and T stc =25℃.

[0108] Based on the solar radiation intensity without considering shading, the solar radiation intensity under standard conditions, and the photovoltaic panel output under standard conditions, the output of the photovoltaic power station without considering shading and temperature influence is constructed;

[0109] Preferably, the factors that affect the actual output of the photovoltaic panels are solar radiation intensity and temperature, so the deterministic factors and uncertain factors in the photovoltaic output are separated to obtain a definition of the output of the photovoltaic power station without considering the effects of shading and temperature.

[0110] Specifically, the output of the photovoltaic power station without considering the influence of shading and temperature is constructed according to the following formula:

[0111]

[0112] Where P c,t It represents the output of the photovoltaic power station without considering the influence of shading and temperature. t It represents the solar radiation intensity without considering the above-mentioned shading, in W / m 2 This value is only related to the longitude and latitude of the area, altitude and time. Its changing pattern is certain and it is the deterministic part of photovoltaic output.

[0113] Based on the solar constant and the annual sequence of days, the radiation intensity of the sun directly shining on the earth's atmosphere is constructed;

[0114] Preferably, since the radiation intensity of the sun directly shining on the earth's atmosphere is only related to the relative position between the sun and the earth, a formula for constructing the radiation intensity of the sun directly shining on the earth's atmosphere can be obtained.

[0115] Specifically, the radiation intensity of the sun directly shining on the earth's atmosphere is constructed according to the following formula:

[0116]

[0117] Wherein, I0 represents the radiation intensity of the sun directly shining on the earth's atmosphere, S0 represents the solar constant, and N represents the day sequence in the year (the value ranges from 1 to 365, and the leap year value ranges from 1 to 366).

[0118] Construct the ground solar radiation intensity based on the sum of direct solar radiation and scattered solar radiation;

[0119] Specifically, for the above-mentioned direct solar radiation, the first influencing factor is the local solar altitude angle at that time, and the second influencing factor is the transparency of the direct solar radiation.

[0120] Specifically, the ground solar radiation intensity is constructed according to the following formula:

[0121] I t =I b,t +I d,t ;

[0122] In the formula, I b,t represents the above direct solar radiation, I d,t represents the above-mentioned solar diffuse radiation.

[0123] Construct a photovoltaic shading factor according to the power temperature coefficient of the photovoltaic panel, the actual temperature at time t, the temperature under the standard conditions, the solar radiation intensity without considering shading, the actual solar radiation intensity at time t, the actual photovoltaic output, and the output of the photovoltaic power station without considering shading and temperature influence.

[0124] Specifically, the photovoltaic shading factor is constructed according to the following formula:

[0125]

[0126] Where η t represents the above photovoltaic shading factor.

[0127] Specifically, it can be seen that the photovoltaic shielding factor describes the ratio of the actual photovoltaic output to the ideal clear sky unobstructed output under the local weather conditions at that time.

[0128] The above-mentioned preset photovoltaic output physical model is obtained based on the above-mentioned actual photovoltaic output, the output of the photovoltaic power station without considering the influence of shielding and temperature, the radiation intensity of the sun directly shining on the earth's atmosphere, the ground solar radiation intensity and the photovoltaic shielding factor.

[0129] Specifically, the solar altitude angle is calculated based on longitude, latitude, solar declination, local solar hour angle and Beijing time during the day:

[0130]

[0131] sinα=sinΦsinδ+cosΦcosω;

[0132] Where ω represents the local hour angle of the sun, t BJ represents Beijing time on the above day, θ represents the above longitude, δ represents the above solar declination angle, Φ represents the above latitude, and α represents the above solar altitude angle.

[0133] Specifically, the transparency of direct solar radiation is calculated based on the atmospheric quality, altitude and solar altitude angle:

[0134]

[0135] Where M h represents the above-mentioned atmospheric quality, which is an analytical function related to the local altitude and the solar altitude angle, h represents the above-mentioned altitude, τ b Representing the above-mentioned transparency of direct solar radiation, the transparency of the atmosphere can be used to describe the attenuation of solar radiation by the atmosphere and depends on the local atmospheric quality.

[0136] Specifically, the direct solar radiation is calculated based on the above-mentioned radiation intensity of the sun directly shining on the earth's atmosphere, the above-mentioned direct solar radiation transparency and the above-mentioned solar altitude angle:

[0137] I b =I0τ b sinα;

[0138] In the formula, I b Represents the above direct solar radiation.

[0139] Specifically, the solar scattered radiation is calculated based on the solar scattered radiation transparency, the above solar direct radiation transparency, the air quality, the parameters related to the air quality, the radiation intensity of the sun directly shining on the earth's atmosphere and the above solar altitude angle:

[0140] τ d =0.271-0.274τ b ;

[0141]

[0142] In the formula, τ d represents the above-mentioned solar scattered radiation transparency, I d represents the above-mentioned solar scattered radiation, k represents the above-mentioned atmospheric quality related parameters, and its value range is [0.6, 0.9]. It takes a lower value when the atmosphere is turbid and a higher value when the quality is good. In this scheme, a fixed value of 0.7 is taken here.

[0143] In this preferred embodiment, the photovoltaic output physical model is obtained based on the actual photovoltaic output, the output of the photovoltaic power station without considering the influence of shielding and temperature, the radiation intensity of the sun directly shining on the earth's atmosphere, the ground solar radiation intensity and the photovoltaic shielding factor. This model can then be used to solve the hourly actual power load corresponding to each of the above sample data in the above historical period according to the above shielding factor, the above installed capacity and the above net load data in the previous real-time example.

[0144] Step S103: obtaining the predicted installed capacity of distributed photovoltaics in the above-mentioned year to be tested, and calculating the distributed photovoltaic deterministic output sequence in the above-mentioned year to be tested according to the above-mentioned predicted installed capacity and a preset photovoltaic output physical model;

[0145] Specifically, the above-mentioned predicted installed capacity is the photovoltaic panel output under the above-mentioned standard conditions. Therefore, the distributed photovoltaic deterministic output sequence of the above-mentioned test year can be obtained by This formula is obtained.

[0146] In this step, by obtaining the predicted installed capacity of distributed photovoltaics in the year to be tested, the deterministic output sequence of distributed photovoltaics in the year to be tested is obtained according to the predicted installed capacity and the preset photovoltaic output physical model.

[0147] Step S104: inputting the above weather forecast data into a preset shading factor prediction model to obtain a shading factor prediction sequence for the above year to be tested;

[0148] Specifically, the weather forecast data is input into a preset shielding factor prediction model for prediction to obtain a shielding factor prediction sequence for the year to be tested.

[0149] In this step, the above weather forecast data is input into a preset shading factor prediction model for prediction, thereby obtaining a shading factor prediction sequence for the year to be tested.

[0150] Step S105: Calculate the net load of the area to be tested in the year to be tested based on the above-mentioned shading factor prediction sequence, the above-mentioned distributed photovoltaic deterministic output sequence and the predicted power load.

[0151] In a preferred embodiment, the net load of the area to be tested in the year to be tested is calculated based on the above-mentioned shading factor prediction sequence, the above-mentioned distributed photovoltaic deterministic output sequence and the predicted power load, including:

[0152] The predicted distributed photovoltaic output is calculated based on the product of the above-mentioned shading factor prediction sequence and the above-mentioned distributed photovoltaic deterministic output sequence;

[0153] The above net load is calculated based on the difference between the above predicted electricity load and the above predicted distributed photovoltaic output.

[0154] Specifically, the above-mentioned shading factor prediction sequence is H, and the above-mentioned distributed photovoltaic deterministic output sequence is P c , the above predicted power load is L pred , so by calculating H×P c This formula gives the predicted distributed photovoltaic output P pred , and then calculate L pred -P pred This formula gives the net load L net .

[0155] In this preferred embodiment, the predicted distributed photovoltaic output is obtained by multiplying the shielding factor prediction sequence and the distributed photovoltaic deterministic output sequence, and the net load is obtained by calculating the difference between the predicted power load and the predicted distributed photovoltaic output.

[0156] In another preferred embodiment, after obtaining the above-mentioned net load, the method further comprises:

[0157] Obtain the preset net load of the above-mentioned area to be tested in the year to be tested;

[0158] The above-mentioned preset net load is compared with the above-mentioned net load, and it is determined whether the above-mentioned net load is within the range of the above-mentioned preset net load. If not, the minimum net load difference when the above-mentioned net load meets the above-mentioned preset net load range is calculated, and the predicted installed capacity of the above-mentioned distributed photovoltaic is adjusted according to the above-mentioned minimum net load difference.

[0159] Specifically, the calculated net load is compared with the preset net load. If the calculated net load is not within the range of the preset net load, the minimum difference when the net load satisfies the preset net load range is calculated, and the predicted installed capacity is adjusted according to the difference.

[0160] In this preferred embodiment, by calculating the minimum difference between the preset net load range and the preset net load range when the preset net load range is not met, the predicted installed capacity is adjusted to make the setting of the installed capacity more reasonable.

[0161] Based on the above method embodiment, the present invention provides a corresponding device embodiment.

[0162] like Figure 2 As shown, an embodiment of the present invention provides a long-term net load prediction device considering distributed photovoltaics, including:

[0163] Hourly average load acquisition module, power load prediction module, distributed photovoltaic deterministic output sequence calculation module, shading factor prediction sequence calculation module and net load calculation module;

[0164] The above-mentioned hourly average load acquisition module is used to obtain the hourly average load of the area to be tested in the year to be tested;

[0165] The power load prediction module is used to obtain the weather forecast data of the year to be tested, so as to obtain the predicted power load of the year to be tested according to the weather forecast data, the hourly average load and the preset power load prediction model;

[0166] The above-mentioned distributed photovoltaic deterministic output sequence calculation module is used to obtain the predicted installed capacity of the distributed photovoltaic in the above-mentioned year to be tested, so as to calculate the distributed photovoltaic deterministic output sequence in the above-mentioned year to be tested based on the above-mentioned predicted installed capacity and the preset photovoltaic output physical model;

[0167] The above-mentioned shading factor prediction sequence calculation module is used to input the above-mentioned weather forecast data into a preset shading factor prediction model to obtain the shading factor prediction sequence of the above-mentioned year to be tested;

[0168] The net load calculation module is used to calculate the net load of the area to be tested in the year to be tested based on the above-mentioned shading factor prediction sequence, the above-mentioned distributed photovoltaic deterministic output sequence and the predicted power load.

[0169] It should be noted that the device embodiments described above are merely schematic, wherein the modules described above as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement it without creative work. The above schematic diagram is only an example of a long-term net load prediction device considering distributed photovoltaics, and does not constitute a limitation on a long-term net load prediction device considering distributed photovoltaics, and may include more or fewer components than shown in the figure, or a combination of certain components, or different components.

[0170] Based on the above method item embodiments, the present invention provides corresponding terminal device item embodiments.

[0171] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements a long-term net load forecasting method considering distributed photovoltaics as described in any embodiment of the present invention when executing the computer program.

[0172] Exemplarily, in this embodiment, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the device;

[0173] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The device may include, but is not limited to, a processor, a memory;

[0174] The processor referred to herein may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The above-mentioned processor is the control center of the above-mentioned device, and uses various interfaces and lines to connect various parts of the entire device;

[0175] The above-mentioned memory can be used to store the above-mentioned computer programs and / or modules. The above-mentioned processor realizes various functions of the above-mentioned device by running or executing the computer programs and / or modules stored in the above-mentioned memory, and calling the data stored in the memory. The above-mentioned memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; in addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0176] Based on the above method item embodiments, the present invention provides a corresponding storage medium item embodiment.

[0177] Another embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute a long-term net load prediction method considering distributed photovoltaics as described in any embodiment of the present invention.

[0178] In this embodiment, the storage medium is a computer-readable storage medium, the computer program includes computer program code, and the computer program code may be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0179] Compared with the prior art, by implementing the above-mentioned embodiments of the present invention, the distributed photovoltaic output can be taken into account when predicting the long-term net load.

[0180] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A long-term net load forecasting method considering distributed photovoltaics, characterized in that: include: Get the hourly average load of the region to be tested in the year to be tested; Obtaining weather forecast data for the year to be tested, and obtaining a predicted power load for the year to be tested based on the weather forecast data, the hourly average load, and a preset power load prediction model; Obtaining the predicted installed capacity of distributed photovoltaics in the year to be tested, and calculating the distributed photovoltaic deterministic output sequence in the year to be tested according to the predicted installed capacity and a preset photovoltaic output physical model; Inputting the weather forecast data into a preset shading factor prediction model to obtain a shading factor prediction sequence for the year to be tested; The net load of the area to be tested in the year to be tested is calculated based on the shading factor prediction sequence, the distributed photovoltaic deterministic output sequence and the predicted power load.

2. A method for long-term net load prediction considering distributed photovoltaics according to claim 1, characterized in that: The step of obtaining the hourly average load of the region to be tested in the year to be tested includes: Obtain the predicted annual electricity consumption of the area to be tested in the year to be tested and the number of hours in the year in which the predicted annual electricity consumption occurs; The hourly average load of the year to be measured is obtained according to the quotient of the predicted annual electricity consumption and the number of hours.

3. A long-term net load forecasting method considering distributed photovoltaics according to claim 2, characterized in that: The obtaining of the weather forecast data for the year to be tested, and obtaining the predicted power load for the year to be tested according to the weather forecast data, the hourly average load and a preset power load prediction model, comprises: Inputting the weather forecast data into the preset power load forecast model to obtain the normalized forecast power load for the year to be tested; The predicted electricity load for the year to be tested is calculated based on the product of the hourly average load and the normalized predicted electricity load; The training of the preset power load prediction model includes: Acquire a plurality of first sample data with first true values ​​within a period of historical time in the area to be tested; wherein the first sample data includes: calendar time information and first historical weather data corresponding to the calendar time information; the first true value is used to represent the normalized hourly actual power load within a period of historical time in the area to be tested; For each first sample data, the first sample data is input into the power load prediction model to be trained to obtain the normalized predicted power load corresponding to each first sample data, and the value of the first loss function is calculated according to the normalized predicted power load and the first true value; After each calculation to obtain a value of the first loss function, it is determined whether the first loss function converges. If so, it is determined that the training of the power load prediction model to be trained is completed. If not, the training of the power load prediction model to be trained continues.

4. A long-term net load forecasting method considering distributed photovoltaics according to claim 3, characterized in that: The acquisition of the normalized hourly actual power load includes: Acquire second sample data of a period of historical time in the area to be tested; wherein the second sample data includes: net load data, the first historical weather data, and the installed capacity of distributed photovoltaics; Inputting the first historical weather data into a preset shading factor prediction model to obtain the shading factor corresponding to each of the second sample data within the historical period; According to the shielding factor, the installed capacity and the preset photovoltaic output physical model, the distributed photovoltaic output corresponding to each sample data within the historical period is obtained; According to the net load data and the sum of the distributed photovoltaic output, the hourly actual power load corresponding to each sample data in the historical period is calculated; Divide the hourly actual power load by year, and add up a number of the hourly actual power loads corresponding to each year to obtain the annual power consumption of each year in the historical period; According to the quotient of the annual power consumption and the number of hours in the year, the hourly average load of the year is obtained, and then the hourly average load corresponding to each year is obtained; Obtaining a normalized hourly actual power load according to the quotient of the actual power load corresponding to each year and the hourly average load corresponding to each year; The training of the preset occlusion factor prediction model includes: Acquire a plurality of second historical weather data with second real values ​​in a historical period of time in any region; wherein the second real values ​​are used to represent each shielding factor corresponding to each second historical weather data; For each second historical weather data, input the second historical weather data into the power load prediction model to be trained to obtain the shielding factor corresponding to each second historical weather data, and calculate the value of the second loss function according to the shielding factor and the second true value; After each calculation to obtain a value of the second loss function, it is determined whether the second loss function converges. If so, it is determined that the training of the occlusion factor prediction model to be trained is completed. If not, the training of the occlusion factor prediction model to be trained continues.

5. A method for long-term net load prediction considering distributed photovoltaics according to claim 4, characterized in that: The preset photovoltaic output physical model includes: The actual photovoltaic output is constructed based on the solar radiation intensity and temperature under standard conditions, the photovoltaic panel output under standard conditions, the power temperature coefficient of the photovoltaic panel, and the actual solar radiation intensity and temperature at a certain moment; Based on the solar radiation intensity without considering shading, the solar radiation intensity under standard conditions, and the photovoltaic panel output under standard conditions, the output of the photovoltaic power station without considering shading and temperature influence is constructed; Based on the solar constant and the annual sequence of days, the radiation intensity of the sun directly shining on the earth's atmosphere is constructed; Construct the ground solar radiation intensity based on the sum of direct solar radiation and scattered solar radiation; Constructing a photovoltaic shading factor according to the power temperature coefficient of the photovoltaic panel, the actual temperature at time t, the temperature under the standard conditions, the solar radiation intensity without considering the shading, the actual solar radiation intensity at time t, the actual photovoltaic output, and the output of the photovoltaic power station without considering the shading and temperature influence; The preset photovoltaic output physical model is obtained according to the actual photovoltaic output, the output of the photovoltaic power station without considering the influence of shielding and temperature, the radiation intensity of the sun directly shining on the earth's atmosphere, the ground solar radiation intensity and the photovoltaic shielding factor.

6. A method for long-term net load prediction considering distributed photovoltaics according to claim 5, characterized in that: The method of calculating the net load of the area to be tested in the year to be tested according to the shading factor prediction sequence, the distributed photovoltaic deterministic output sequence and the predicted power load includes: The predicted distributed photovoltaic output is calculated according to the product of the shading factor prediction sequence and the distributed photovoltaic deterministic output sequence; The net load is calculated based on the difference between the predicted power load and the predicted distributed photovoltaic output.

7. A method for long-term net load prediction considering distributed photovoltaics according to claim 6, characterized in that: After obtaining the net load, the method further includes: Obtaining a preset net load of the region to be tested in the year to be tested; The preset net load is compared with the net load, and it is determined whether the net load is within the range of the preset net load. If not, the minimum net load difference when the net load meets the preset net load range is calculated, and the predicted installed capacity of the distributed photovoltaic is adjusted according to the minimum net load difference.

8. A long-term net load prediction device considering distributed photovoltaics, characterized in that: include: Hourly average load acquisition module, power load prediction module, distributed photovoltaic deterministic output sequence calculation module, shading factor prediction sequence calculation module and net load calculation module; The hourly average load acquisition module is used to acquire the hourly average load of the area to be tested in the year to be tested; The power load prediction module is used to obtain weather forecast data for the year to be tested, so as to obtain the predicted power load for the year to be tested based on the weather forecast data, the hourly average load and a preset power load prediction model; The distributed photovoltaic deterministic output sequence calculation module is used to obtain the predicted installed capacity of the distributed photovoltaic in the year to be tested, so as to calculate the distributed photovoltaic deterministic output sequence in the year to be tested according to the predicted installed capacity and a preset photovoltaic output physical model; The shading factor prediction sequence calculation module is used to input the weather forecast data into a preset shading factor prediction model to obtain the shading factor prediction sequence of the year to be tested; The net load calculation module is used to calculate the net load of the area to be tested in the year to be tested based on the shielding factor prediction sequence, the distributed photovoltaic deterministic output sequence and the predicted power load.

9. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a long-term net load forecasting method considering distributed photovoltaics as described in any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute a long-term net load prediction method considering distributed photovoltaics as described in any one of claims 1 to 7.