Photovoltaic power generation prediction method, device, electronic equipment and storage medium

Abstract

The present disclosure provides a photovoltaic power prediction method, device, electronic device and storage medium, including: obtaining power data and first power data of a distributed photovoltaic power station collection meter, wherein the first power data is the difference between every two adjacent table code values ​​in the collection meter; determining the second power data of the distributed photovoltaic power station based on the power data, and determining the power error value of the distributed photovoltaic power station based on the first power data and the second power data; filtering the first power data and the second power data based on the power error value to obtain filtered power data; interpolating the filtered power data to obtain interpolated power data; and performing power prediction based on the interpolated power data to obtain prediction results.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, electronic equipment and storage medium for predicting photovoltaic power generation. Background Art

[0002] In order to achieve the dual carbon goals, the installed capacity of photovoltaic power generation has increased year by year in recent years, and the proportion of distributed photovoltaic power station installed capacity has also increased year by year. The power generation of photovoltaic power stations is highly volatile and intermittent, which poses a great challenge to the safety of photovoltaic energy access to the power grid. In actual production, photovoltaic power generation prediction technology is mainly used to improve the safety of photovoltaic energy access to the power grid. Its main principle is to perform mathematical modeling based on the basic information of photovoltaic power stations and historical power generation, and then predict future photovoltaic power.

[0003] In the existing technology, photovoltaic power generation prediction technology is mainly used in centralized photovoltaic power stations with accurate power value collection. It is not applicable to distributed photovoltaic power stations with a large number of power stations, wide distribution, uneven density and poor data quality. Summary of the invention

[0004] In view of this, the purpose of the present disclosure is to provide a photovoltaic power prediction method, device, electronic device and storage medium.

[0005] As one aspect of the present disclosure, a method for predicting photovoltaic power generation is provided, comprising:

[0006] Acquire power data and first electric quantity data of a collection meter of a distributed photovoltaic power station, wherein the first electric quantity data is a difference between every two adjacent meter code values ​​in the collection meter;

[0007] Determine second power data of the distributed photovoltaic power station based on the power data, and determine a power error value of the distributed photovoltaic power station based on the first power data and the second power data;

[0008] Filtering the first power data and the second power data based on the power error value to obtain filtered power data;

[0009] Performing interpolation processing on the filtered power data to obtain interpolated power data;

[0010] Power prediction is performed based on the interpolated electrical quantity data to obtain a prediction result.

[0011] Optionally, determining the second power data of the distributed photovoltaic power station based on the power data includes:

[0012] Determine the number of sampling points of the power data, and determine second electric quantity data of the distributed photovoltaic power station based on the number of sampling points;

[0013] Wherein, the second power data is expressed as:

[0014]

[0015] in, Represented as the second power data, Indicates the power value of the electric meter at the i-th point on the k-th day, n sample is the number of power sampling points in a day.

[0016] Optionally, determining the power error value of the distributed photovoltaic power station based on the first power data and the second power data includes:

[0017] Performing difference calculation on the first point quantity data and the second electric quantity data, and obtaining a result of the difference calculation;

[0018] Determine the power error value of the distributed photovoltaic power station based on the result of the difference calculation;

[0019] Wherein, the power error value is expressed as:

[0020]

[0021] in, Indicates the power error value. Indicates the second power data, Indicates the first power data.

[0022] Optionally, filtering the first power data and the second power data based on the power error value to obtain filtered power data includes:

[0023] generating an electric quantity data set based on the first electric quantity data and the second electric quantity data;

[0024] Traversing the electric quantity data set, and determining the electric quantity data corresponding to the electric quantity error value greater than the target threshold in the electric quantity data set as abnormal data;

[0025] The abnormal data in the electric quantity data set is filtered to obtain filtered electric quantity data.

[0026] Optionally, performing interpolation processing on the filtered power data to obtain interpolated power data includes:

[0027] Determining the location of missing power values ​​in the filtered power data;

[0028] In response to determining that the position is a continuous position, interpolating the filtered power data by screening out the filtered power data to obtain first power data after interpolation;

[0029] In response to determining that the position is a discontinuous position, the filtered power data is interpolated by linear interpolation to obtain second power data after interpolation.

[0030] Optionally, performing power prediction based on the interpolated power data and obtaining a prediction result includes:

[0031] generating a target power prediction model based on the first electric quantity data after the interpolation processing and the second electric quantity data after the interpolation processing;

[0032] The photovoltaic power generation power is predicted based on the target power prediction model, and a prediction result is obtained.

[0033] Optionally, generating a power prediction model based on the first electric quantity data after the interpolation processing and the second electric quantity data after the interpolation processing includes:

[0034] Generate a power prediction model based on the first electric quantity data after the interpolation processing, the second electric quantity data after the interpolation processing, and a preset encoder and decoder;

[0035] The power prediction model is optimized based on the power error value to obtain a target power prediction model.

[0036] As a second aspect of the present disclosure, the present disclosure further provides a photovoltaic power prediction device, comprising:

[0037] The data acquisition module is configured to: acquire power data and first electric quantity data of a distributed photovoltaic power station collection meter, wherein the first electric quantity data is a difference between every two adjacent meter code values ​​in the collection meter;

[0038] an error value determination module, configured to: determine second power data of the distributed photovoltaic power station based on the power data, and determine a power error value of the distributed photovoltaic power station based on the first power data and the second power data;

[0039] A data filtering module is configured to: filter the first power data and the second power data based on the power error value to obtain filtered power data;

[0040] The interpolation processing module is configured to: perform interpolation processing on the filtered power data to obtain interpolated power data;

[0041] The power prediction module is configured to: perform power prediction based on the interpolated power data and obtain a prediction result.

[0042] Optionally, determining the second power data of the distributed photovoltaic power station based on the power data includes:

[0043] Determine the number of sampling points of the power data, and determine second electric quantity data of the distributed photovoltaic power station based on the number of sampling points;

[0044] Wherein, the second power data is expressed as:

[0045]

[0046] in, Represented as the second power data, Indicates the power value of the electric meter at the i-th point on the k-th day, n sample is the number of power sampling points in a day.

[0047] Optionally, determining the power error value of the distributed photovoltaic power station based on the first power data and the second power data includes:

[0048] Performing difference calculation on the first point quantity data and the second electric quantity data, and obtaining a result of the difference calculation;

[0049] Determine the power error value of the distributed photovoltaic power station based on the result of the difference calculation;

[0050] Wherein, the power error value is expressed as:

[0051]

[0052] in, Indicates the power error value. Indicates the second power data, Indicates the first power data.

[0053] Optionally, filtering the first power data and the second power data based on the power error value to obtain filtered power data includes:

[0054] generating an electric quantity data set based on the first electric quantity data and the second electric quantity data;

[0055] Traversing the electric quantity data set, and determining the electric quantity data corresponding to the electric quantity error value greater than the target threshold in the electric quantity data set as abnormal data;

[0056] The abnormal data in the electric quantity data set is filtered to obtain filtered electric quantity data.

[0057] Optionally, performing interpolation processing on the filtered power data to obtain interpolated power data includes:

[0058] Determining the location of missing power values ​​in the filtered power data;

[0059] In response to determining that the position is a continuous position, interpolating the filtered power data by screening out the filtered power data to obtain first power data after interpolation;

[0060] In response to determining that the position is a discontinuous position, the filtered power data is interpolated by linear interpolation to obtain second power data after interpolation.

[0061] Optionally, performing power prediction based on the interpolated power data and obtaining a prediction result includes:

[0062] generating a target power prediction model based on the first electric quantity data after the interpolation processing and the second electric quantity data after the interpolation processing;

[0063] The photovoltaic power generation power is predicted based on the target power prediction model, and a prediction result is obtained.

[0064] As a third aspect of the present disclosure, the present disclosure further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the above-mentioned photovoltaic power generation prediction method provided by the present disclosure is implemented.

[0065] As a fourth aspect of the present disclosure, the present disclosure further provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute any of the methods described above.

[0066] As described above, in the present disclosure, the first power data is first determined based on the meter code value of the photovoltaic power station collection meter, and the second power data is determined based on the power data of the photovoltaic power station. Then, the power error value of the photovoltaic power station is determined based on the first power data and the second power data, and then the first power data and the second power data are optimized based on the power error value. Finally, power prediction is performed based on the optimized data and the prediction result is obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions in the present disclosure or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0068] Figure 1A A schematic diagram of a photovoltaic power prediction method provided in an embodiment of the present disclosure.

[0069] Figure 1B A schematic diagram of a method for filtering power data provided by an embodiment of the present disclosure.

[0070] Figure 1C A schematic diagram of a method for interpolating electrical quantity data provided in an embodiment of the present disclosure.

[0071] Figure 2 A schematic diagram of the structure of a photovoltaic power prediction device provided in an embodiment of the present disclosure.

[0072] Figure 3 A schematic diagram of the structure of an electronic device for a photovoltaic power prediction method provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0073] In order to make the objectives, technical solutions and advantages of the present disclosure more clearly understood, the present disclosure is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0074] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present disclosure should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in the embodiments of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0075] In order to achieve the dual carbon goals, the installed capacity of photovoltaic power generation has increased year by year in recent years, and the proportion of distributed photovoltaic power station installed capacity has also increased year by year. The power generation of photovoltaic power stations is highly volatile and intermittent, which poses a great challenge to the safety of photovoltaic energy access to the power grid. In actual production, photovoltaic power generation prediction technology is mainly used to improve the safety of photovoltaic energy access to the power grid. Its main principle is to perform mathematical modeling based on the basic information of photovoltaic power stations and historical power generation, and then predict future photovoltaic power.

[0076] In the existing technology, photovoltaic power generation prediction technology is mainly used in centralized photovoltaic power stations with accurate power value collection. It is not applicable to distributed photovoltaic power stations with a large number of power stations, wide distribution, uneven density and poor data quality.

[0077] The existing technologies for distributed photovoltaic power prediction have some significant shortcomings in various links. First, the existing technologies have defects in model anomaly identification and reconstruction. Specifically, the model anomaly identification mainly relies on photovoltaic power data and irradiance data. For centralized photovoltaic power stations, their photovoltaic power prediction curves need to be reported to the power grid company, and there are accuracy index assessment requirements, so the power record values ​​are generally more accurate.

[0078] However, for distributed photovoltaic power stations, there are a large number of household photovoltaics, which are large in number, low in installed capacity, and completely maintained by users. Users do not pay attention to their daily power data, but only pay attention to the meter code value related to the economic benefits of power generation. Therefore, their power values ​​may be completely wrong. In this case, relying on power values ​​for outlier identification is no longer reliable. Taking the paper "Photovoltaic Power Restriction Abnormal Data Identification Algorithm Based on Mathematical Morphology Denoising" as an example, if the power value itself is incorrect, such as the CT and PT ratio settings are wrong, it will cause it to have a consistent change trend with the irradiance. At this time, this method will be completely ineffective.

[0079] In addition, distributed photovoltaic reconstruction has a large reconstruction error. According to existing literature, the relative error of photovoltaic power reconstruction is as high as 26%. If training is carried out based on such data, the prediction accuracy will be greatly reduced.

[0080] Secondly, the existing technology has defects in the construction of prediction models. Specifically, the current main distributed photovoltaic prediction model construction generally requires the use of distributed photovoltaic power station historical power data and numerical weather forecast data during training and prediction. A small number of them also add meteorological observation data and historical and predicted data of centralized photovoltaic power stations. Although these models have achieved good accuracy, they have great limitations in practical promotion. There are several main problems: First, it is difficult to obtain distributed photovoltaic data in real time. At present, there are two main ways to obtain distributed photovoltaic data in the power grid dispatching system: direct access and forwarding through the system.

[0081] Among them, in the case of direct access, the data of photovoltaic power stations with higher installed capacity can be reported by the station fusion terminal / concentrator, while the household photovoltaic with lower installed capacity is reported by installing a special collection device. The equipment investment is large and it is difficult to promote it in household photovoltaic power stations with large numbers and small capacities. In the system forwarding mode, the data is forwarded through the distribution automation system, the electricity information collection system or the aggregator platform. The dispatching system can obtain more comprehensive data, but the real-time performance is poor. Second, the operation and maintenance cost is high. The existing power forecast needs to use numerical weather forecast data (NWP) or obtain data from other data interfaces in real time. A certain data interface fee needs to be paid every year. This method is unsustainable for distributed photovoltaic forecasting. Third, the existing technology focuses on the accuracy of power prediction, and pays insufficient attention to the accuracy of daily power prediction. The large-scale access of household distributed photovoltaics makes the quality supervision of power data obviously insufficient, and model training based only on power data may have large deviations.

[0082] Finally, the existing technology has defects in model training and optimization. Specifically, in terms of model training and optimization, the existing technology uses power data error as the loss function.

[0083] In order to solve the above problems, the present disclosure provides a method, device, electronic device and storage medium for predicting photovoltaic power generation. Through the above method, the present disclosure first determines the first power data based on the meter code value of the photovoltaic power station collection meter, and determines the second power data based on the power data of the photovoltaic power station. Then, the power error value of the photovoltaic power station is determined based on the first power data and the second power data, and then the first power data and the second power data are optimized based on the power error value. Finally, power prediction is performed based on the optimized data and the prediction result is obtained.

[0084] In the present disclosure, the first power data is generated by collecting the meter code value of the electric meter, and the second power data is generated by the power data of the photovoltaic power station, so that the method in the present disclosure can more accurately collect the actual power consumption of the distributed photovoltaic power station. After that, the present disclosure optimizes the difference (power error value) between the first power data and the second power data to eliminate the bad data (i.e., error data) in the historical data of the distributed photovoltaic power station. Finally, the target power prediction model is generated by the optimized power data, thereby realizing the prediction of the power of the distributed photovoltaic power station.

[0085] After introducing the basic principles of the present disclosure, various non-limiting embodiments of the present disclosure are described in detail below.

[0086] Figure 1A A schematic diagram of a photovoltaic power prediction method provided in an embodiment of the present disclosure.

[0087] Figure 1A The photovoltaic power generation prediction method shown further comprises the following steps:

[0088] Step S10: acquiring power data and first electric quantity data of a distributed photovoltaic power station collection meter, wherein the first electric quantity data is a difference between every two adjacent meter code values ​​in the collection meter.

[0089] In some optional embodiments, the photovoltaic power generation power prediction method disclosed in the present invention can be used to predict the power generation power of a distributed photovoltaic power station. When performing power prediction of a distributed photovoltaic power station, it can be implemented through a power prediction system.

[0090] In some optional embodiments, the power prediction system may first obtain power data and first power data of the distributed photovoltaic power station collection meter, wherein the first power data may be the difference between every two adjacent meter code values ​​in the collection meter.

[0091] In some optional embodiments, the aforementioned collection meter may include meter code values ​​of several time periods, and the meter code values ​​are used to represent the power data of the photovoltaic power station. For example, the meter code values ​​may include "23.5.6-00:00-154" and "23.5.9-00:00-298", etc. In the present disclosure, the collection meter may freeze the power data of the photovoltaic power station at 00:00 every day and obtain the meter code value information of the day.

[0092] Step S20: determining second power data of the distributed photovoltaic power station based on the power data, and determining a power error value of the distributed photovoltaic power station based on the first power data and the second power data.

[0093] In some optional embodiments, step S20 further includes:

[0094] S201: Determine the number of sampling points of the power data, and determine second power data of the distributed photovoltaic power station based on the number of sampling points.

[0095] S202: performing a difference calculation on the first point quantity data and the second power quantity data, and obtaining a result of the difference calculation.

[0096] S203: Determine an electric quantity error value of the distributed photovoltaic power station based on a result of the difference calculation.

[0097] In some optional embodiments, after obtaining the power data and the first power data of the electric meter collected by the distributed photovoltaic power station, the power prediction system can determine the number of sampling points of the power data, and determine the second power data of the distributed photovoltaic power station based on the number of sampling points. Then, the difference calculation is performed on the first point quantity data and the second power data to obtain the result of the difference calculation.

[0098] In some optional embodiments, after obtaining the difference calculation result, the power prediction system can determine the power error value of the distributed photovoltaic power station based on the difference calculation result, and then optimize and correct the first power data and the second power data by the power error value.

[0099] In some optional embodiments, the aforementioned second power data may be expressed as:

[0100]

[0101] in, Represented as the second power data, Indicates the power value of the electric meter at the i-th point on the k-th day, n sample is the number of power sampling points in a day, which is 96 in the present disclosure.

[0102] In some optional embodiments, the aforementioned power error value can be expressed as:

[0103]

[0104] in, Indicates the power error value. Indicates the second power data, Indicates the first power data.

[0105] Step S30: filtering the first power data and the second power data based on the power error value to obtain filtered power data.

[0106] Figure 1BA schematic diagram of a method for filtering power data provided by an embodiment of the present disclosure.

[0107] In some optional embodiments, such as Figure 1B As shown, step S30 further includes:

[0108] S301: Generate an electric quantity data set based on the first electric quantity data and the second electric quantity data.

[0109] S302: traverse the power data set, and determine the power data corresponding to the power error value greater than the target threshold in the power data set as abnormal data.

[0110] S303: Filter the abnormal data in the power data set to obtain filtered power data.

[0111] In some optional embodiments, after obtaining the power error value, the power prediction system may first generate a power data set based on the first power data and the second power data. The power prediction system may then correct and optimize the power data set. This process avoids the process of the power prediction system correcting and optimizing the first power data and the second power data respectively, thereby improving the efficiency of correction and optimization.

[0112] In some optional embodiments, after generating the power data set, the power prediction system may first traverse the power data set, and determine the power data corresponding to the power error value greater than the target threshold in the power data set as abnormal data, and then filter the abnormal data in the power data set to obtain filtered power data.

[0113] In some optional embodiments, the aforementioned filtering process may be to reset the abnormal data in the power data set to zero, and retain the position information of the abnormal data after the zeroing process in the power data set.

[0114] Step S40: performing interpolation processing on the filtered power data to obtain interpolated power data.

[0115] Figure 1C A schematic diagram of a method for interpolating power data provided by an embodiment of the present disclosure

[0116] In some optional embodiments, such as Figure 1C As shown, step S40 further includes:

[0117] S401: Determine the location of missing power values ​​in the filtered power data.

[0118] S402: In response to determining that the position is a continuous position, interpolate the filtered power data by screening out the filtered power data, and obtain first power data after interpolation.

[0119] S403: In response to determining that the position is a discontinuous position, interpolate the filtered power data by linear interpolation to obtain second power data after interpolation.

[0120] In some optional embodiments, after filtering the abnormal data and obtaining the filtered power data, the power prediction system can determine the location of the missing power value in the filtered power data, that is, the location information of the abnormal data mentioned above.

[0121] In some optional embodiments, the power prediction system can determine the location information of the abnormal data. If the location is a continuous location, the filtered power data can be interpolated by screening to obtain the first power data after interpolation.

[0122] In some optional embodiments, if the position is a discontinuous position, the filtered power data may be interpolated by linear interpolation to obtain the second power data after interpolation.

[0123] Step S50: performing power prediction based on the interpolated power data and obtaining a prediction result.

[0124] In some optional embodiments, step S50 further includes:

[0125] S501: Generate a target power prediction model based on the first power data after the interpolation processing and the second power data after the interpolation processing.

[0126] In some optional embodiments, step S501 further includes:

[0127] S5011: Generate a power prediction model based on the first power data after the interpolation processing, the second power data after the interpolation processing, and a preset encoder and decoder.

[0128] S5012: Optimize the power prediction model based on the power error value and obtain a target power prediction model.

[0129] S502: Perform power prediction on photovoltaic power generation based on the target power prediction model and obtain a prediction result.

[0130] In some optional embodiments, after obtaining the interpolated power data, the power prediction system can generate a target power prediction model based on the interpolated first power data and the interpolated second power data. Then, the photovoltaic power generation power is predicted based on the target power prediction model to obtain a prediction result.

[0131] In some optional embodiments, the process of generating the target power prediction model by the power prediction system can be specifically as follows: the power prediction system first generates the power prediction model based on the first power data after interpolation processing, the second power data after interpolation processing, and the preset encoder and decoder. Then, the power prediction model is optimized based on the aforementioned power error value, and the target power prediction model can be obtained.

[0132] In some optional embodiments, the method disclosed in the present invention first screens the meter power acquisition data and the daily power freeze data; uses the more reliable power freeze data to screen out inaccurate power data, thereby improving the reliability of the training data set.

[0133] In some optional embodiments, the method disclosed in the present invention secondly merges and normalizes the power and electricity data of regional distributed photovoltaic power stations. Although outliers are more common for single household distributed photovoltaic power stations, for areas with more power stations, the outliers of each power station in each time period are relatively scattered, and the influence of outliers on the overall output of regional photovoltaic power at the same time is reduced. This method can obtain an accurate regional photovoltaic total output curve, and abandons the method of reconstructing data, eliminating the large errors introduced by the reconstruction link;

[0134] In some optional embodiments, the method in the present disclosure finally combines the historical data and predicted data of the regional centralized photovoltaic power station to build a power and electricity joint prediction model with an encoder-decoder structure to achieve accurate prediction of photovoltaic power and electricity on the day before. During the inference process of the model, its input data is the historical and predicted data of the centralized power station that can be directly obtained by the dispatching system, and does not rely on NWP data and power data of distributed photovoltaic power stations, thereby solving the problems of high operation and maintenance costs of NWP data and low real-time performance of distributed photovoltaic power station power data.

[0135] In some optional embodiments, the method disclosed in the present invention mainly includes three steps. The first step is to jointly construct a distributed photovoltaic power and electricity data set, and construct multiple data sets according to different screening error limits. The second step is to build and train a regional distributed photovoltaic power and electricity joint prediction model. The third step is to select the model with the smallest final loss function among multiple data sets as the final model.

[0136] In some optional embodiments, the power error value caused The main reasons for the value being too large are as follows: Data loss, which is manifested as missing data, including missing power values ​​and missing meter frozen values. Large zero drift value. The power generation of the photovoltaic power station should be 0 outside the sunrise and sunset interval. Since the zero drift of the meter collection value at some sites is large, it ultimately affects the accuracy of power generation. Fixed system errors or deviations caused by system configuration. For example, the total active power is not added up by the three-phase power. For this type of data, the total active power is processed as the three-phase addition. There are also some power stations whose error causes are unknown, and their power integrated daily power is Difference from power freeze value The difference is large, but the ratio range is relatively fixed. Random error. In most cases, the relative error of daily power generation at some sites is small, but the relative error is large on some days.

[0137] In some optional embodiments, The meter's electric energy indication is derived from the meter's electric energy frozen value and is used for power generation fee settlement, which has high reliability. However, the power value is used very rarely and its reliability cannot be guaranteed. As a reference value, data is screened and merged.

[0138] In some optional embodiments, there are many regional distributed photovoltaic sites, poor data quality, and various types of outliers are also common problems. In the research related to new energy prediction, many studies identify outliers and interpolate according to statistical laws, but for the case of a large number of outliers, this approach may introduce additional errors to the data set.

[0139] In some optional embodiments, the purpose of regional distributed photovoltaic power prediction is to provide an accurate regional photovoltaic total output curve. Although abnormal values ​​are more common for a single power station, for a region containing more power stations, the abnormal points of each power station in each time period are relatively scattered, and the influence of abnormal value points on the overall regional photovoltaic output at the same time is reduced. Therefore, the present disclosure adopts the overall idea of ​​screening and merging to obtain a regional photovoltaic combined output curve with higher numerical quality.

[0140] In some optional embodiments, for each meter's power data and daily electricity data, a meter history data set is constructed. For meter j, its power data set Freeze the power difference data set where t pi ,t Ei are the corresponding times of power and daily energy, t pi Accurate to the minute, t Ei Accurate to the day, C j is the meter capacity value, i is the serial number, n is sampleis the number of sampling points in a day, and m is the number of days included in the data set. ij or E dij Set to 0, metering capacity C ij Set to 0.

[0141] In some optional embodiments, for filtering data during maintenance, data within the maintenance period of the electric meter may be filtered out based on the maintenance records and the line topology diagram.

[0142] In some optional embodiments, with respect to power-limiting data elimination, if there is power limitation in the area where the distributed photovoltaic system is located, abnormal data is eliminated based on the power limitation information record.

[0143] In some optional embodiments, for power zero drift data correction, all negative values ​​in the original power value can be set to 0, and then the sunrise and sunset times are obtained according to the power station coordinates and date, and the power values ​​outside the sunrise and sunset interval are set to 0.

[0144] In some optional embodiments, for missing value processing, linear interpolation can be performed for a single missing value point in the power collection value, and screening can be performed for multiple consecutive missing value points. Missing values ​​in the power collection value are screened out.

[0145] In some optional embodiments, for the total active power error data recovery of the three-phase electric meter, the power data after the above 4 steps can be processed. Daily frozen value difference data According to the above process, the relative error of daily power consumption can be calculated day by day to obtain the error data set If the meter j is a three-phase energy meter, the relative error for the three-phase meter on the date is greater than the total active power recognition threshold e T1 The date of acquisition Get the phase power value of meter j on day k Then the maximum value of the sum of the phase powers on the kth day is obtained according to the following formula.

[0146]

[0147]

[0148] In some optional embodiments, the phase power maximum point P is calculated. i sum The total active power P at the corresponding time i Relative error Such as pr Greater than the relative error threshold e of the maximum value within the day T2 , then all power points on the kth day are replaced by the sum of the phase-by-phase active powers, that is,

[0149]

[0150]

[0151] In some optional embodiments, the power integrated daily electricity quantity E of the kth day is recalculated. p , and the updated daily power relative error e' r If e' r Less than the total active reconstruction verification threshold e T3 , then keep the updated power data and update the data set Otherwise, the power of day k is and daily electricity Screen out.

[0152] In some optional embodiments, for power and electricity joint screening, for meter j, if the relative error of daily electricity consumption on the kth day is e rk Greater than the threshold e T4 , then filter out the power data of the kth day The difference between the frozen power consumption data on the kth day

[0153] In some optional embodiments, for regional power and electricity merging and normalization, the regional normalized power is calculated by the following formula.

[0154] In some optional embodiments, the regional normalized daily electricity consumption can be calculated by the following formula.

[0155] Through the above method, the present disclosure first determines the first power data based on the meter code value of the photovoltaic power station collection meter, and determines the second power data based on the power data of the photovoltaic power station. Then, the power error value of the photovoltaic power station is determined based on the first power data and the second power data, and then the first power data and the second power data are optimized based on the power error value. Finally, power prediction is performed based on the optimized data and the prediction result is obtained.

[0156] Based on the same technical concept, corresponding to any of the above-mentioned embodiments and methods, the present disclosure also provides a photovoltaic power generation prediction device. The photovoltaic power generation prediction method described in any of the above-mentioned embodiments can be implemented through the photovoltaic power generation prediction device provided by the present disclosure.

[0157] Figure 2 A schematic diagram of the structure of a photovoltaic power prediction device provided in an embodiment of the present disclosure.

[0158] Figure 2 The photovoltaic power generation prediction device shown further includes the following modules:

[0159] A data acquisition module 10, an error value determination module 20, a data filtering module 30, an interpolation processing module 40 and a power prediction module 50;

[0160] The data acquisition module 10 is configured to acquire power data and first power data from a distributed photovoltaic power station collection meter, wherein the first power data is a difference between every two adjacent meter code values ​​in the collection meter.

[0161] The error value determination module 20 is configured to: determine the second power data of the distributed photovoltaic power station based on the power data, and determine the power error value of the distributed photovoltaic power station based on the first power data and the second power data. Specifically, the following steps are performed:

[0162] Determine the number of sampling points of the power data, and determine second electric quantity data of the distributed photovoltaic power station based on the number of sampling points;

[0163] Wherein, the second power data is expressed as:

[0164]

[0165] in, Represented as the second power data, Indicates the power value of the electric meter at the i-th point on the k-th day, n sample is the number of power sampling points in a day;

[0166] Performing difference calculation on the first point quantity data and the second electric quantity data, and obtaining a result of the difference calculation;

[0167] Determine the power error value of the distributed photovoltaic power station based on the result of the difference calculation;

[0168] Wherein, the power error value is expressed as:

[0169]

[0170] in, Indicates the power error value. Indicates the second power data, Indicates the first power data.

[0171] The data filtering module 30 is configured to filter the first power data and the second power data based on the power error value to obtain filtered power data. Specifically, the following steps are performed:

[0172] generating an electric quantity data set based on the first electric quantity data and the second electric quantity data;

[0173] Traversing the electric quantity data set, and determining the electric quantity data corresponding to the electric quantity error value greater than the target threshold in the electric quantity data set as abnormal data;

[0174] The abnormal data in the electric quantity data set is filtered to obtain filtered electric quantity data.

[0175] The interpolation processing module 40 is configured to perform interpolation processing on the filtered power data to obtain interpolated power data. Specifically, the following steps are performed:

[0176] Determining the location of missing power values ​​in the filtered power data;

[0177] In response to determining that the position is a continuous position, interpolating the filtered power data by screening out the filtered power data to obtain first power data after interpolation;

[0178] In response to determining that the position is a discontinuous position, the filtered power data is interpolated by linear interpolation to obtain second power data after interpolation.

[0179] The power prediction module 50 is configured to perform power prediction based on the interpolated power data and obtain a prediction result. Specifically, the following steps are performed:

[0180] Generating a target power prediction model based on the first power data after the interpolation processing and the second power data after the interpolation processing includes:

[0181] Generate a power prediction model based on the first electric quantity data after the interpolation processing, the second electric quantity data after the interpolation processing, and a preset encoder and decoder;

[0182] Optimizing the power prediction model based on the power error value to obtain a target power prediction model;

[0183] The photovoltaic power generation power is predicted based on the target power prediction model, and a prediction result is obtained.

[0184] Based on the same technical concept, corresponding to any of the above-mentioned embodiments and methods, the present disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the photovoltaic power generation power prediction method described in any of the above-mentioned embodiments is implemented.

[0185] Figure 3A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0186] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0187] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0188] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0189] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0190] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0191] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0192] The electronic device of the above embodiment is used to implement the corresponding photovoltaic power prediction method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0193] Based on the same technical concept, corresponding to any of the above-mentioned embodiment methods, the present disclosure also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the photovoltaic power generation power prediction method described in any of the above embodiments.

[0194] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0195] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the photovoltaic power prediction method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0196] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples. Based on the concept of the present disclosure, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present disclosure as described above, which are not provided in detail for the sake of simplicity.

[0197] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present disclosure difficult to understand, the known power / ground connections to the integrated circuit (IC) chips and other components may or may not be shown in the provided figures. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present disclosure difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present disclosure will be implemented (that is, these details should be fully within the scope of understanding of those skilled in the art). Where specific details (e.g., circuits) are set forth to describe exemplary embodiments of the present disclosure, it is apparent to those skilled in the art that the embodiments of the present disclosure can be implemented without these specific details or with changes in these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0198] Although the present disclosure has been described in conjunction with specific embodiments of the present disclosure, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0199] The embodiments of the present disclosure are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present disclosure should be included in the scope of protection of the present disclosure.

Claims

1. A method for predicting photovoltaic power generation, characterized in that: include: Acquire power data and first electric quantity data of a collection meter of a distributed photovoltaic power station, wherein the first electric quantity data is a difference between meter code values ​​of each two adjacent time periods in the collection meter; The table code value is used to represent the power data of the distributed photovoltaic power station; Determine second power data of the distributed photovoltaic power station based on the power data, and determine a power error value of the distributed photovoltaic power station based on the first power data and the second power data; Filtering the first power data and the second power data based on the power error value to obtain filtered power data; Performing interpolation processing on the filtered power data to obtain interpolated power data; Perform power prediction based on the interpolated power data and obtain a prediction result; Wherein, determining the second power data of the distributed photovoltaic power station based on the power data includes: Determine the number of sampling points of the power data, and determine second electric quantity data of the distributed photovoltaic power station based on the number of sampling points; Wherein, the second power data is expressed as: in, Represented as the second power data, Indicates the power value of the electric meter at the i-th point on the k-th day, n sample is the number of power sampling points in a day.

2. The method according to claim 1, characterized in that The determining the power error value of the distributed photovoltaic power station based on the first power data and the second power data includes: Performing a difference calculation on the first power data and the second power data, and obtaining a result of the difference calculation; Determine the power error value of the distributed photovoltaic power station based on the result of the difference calculation; Wherein, the power error value is expressed as: in, Indicates the power error value. Indicates the second power data, Indicates the first power data.

3. The method according to claim 1, characterized in that The filtering the first power data and the second power data based on the power error value to obtain filtered power data includes: generating an electric quantity data set based on the first electric quantity data and the second electric quantity data; Traversing the electric quantity data set, and determining the electric quantity data corresponding to the electric quantity error value greater than the target threshold in the electric quantity data set as abnormal data; The abnormal data in the electric quantity data set is filtered to obtain filtered electric quantity data.

4. The method according to claim 1, characterized in that: The interpolation processing is performed on the filtered power data to obtain the interpolated power data, including: Determining the location of missing power values ​​in the filtered power data; In response to determining that the position is a continuous position, interpolating the filtered power data by screening out the filtered power data to obtain first power data after interpolation; In response to determining that the position is a discontinuous position, the filtered power data is interpolated by linear interpolation to obtain second power data after interpolation.

5. The method according to claim 4, characterized in that The performing power prediction based on the interpolated power data and obtaining the prediction result comprises: generating a target power prediction model based on the first electric quantity data after the interpolation processing and the second electric quantity data after the interpolation processing; The photovoltaic power generation power is predicted based on the target power prediction model, and a prediction result is obtained.

6. The method according to claim 5, characterized in that The generating a power prediction model based on the first electric quantity data after the interpolation processing and the second electric quantity data after the interpolation processing includes: Generate a power prediction model based on the first electric quantity data after the interpolation processing, the second electric quantity data after the interpolation processing, and a preset encoder and decoder; The power prediction model is optimized based on the power error value to obtain a target power prediction model.

7. A photovoltaic power generation prediction device, characterized in that: include: The data acquisition module is configured to: acquire power data and first power data of a distributed photovoltaic power station collection meter, wherein the first power data is the difference between table code values ​​of every two adjacent time periods in the collection meter; the table code value is used to represent the power data of the distributed power station; an error value determination module, configured to: determine second power data of the distributed photovoltaic power station based on the power data, and determine a power error value of the distributed photovoltaic power station based on the first power data and the second power data; A data filtering module is configured to: filter the first power data and the second power data based on the power error value to obtain filtered power data; The interpolation processing module is configured to: perform interpolation processing on the filtered power data to obtain interpolated power data; A power prediction module is configured to: perform power prediction based on the interpolated power data and obtain a prediction result; Wherein, the error value determination module is further configured as follows: Determine the number of sampling points of the power data, and determine second electric quantity data of the distributed photovoltaic power station based on the number of sampling points; Wherein, the second power data is expressed as: in, Represented as the second power data, Indicates the power value of the electric meter at the i-th point on the k-th day, n sample is the number of power sampling points in a day.

8. The device according to claim 7, characterized in that The determining the power error value of the distributed photovoltaic power station based on the first power data and the second power data includes: Performing a difference calculation on the first power data and the second power data, and obtaining a result of the difference calculation; Determine the power error value of the distributed photovoltaic power station based on the result of the difference calculation; Wherein, the power error value is expressed as: in, Indicates the power error value. Indicates the second power data, Indicates the first power data.

9. The device according to claim 7, characterized in that The filtering the first power data and the second power data based on the power error value to obtain filtered power data includes: generating an electric quantity data set based on the first electric quantity data and the second electric quantity data; Traversing the electric quantity data set, and determining the electric quantity data corresponding to the electric quantity error value greater than the target threshold in the electric quantity data set as abnormal data; The abnormal data in the electric quantity data set is filtered to obtain filtered electric quantity data.

10. The device according to claim 7, characterized in that The interpolation processing is performed on the filtered power data to obtain the interpolated power data, including: Determining the location of missing power values ​​in the filtered power data; In response to determining that the position is a continuous position, interpolating the filtered power data by screening out the filtered power data to obtain first power data after interpolation; In response to determining that the position is a discontinuous position, the filtered power data is interpolated by linear interpolation to obtain second power data after interpolation.

11. The device according to claim 10, characterized in that The performing power prediction based on the interpolated power data and obtaining the prediction result comprises: generating a target power prediction model based on the first electric quantity data after the interpolation processing and the second electric quantity data after the interpolation processing; The photovoltaic power generation power is predicted based on the target power prediction model, and a prediction result is obtained.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 6 is implemented.

13. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 6.

Images