CN-120013265-B - Photovoltaic string low-performance operation insight analysis system and method based on big data

CN120013265BCN 120013265 BCN120013265 BCN 120013265BCN-120013265-B

Abstract

The invention provides a photovoltaic string performance attenuation performance operation insight analysis system and method based on big data, wherein the system collects the electrical number and production state data of a string inverter or a direct current combiner box and a photovoltaic string; the method comprises the steps of determining a power generation analysis unit according to the electrical relation between a serial number analysis photovoltaic string and upper equipment, preprocessing after aligning data time, checking effective model input data for preparing a performance analysis operation model, calculating power generation performance evaluation indexes corresponding to each power generation analysis unit by using the performance analysis operation model, optimizing a power generation service string structure based on the calculated power generation performance evaluation indexes, and deciding operation and maintenance processing schemes with different evaluation index level matching. The scheme can overcome the defect of insufficient reliability of the analysis result in the prior art, adopts an special data processing and analysis algorithm to analyze abnormal low-efficiency states in a plurality of groups of strings, and accordingly provides improvement measures, thereby being beneficial to realizing effective improvement of the overall performance of the photovoltaic system.

Inventors

LIU GUOTONG
LI YE
ZHANG LINGYUN
ZHANG SITAO
JI XIANGYU

Assignees

中国铁道科学研究院集团有限公司
铁科院(北京)工程咨询有限公司
中国铁道科学研究院集团有限公司城市轨道交通中心

Dates

Publication Date: 20260505
Application Date: 20241220

Claims (10)

1. A photovoltaic string performance degradation performance insight analysis system based on big data, the system comprising: the data collection module is configured to acquire the electrical number and production state data of the full-field string type inverter or the direct current combiner box of the photovoltaic field station to be analyzed and the photovoltaic string; The unit structure determining module is configured to analyze the electrical relation between the photovoltaic string and the superior string inverter or the direct current combiner box according to the electrical number, and determine a power generation analysis unit based on a power generation unit digital model corresponding to the photovoltaic string and superior equipment; the data processing module is configured to pre-process the data after aligning the time of the acquired data and check the data based on the pre-processed data to prepare effective model input data of the performance analysis operation model; The serial performance analysis module is configured to start the constructed performance analysis operation model according to the set analysis operation starting time, input the effective model input data of the station to be analyzed and determine the power generation performance evaluation index corresponding to each power generation analysis unit; the business optimization application module is configured to optimally mark a power generation business string structure based on the calculated power generation performance evaluation index and to decide an operation and maintenance processing scheme with different evaluation index grades matched; The string performance analysis module calculates the average value and standard deviation of the capacity utilization rates of the upper-level different power generation analysis units of the multi-photovoltaic string based on the performance analysis operation model to serve as a power generation performance evaluation index; The business optimization application module comprises a string structure optimization unit, wherein the string structure optimization unit is configured to reject low-performance photovoltaic strings by combining a power generation performance evaluation index and a set optimization standard comparison analysis result, so as to realize optimization and marking of a power generation operation string structure; If the calculated standard deviation conclusion does not meet the optimization standard, marking the obtained standard deviation conclusion on the group string with the lowest daily average capacity utilization rate, then removing the group string, calculating again based on the remaining group string in the power generation analysis unit, comparing the circulation with the set optimization standard, and marking the obtained standard deviation conclusion on the group string with the lowest daily average capacity utilization rate in the calculation until the last standard deviation conclusion meets the optimization standard, and stopping the circulation.
2. The system of claim 1, wherein the data processing module comprises a data time alignment and partitioning unit configured to time align the collected device time series production status data using the same proximity time data information model, and select a power generation status data object for storage according to a preset valid data period and analysis node.
3. The system of claim 2, wherein the data processing module comprises a data preprocessing unit for performing a data cleansing process and a data completion process on the data object, the data completion process comprising supplementing the data missing values using interpolation.
4. A system according to claim 3, wherein the data processing module comprises a validity checking unit that checks based on the preprocessed data to prepare valid model input data for the performance analysis operational model according to the following logic: judging whether the data quantity of each photovoltaic group string corresponding to different analysis nodes meets the set data quantity condition, selecting the photovoltaic group string data which does not meet the data quantity condition, and taking the data of the rest photovoltaic group strings as standby power generation state data; and judging whether the number of the photovoltaic strings under the standby power generation state data corresponding to the power generation analysis units meets the set string number condition, selecting the power generation analysis units meeting the string number condition, and taking the standby power generation state data as effective model input data.
5. The system according to claim 1 or 4, wherein in the process of calculating the power generation performance evaluation indexes of the different power generation analysis units by the serial performance analysis module, the starting time of the analysis model is set first, and the time when the irradiance of the photovoltaic station is greater than the set condition for the last time a day is set as the starting time point of the calculation of the daily performance analysis operation model.
6. The system of claim 1, wherein the performance analysis operational model employs the following logic: , , , Wherein, the The standard deviation of the daily average capacity utilization rate of the photovoltaic strings corresponding to the power generation analysis unit is calculated at present, m is the number of strings under the power generation analysis unit, Represents the average value of the daily capacity utilization of the ith group string in the power generation analysis unit, As a result of averaging again the daily capacity utilization average value of all the strings under the power generation analysis unit, Representing the average value of the daily capacity utilization of each string, n representing the number of string capacity utilizations calculated by a single analysis node of each day of the string, Capacity utilization data representing a daily jth analysis node, And (3) representing the capacity utilization rate of a single analysis node of the photovoltaic string, wherein VI represents the direct-current side power of the photovoltaic string of the single analysis node, and Wp represents the installed capacity of the photovoltaic string.
7. The system of claim 1, wherein the business optimization application module comprises an operation and maintenance scheme decision unit configured to determine performance evaluation levels of different levels based on a set performance index dividing rule, and the performance evaluation levels are corresponding to performance attenuation states of different degrees, and the performance evaluation levels are pre-warned and displayed in different forms to match operation and maintenance processing schemes of different emergency degrees.
8. The system of claim 7, further comprising an operation and maintenance effect analysis module configured to mark the photovoltaic string adopting the operation and maintenance scheme and record operation and maintenance information, update the data label of the operation and maintenance string and compare the power generation performance evaluation index of the power generation analysis unit before and after operation and maintenance, analyze the optimization effect of the operation and maintenance scheme, and provide data support for the operation and maintenance scheme decision unit.
9. A photovoltaic string performance degradation performance operation insight analysis method based on big data, characterized in that the method is applied to the system of any one of claims 1-8, the method comprising: the method comprises the steps that a data collection module is used for obtaining the electrical number of a string inverter or a direct current combiner box of the whole field of a photovoltaic field station to be analyzed and photovoltaic string production state data corresponding to the whole field measuring point; analyzing the electrical relation between the photovoltaic string and the upper-level string inverter or the direct current combiner box according to the electrical number, and determining a power generation analysis unit based on a power generation unit digital model corresponding to the photovoltaic string and the upper-level equipment; After aligning the time of the acquired different electrical serial numbers, preprocessing the data and checking based on the preprocessed data to prepare effective model input data of the performance analysis operation model; Starting a constructed performance analysis operation model according to the set analysis operation starting time, inputting effective model input data of a station to be analyzed, and determining power generation performance evaluation indexes corresponding to each power generation analysis unit; the low-performance photovoltaic string is removed by combining the power generation performance evaluation index and a set optimization standard comparison analysis result, so that the optimization and marking of the power generation operation string structure are realized; If the calculated standard deviation conclusion does not meet the optimization standard, marking the obtained standard deviation conclusion on the group string with the lowest daily average capacity utilization rate, then removing the group string, calculating again based on the remaining group string in the power generation analysis unit, comparing the circulation with the set optimization standard, and marking the obtained standard deviation conclusion on the group string with the lowest daily average capacity utilization rate in the calculation until the last standard deviation conclusion meets the optimization standard, and stopping the circulation.
10. A storage medium having stored thereon program code for implementing the method of claim 9.

Description

Photovoltaic string low-performance operation insight analysis system and method based on big data Technical Field The invention relates to the technical field of photovoltaic string performance analysis, in particular to a photovoltaic string low-performance operation insight analysis system and method based on big data. Background Under illumination conditions, the panel of the photovoltaic cell absorbs light energy, and different numbers of charges are accumulated at two ends of the panel, so that electromotive force is generated at two ends of the panel, and the conversion from light energy to electric energy is realized. As an important device for energy conversion, the photovoltaic module may affect the power generation benefits due to factors such as surface pollution and module performance attenuation. In theory, the strings under the same power generation unit are under the same working condition, the power generation power (or current/voltage value) of the strings should always be consistent or have only slight difference, but some environmental and human factors can cause the strings to fail to fully generate, which obviously reduces the power generation efficiency of the photovoltaic power station, and effectively identifies the low-efficiency strings by a professional string low-efficiency algorithm has an important effect on improving the power generation efficiency. The main recognition thinking is that starting from the actual historical period string data of the photovoltaic string, the string data with set characteristics is recognized by combining a clustering algorithm, so that the corresponding low-efficiency string is estimated, and the abnormal photovoltaic string is recognized for processing based on the low-efficiency string, so that the economic and efficient operation and maintenance of the power station are promoted to a certain extent. For example, a method for identifying photovoltaic strings in low efficiency is provided in some prior art, the method comprises the steps of obtaining current and voltage string data of each photovoltaic string in a history period for a preset time period, identifying abnormal data after processing and removing the abnormal data, screening string data contained in a first preset period from the removed string data, obtaining the number of the screened photovoltaic strings, and selecting a clustering object to identify the photovoltaic strings in low efficiency in photovoltaic equipment. Although the influence of abnormal values on the identification result is overcome to a certain extent, the current and voltage data of the photovoltaic string are limited, and a comprehensive data base is difficult to provide for a clustering algorithm, so that the defect in the accuracy of the identification result is caused. In addition, researches on low-efficiency identification and electric quantity lifting of a photovoltaic module are related, for example, a certain prior art provides a low-efficiency identification and electric quantity lifting method of the photovoltaic module under multi-orientation and inclination angles, the scheme stores preprocessed data into a cloud space, and an MLP-Mixer model is built by adopting a TensorFlow deep learning framework; integrating the output result of the MLP-Mixer model with the priori knowledge of the photovoltaic power station to form a second data set, further establishing a XGBoost model, collecting real-time operation data of the photovoltaic power station when the model is applied, and transmitting the real-time operation data into the MLP-Mixer model. The inefficient group strings are identified by the MLP-Mixer model and maintenance recommendations are generated by the XGBoost model. The abnormal maintenance proposal analysis model is established by combining the identification result and the priori knowledge, but the analysis result depends on the identification result, the accuracy of the low-efficiency group string information which is only identified based on the history period group string data is difficult to guarantee, and the reliability of the maintenance proposal analysis result is insufficient. In addition, the individual photovoltaic equipment adopts alarm information to realize the low-efficiency identification of the photovoltaic module or manually identify the low-efficiency photovoltaic module based on the power generation curve of the equipment, however, the method for utilizing the alarm information lacks accuracy, the general equipment remote signaling alarm usually does not report the state of the low-efficiency operation of the group string, a great deal of labor cost and on-site operation time cost are consumed when the power generation curve of the equipment is checked, and the traditional methods can cause missed judgment and misjudgment and can not well meet the operation requirements of the photoelectric field. The information disclosed in the background section of the invention i