CN-122021599-A - Photovoltaic string convenient configuration method for new energy station system

CN122021599ACN 122021599 ACN122021599 ACN 122021599ACN-122021599-A

Abstract

The invention relates to a photovoltaic string convenient configuration method for a new energy station system, in particular to the technical field of intelligent operation and maintenance of a photovoltaic power station, which realizes the technical span from traditional manual experience configuration to data-driven automatic optimization by constructing an intelligent photovoltaic string configuration closed-loop system, integrates parameter rule automatic extraction, historical data strategy mining, multi-objective optimization generation and online closed-loop verification, the system can adapt to environmental changes and equipment aging, continuously optimize configuration templates, ensure that the photovoltaic strings are in an optimal running state for a long time, and effectively improve the power generation efficiency, the equipment reliability and the full life cycle economic benefit of the new energy station.

Inventors

SHEN BO
ZHAO TIANYE
CHEN DEBIN
CHEN YANLEI
SU RUIZHI
LI ZHOUYU
SANG NAN
LIU XIN

Assignees

北京华能新锐控制技术有限公司
西安热工研究院有限公司
华能吉林发电有限公司新能源分公司

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (10)

1. The photovoltaic group string convenient configuration method for the new energy station system is characterized by comprising the following steps of: Step S1, a server receives a technical document of photovoltaic string equipment, analyzes a parameter description text in the technical document by adopting a natural language processing technology, extracts equipment parameter rules, and generates a structured parameter rule base containing parameter names, parameter value ranges and parameter association relations; S2, the server adopts a deep learning model to mine a parameter optimization strategy based on the structured parameter rule base generated in the step S1 and combines the historical configuration data and the corresponding operation performance data to generate an optimization strategy base containing parameter configuration priority and parameter optimization direction; Step S3, the server generates an adaptive configuration template adapting to different environmental conditions by adopting a multi-objective optimization algorithm based on the optimization strategy library generated in the step S2, carries out simulation verification on the adaptive configuration template, and outputs a final configuration template passing the verification; and S4, monitoring actual operation data of the deployed configuration templates by the server, and updating the deep learning model in the step S2 and the multi-objective optimization algorithm in the step S3 by adopting an online learning algorithm when the deviation between the actual operation data and the simulation data exceeds a preset threshold value, so as to realize closed-loop optimization of the configuration templates.
2. The method for conveniently configuring the photovoltaic string for the new energy station system according to claim 1, wherein in step S1, a natural language processing technology is adopted to analyze a parameter description text in a technical document, and the specific operation of extracting the device parameter rule is as follows: Firstly, carrying out optical character recognition processing on a technical document, and converting document content into machine-readable text data; then adopting a named entity recognition technology to locate a parameter name entity and a parameter value entity in the text; And finally, identifying the corresponding relation between the parameter name entity and the parameter value entity by adopting a relation extraction technology, and extracting a complete equipment parameter rule based on the identified corresponding relation.
3. The photovoltaic string convenient configuration method for the new energy station system according to claim 2, wherein in step S1, the specific operation of generating the structured parameter rule base comprising the parameter name, the parameter value range and the parameter association relation is as follows: the extracted device parameter rules are organized and stored according to a preset rule mode, the rule mode distributes specific data types for parameter names, a standard expression format is defined for parameter value ranges, a logic expression form is specified for parameter association relations, and a machine-readable structured parameter rule base is constructed by systematically mapping and storing all parameter rule examples.
4. The photovoltaic string convenient configuration method for the new energy station system according to claim 3, wherein in step S2, the specific operation of adopting the deep learning model to mine the parameter optimization strategy is as follows: The server performs multi-source data fusion processing on the structured parameter rule base generated in the step S1 and the historical configuration data, wherein the multi-source data fusion processing comprises characteristic alignment processing and time sequence alignment processing, and a standardized training sample set is formed; Training the standardized training sample set by adopting a deep reinforcement learning algorithm to generate a parameter configuration priority grade and a parameter optimization direction vector; And constructing an optimization strategy library based on the parameter configuration priority scores and the parameter optimization direction vectors.
5. The photovoltaic string convenient configuration method for the new energy station system according to claim 4, wherein in step S2, the specific operation of constructing the optimization strategy library is as follows: Normalizing the parameter configuration priority scores to obtain normalized priority weights; performing cluster analysis processing on the parameter optimization direction vector to generate a typical optimization mode set; and carrying out association mapping on the standardized priority weights and the typical optimization mode set to form an optimization strategy library comprising weight distribution and mode identification functions.
6. The photovoltaic string convenient configuration method for the new energy station system according to claim 5, wherein in step S3, the specific operation of generating the adaptive configuration template by adopting the multi-objective optimization algorithm is as follows: the server establishes a multi-objective optimization model based on the optimization strategy library generated in the step S2, taking the maximized power generation efficiency, the maximized equipment life and the maximized voltage level as three optimization targets and taking the equipment safe operation boundary and the site physical constraint as limiting conditions; And obtaining the self-adaptive configuration template capable of keeping balanced performance under various environmental conditions by solving the multi-objective optimization model.
7. The photovoltaic string convenient configuration method for the new energy station system according to claim 6, wherein in step S3, the specific operation of performing simulation verification on the adaptive configuration template is as follows: Simulating conditions of irradiance, ambient temperature and component temperature which change all the year round by adopting a photovoltaic system simulator, and verifying the power generation efficiency, equipment heat loss and voltage fluctuation rate of the self-adaptive configuration template under the simulation conditions; And when all the verification indexes meet the preset standards, outputting the self-adaptive configuration template as a final configuration template.
8. The photovoltaic string convenient configuration method for the new energy station system according to claim 7, wherein in step S4, the specific operation of monitoring the actual operation data of the deployed configuration template is: The server continuously collects power generation efficiency data, equipment temperature data and voltage fluctuation data of the photovoltaic string in an operation state as actual operation data; real-time comparison is carried out on actual operation data and simulation data corresponding to a simulation verification stage, and deviation values of all data items are calculated through a deviation calculation model; and triggering a model updating process when the deviation value of any data item exceeds a corresponding preset threshold value.
9. The photovoltaic string convenient configuration method for the new energy station system according to claim 8, wherein in step S4, the specific operations of updating the deep learning model and the multi-objective optimization algorithm by adopting the online learning algorithm are as follows: Adopting an incremental learning algorithm, taking actual running data as a new training sample, and performing parameter fine adjustment on the deep learning model in the step S2; Meanwhile, a self-adaptive optimization algorithm is adopted, and the weight distribution strategy of the multi-objective optimization algorithm in the step S3 is adjusted according to the actual operation effect; and using the updated deep learning model and the multi-objective optimization algorithm to generate a new round of self-adaptive configuration template to realize closed-loop optimization.
10. A photovoltaic string convenient configuration system for a new energy station system of the method of claim 1, comprising a structured parameter module, a parameter optimization module, a simulation verification module and a configuration module: the server receives the technical document of the photovoltaic string device, analyzes the parameter description text in the technical document by adopting a natural language processing technology, extracts the device parameter rules, and generates a structured parameter rule base containing parameter names, parameter value ranges and parameter association relations; the parameter optimization module is used for the server to mine a parameter optimization strategy by adopting a deep learning model based on the generated structured parameter rule base and combining the historical configuration data with the corresponding operation performance data to generate an optimization strategy base containing parameter configuration priority and parameter optimization direction; The simulation verification module is used for generating self-adaptive configuration templates adapting to different environmental conditions by adopting a multi-objective optimization algorithm based on the generated optimization strategy library, performing simulation verification on the self-adaptive configuration templates, and outputting a final configuration template passing the verification; The configuration module is used for monitoring actual operation data of the deployed configuration template by the server, and when the deviation between the actual operation data and the simulation data exceeds a preset threshold value, updating the deep learning model and the multi-objective optimization algorithm by adopting an online learning algorithm to realize closed-loop optimization of the configuration template.

Description

Photovoltaic string convenient configuration method for new energy station system Technical Field The invention relates to the field of intelligent operation and maintenance of photovoltaic power stations, in particular to a convenient configuration method for a photovoltaic string of a new energy station system. Background With the rapid iteration and large-scale application of the photovoltaic power generation technology, the serial string type inverter adopted by the modern photovoltaic field station and the types of the matched photovoltaic modules thereof are diversified day by day. Every year, a plurality of photovoltaic string devices with new models and new specifications are put into the market, and the devices have differences in electrical parameters, operation characteristics, communication protocols and the like. In the daily operation and maintenance, capacity expansion transformation or equipment replacement process of a field station, operation and maintenance personnel face the challenge of quickly and accurately incorporating new model string equipment into an existing monitoring system. At present, complete equipment parameter information is usually only recorded in unstructured documents such as technical manuals, data sheets and the like provided by manufacturers, and operation and maintenance personnel need to manually review, understand and extract key parameters. At present, a template library for assisting in configuration of string equipment information in a photovoltaic station monitoring system is established and updated by severely relying on manual operation. The operation and maintenance engineer needs to manually enter information of various technical parameters, operation threshold values, alarm conditions and the like into the system one by one according to the paper or electronic version technical document of the new equipment so as to create a new configuration template. The template updating mode based on manual maintenance has obvious hysteresis, cannot keep up with the updating speed of equipment, and a large number of newly commissioned equipment cannot obtain accurate model support in a monitoring system. The fundamental problem is that the prior art lacks the ability to automatically extract, parse and format device parameters from unstructured data sources, failing to construct a self-evolving intelligent knowledge system. The technical defect directly causes that the configuration information updating period of the photovoltaic station monitoring system is far longer than the actual operation period of the equipment, so that the integrity of the monitoring data of the new equipment is affected, the hidden danger of equipment operation is covered due to untimely or inaccurate parameter configuration, and the potential risk is finally formed for the overall power generation efficiency and the safe operation of the station. Disclosure of Invention Aiming at the technical problems in the prior art, the invention provides a convenient configuration method for a photovoltaic string of a new energy station system, which aims to solve the problems in the background art. The technical scheme for solving the technical problems is as follows, and the photovoltaic string convenient configuration method for the new energy station system comprises the following steps: Step S1, a server receives a technical document of photovoltaic string equipment, analyzes a parameter description text in the technical document by adopting a natural language processing technology, extracts equipment parameter rules, and generates a structured parameter rule base containing parameter names, parameter value ranges and parameter association relations; S2, the server adopts a deep learning model to mine a parameter optimization strategy based on the structured parameter rule base generated in the step S1 and combines the historical configuration data and the corresponding operation performance data to generate an optimization strategy base containing parameter configuration priority and parameter optimization direction; Step S3, the server generates an adaptive configuration template adapting to different environmental conditions by adopting a multi-objective optimization algorithm based on the optimization strategy library generated in the step S2, carries out simulation verification on the adaptive configuration template, and outputs a final configuration template passing the verification; and S4, monitoring actual operation data of the deployed configuration templates by the server, and updating the deep learning model in the step S2 and the multi-objective optimization algorithm in the step S3 by adopting an online learning algorithm when the deviation between the actual operation data and the simulation data exceeds a preset threshold value, so as to realize closed-loop optimization of the configuration templates. In a preferred embodiment, in step S1, the specific operation of parsing the par