Photovoltaic power automatic adjusting method and system

Abstract

The invention relates to the technical field of photovoltaic power adjustment, and discloses a photovoltaic power automatic adjustment method and system. The method comprises the steps of collecting real-time parameters of an inverter through distributed communication, performing multi-dimensional evaluation sorting based on equipment efficiency, load rate and operation time, executing adaptive power distribution fusing edge calculation and cloud optimization, generating standardized instruction time-sharing issuing, and storing control logic by using a block chain. And an encrypted link is established to realize bidirectional data interaction between the power station and the dispatching system to form a closed-loop control system. According to the invention, high-efficiency, precise and reliable dynamic power adjustment of the inverter level is realized.

Description

Technical Field

[0001] The present application relates to the technical field of photovoltaic power regulation, and in particular to a photovoltaic power automatic regulation method and system. Background Art

[0002] Currently, photovoltaic power plants, as a crucial component of renewable energy generation, are experiencing a clear trend toward scale and intensification. In the daily operation of a PV power plant, individual inverters, the core energy converter between the PV panels and the grid, fulfill the crucial function of converting direct current (DC) to alternating current (AC) and stabilizing its output. Inverter power regulation in existing PV power plants primarily relies on a centralized monitoring system that collects inverter operating data through polling. This system then implements simple, pre-set control logic, such as issuing unified commands based on the rated power ratio of each inverter group or uniformly reducing overall output when grid dispatch demands arise. To ensure real-time communication and data integrity, standard communication protocols such as MODBUS, IEC60870-5-104, or custom protocols are typically used for data exchange. Furthermore, some power plants deploy simple dispatch response modules to implement power reduction or increase operations. These traditional solutions, characterized by periodic data collection, static configuration, and unified command issuance, generally meet the regulation requirements of smaller PV power plants.

[0003] However, with the continued growth of installed capacity in photovoltaic power plants and the resulting proliferation and diversification of inverter models, existing technologies have gradually exposed a series of problems. First, traditional data collection and control systems lack a flexible multi-level communication architecture, resulting in limited data collection frequency, severe communication conflicts, and the risk of information loss or delays. This makes them unable to support high-frequency and precise inverter status monitoring. Second, power allocation strategies are simplistic and fail to intelligently evaluate and prioritize inverters based on their real-time operating status, leading to low overall operating efficiency and exacerbated equipment aging caused by chronic overload in individual inverters. Third, the lack of an effective mechanism for personalized processing and differentiated distribution of power regulation commands leads to delayed regulation responses and the inability to accurately and quickly implement grid dispatch commands. More importantly, traditional control logic lacks trusted evidence and tamper-proof mechanisms, posing security risks such as system tampering and untraceable command execution. This makes it difficult to meet the increasing demands for safe, compliant, and intelligent operation and maintenance of new energy power plants. Summary of the Invention

[0004] The present application provides a photovoltaic power automatic adjustment method and system for achieving high-efficiency, precise and reliable dynamic power adjustment at the inverter level.

[0005] In the first aspect, the present application provides a photovoltaic power automatic regulation method, which includes: establishing a connection with all inverters in a photovoltaic power station through a distributed communication network, collecting real-time operating parameters of each inverter, and obtaining an inverter information data set; performing a multi-dimensional evaluation and sorting of the input power of each inverter according to the inverter information data set to obtain an inverter power priority sequence, wherein the multi-dimensional evaluation includes equipment efficiency index, load rate and operating time analysis; executing an adaptive iterative algorithm based on the inverter power priority sequence to perform power allocation calculation to obtain a target power value for each inverter, wherein the adaptive iterative algorithm integrates edge computing and cloud collaborative processing; encoding the target power value into a standardized control instruction, and issuing the control instruction to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data; constructing a control logic security storage system through blockchain technology, converting the control logic into a smart contract form for storage, and obtaining a tamper-proof control strategy library; establishing an encrypted communication link based on the power regulation response data to realize two-way data exchange between the power station and the power grid dispatching system, obtaining a closed-loop control system, and transmitting the target power value to the on-site execution terminal.

[0006] In a second aspect, the present application provides a photovoltaic power automatic regulation system, the photovoltaic power automatic regulation system comprising:

[0007] A connection module is used to establish a connection with all inverters in the photovoltaic power station through a distributed communication network, collect real-time operating parameters of each inverter, and obtain an inverter information data set;

[0008] a sorting module, configured to perform a multi-dimensional evaluation and sorting of the input power of each inverter according to the inverter information data set to obtain an inverter power priority sequence, wherein the multi-dimensional evaluation includes equipment efficiency index, load rate and operating time analysis;

[0009] An allocation module, configured to execute an adaptive iterative algorithm based on the inverter power priority sequence to perform power allocation calculations to obtain a target power value for each inverter, wherein the adaptive iterative algorithm integrates edge computing and cloud collaborative processing;

[0010] a sending module, configured to encode the target power value into a standardized control instruction, and send the control instruction to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data;

[0011] A conversion module is used to build a control logic security storage system using blockchain technology, convert the control logic into a smart contract format for storage, and obtain a tamper-proof control policy library;

[0012] The encryption module is used to establish an encrypted communication link based on the power regulation response data, realize two-way data exchange between the power station and the power grid dispatching system, obtain a closed-loop control system, and transmit the target power value to the on-site execution terminal.

[0013] In a third aspect, a photovoltaic power automatic adjustment device is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the photovoltaic power automatic adjustment device executes the above-mentioned photovoltaic power automatic adjustment method.

[0014] In a fourth aspect, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, and when the computer-readable storage medium is run on the computer, the computer executes the above-mentioned photovoltaic power automatic adjustment method.

[0015] In the technical solution provided by this application, a connection is established with all inverters in the photovoltaic power station through a distributed communication network, ensuring the system's wide adaptability to inverters of different manufacturers and models, structurally avoiding the bottlenecks and fault concentration problems existing in the centralized communication architecture, and making the data collection process more concurrency, stable and scalable. Secondly, the collected real-time operating parameters of the inverter are systematically organized into a complete data set, and further sorted by multi-dimensional evaluation such as equipment efficiency index, load rate and operating time analysis of the inverter input power, so that the inverter power regulation priority is no longer based on static configuration, but is dynamically adjusted based on the operating performance and health status of each device, significantly improving the rationality and pertinence of the regulation. On this basis, the system executes an adaptive iterative algorithm that integrates edge computing and cloud collaborative processing for power allocation. Compared with the traditional single-point computing mode, it has stronger real-time response capabilities and algorithm optimization capabilities, and can dynamically balance the relationship between the total power control target and the power constraints of each inverter, reduce the number of invalid iterations, and improve convergence efficiency and computing stability. Furthermore, by encoding the target power value into standardized control instructions and distributing them to each inverter using a time-sharing and partitioning strategy, network congestion and communication conflicts that may be caused by batch transmission are avoided. The controllability and distribution efficiency of the control instructions are also improved, enabling the system to maintain high reliability and synchronization when executing large-scale instruction scheduling. More importantly, this method introduces blockchain technology to securely store the control logic in the form of smart contracts. Each version of the control logic is cryptographically signed and timestamped, ensuring that the execution process of the control instructions is traceable and tamper-proof. This effectively addresses the issues of control logic being easily modified and execution results being difficult to verify in existing systems, significantly enhancing the system's reliability, security, and management transparency. Finally, by analyzing the power regulation response data and establishing an encrypted communication link to enable two-way data exchange between the power plant and the grid dispatch system, the system can meet the real-time dispatch requirements of the grid while providing feedback on the execution results. This truly implements a closed-loop response structure for dispatch control, improving the overall automation level, response speed, and grid adaptability of the photovoltaic power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 This is a schematic diagram of an embodiment of a method for automatic photovoltaic power regulation in an embodiment of the present application;

[0018] Figure 2This is a schematic diagram of an embodiment of a photovoltaic power automatic adjustment system in an embodiment of the present application;

[0019] Figure 3 It is a schematic block diagram of the structure of the photovoltaic power automatic adjustment device in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The embodiments of the present application provide a method and system for automatic regulation of photovoltaic power. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or devices.

[0021] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In one embodiment of the photovoltaic power automatic adjustment method of the present application, the method includes:

[0022] Step S101: establishing a connection with all inverters in a photovoltaic power station through a distributed communication network, collecting real-time operating parameters of each inverter, and obtaining an inverter information data set;

[0023] Step S102: performing a multi-dimensional evaluation and sorting of the input power of each inverter according to the inverter information data set to obtain an inverter power priority sequence, wherein the multi-dimensional evaluation includes equipment efficiency index, load rate, and operating time analysis;

[0024] Step S103: Executing an adaptive iterative algorithm based on the inverter power priority sequence to perform power allocation calculation to obtain a target power value for each inverter, wherein the adaptive iterative algorithm integrates edge computing and cloud collaborative processing;

[0025] Step S104: Encode the target power value into a standardized control instruction, and send the control instruction to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data;

[0026] Step S105: construct a control logic security storage system using blockchain technology, convert the control logic into a smart contract format for storage, and obtain a tamper-proof control policy library;

[0027] Step S106: establishing an encrypted communication link based on the power regulation response data, realizing two-way data exchange between the power station and the grid dispatching system, obtaining a closed-loop control system, and transmitting the target power value to the on-site execution terminal.

[0028] It is understandable that the execution subject of the present application can be a photovoltaic power automatic regulation system, or a terminal or a server, which is not limited here. The embodiment of the present application is described by taking a server as the execution subject as an example.

[0029] Specifically, a distributed communication network connects all inverters in the PV power plant, collecting real-time operating parameters from each inverter and generating an inverter information dataset. A distributed communication network divides multiple inverters into several zones based on geographic location. Inverters within each zone are connected via local communication nodes, forming a local subnet. Each subnet is then connected to a central control server via a master communication node. The communication protocol can utilize the RS485 bus or be based on the MODBUS or IEC61850 standards to ensure interoperability with inverters from different manufacturers. During data collection, the system sends data requests to each inverter based on a pre-defined polling cycle and event trigger conditions. The inverter responds with real-time parameters, including input power, output power, DC voltage, AC current, and housing temperature. The collected raw data is first standardized, converting data from different protocol formats into a unified internal field structure. For example, the voltage field defined in the MODBUS protocol and the voltage field in IEC61850 are mapped to the internal unified field "DC_Voltage." After data standardization, outlier removal is performed. For example, if the input power in a data acquisition is negative or exceeds twice the rated power of the inverter, it is determined to be an outlier and removed. Subsequently, the standardized data is aligned according to the acquisition timestamp to ensure the consistency of data from multiple inverters within the same time window, forming the original data matrix. To reduce storage pressure, dimensionality reduction and compression algorithms are applied. For example, principal component analysis (PCA) is used to extract the three main components of input power, conversion efficiency, and device temperature, compressing redundant features and retaining only the parameters with the greatest impact on subsequent evaluation to obtain a feature parameter matrix. This matrix is ​​then stored in a distributed database according to the time series index, and a device ID and timestamp index is established for each record.

[0030] Based on the inverter information dataset, a multi-dimensional evaluation and ranking of each inverter's input power is performed to determine the inverter power priority sequence. This multi-dimensional evaluation includes analysis of the device efficiency index, load factor, and operating time. The device efficiency index is the ratio of the inverter's output power to its input power, normalized using a temperature correction factor. For example, when an inverter has an input power of 5000W, an output power of 4800W, and a temperature of 45°C, the normalized efficiency score is 0.95, corrected for high-temperature efficiency degradation based on an empirical model. The load factor is calculated as the ratio of real-time output power to the device's rated power. For example, if an inverter with a rated power of 6000W and a current output power of 4800W has a load factor of 80%, this is mapped to a load score using a piecewise function. The operating time analysis is based on a weighted sum of the device's total operating hours and the number of historical failures. For example, if the device has operated for 5000 hours and experienced two failures, a weighted formula is used to calculate the life balance factor. All scores are combined into a comprehensive score according to preset weights. For example, a weight of 0.5 is set for the efficiency score, 0.3 for the load score, and 0.2 for the lifetime balance coefficient. The final weighted sum is used to obtain the comprehensive score. An improved quick sorting algorithm is executed based on the comprehensive score. If the current regulation direction is to increase power, the inverters are sorted from high to low according to the comprehensive score, forming a power priority sequence.

[0031] An adaptive iterative algorithm is used to calculate power allocation based on the inverter power priority sequence, determining the target power value for each inverter. This adaptive iterative algorithm integrates edge computing and cloud-based collaborative processing. The algorithm first receives a grid dispatch instruction containing the target total power value. For example, if the current total power is to be adjusted from 600kW to 550kW, the power adjustment is -50kW. Based on the inverter sorting order, the power of the front-row inverters is initially reduced proportionally to form an initial power allocation table. This initial allocation table is transmitted to the edge computing node, where power constraints are checked. For example, if the minimum output power of an inverter must not fall below 1000W, the initial allocation is checked for violations. Inverters that do not meet the constraints are marked as abnormal and a secondary allocation correction is performed. Subsequently, an adaptive step-size gradient descent optimization algorithm is run on a cloud-based high-performance computing cluster. With the objective function of reducing the deviation between the current total power and the target power, the allocation to each inverter is dynamically adjusted until convergence conditions are met. To improve accuracy, the historical response characteristics of the inverter are combined for response prediction during iteration. For example, if the historical response of an inverter lags by 30 seconds, a buffer zone is reserved in the allocation to fine-tune its target power.

[0032] The target power value is encoded into a standardized control instruction, and the instruction is sent to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data. First, the target power value is adapted into a corresponding format instruction set based on the inverter manufacturer and communication protocol. For example, if manufacturer A uses the MODBUS standard and manufacturer B uses a proprietary protocol, corresponding instruction packets are generated for each. Each instruction packet is attached with a timestamp, CRC checksum, and validity period field. It is sent in batches and zones based on the geographical location and inverter load. For example, the inverters in area 1 are received within 0-10 seconds to avoid centralized communication conflicts. After the instruction is sent, the inverter is required to return a confirmation signal. If the confirmation fails, it will be resent. The monitoring module deployed on site records the power change data of each inverter in real time, monitors the response delay, change slope, and stable value, and forms a response data stream.

[0033] A secure control logic storage system is constructed using blockchain technology. This control logic is converted into smart contracts for storage, creating a tamper-resistant control policy library. The control logic is functionally divided into five modules: data acquisition, sorting calculations, power distribution, command issuance, and response monitoring. Standardized interfaces define data flow relationships within each module. The logic flow graph is translated into an intermediate language instruction set, which then generates executable code. Each code segment is hashed and digitally signed, forming a transaction unit and uploaded to the blockchain network. Verified by consensus across all nodes, the transaction is then linked into the blockchain, ensuring the immutability and traceability of the control logic version.

[0034] An encrypted communication link is established based on the power regulation response data, enabling two-way data exchange between the power plant and the grid dispatching system, and transmitting the target power value to the on-site execution terminal. A two-way authenticated communication channel is established using the TLS protocol, and regulation response data is encrypted according to data sensitivity. For example, grid commands and inverter response data are encrypted using AES-256. The power plant's current status report is regularly uploaded, and new commands from the dispatching system are received. After the command's legitimacy is verified, it is integrated and updated with the local target power. The final command is then distributed to on-site I / O measurement points and power execution terminals via industrial bus protocols such as CANopen or PROFIBUS, forming a complete closed loop.

[0035] In the embodiment of the present application, a connection is established with all inverters in the photovoltaic power station through a distributed communication network, which ensures the system's wide adaptability to inverters of different manufacturers and models, structurally avoiding the bottlenecks and fault concentration problems existing in the centralized communication architecture, making the data collection process more concurrency, stability and scalability. Secondly, the collected real-time operating parameters of the inverter are systematically organized into a complete data set, and further sorted by multi-dimensional evaluation such as equipment efficiency index, load rate and operating time analysis of the inverter input power, so that the inverter power regulation priority is no longer based on static configuration, but is dynamically adjusted based on the operating performance and health status of each device, which significantly improves the rationality and pertinence of the regulation. On this basis, the system executes an adaptive iterative algorithm that integrates edge computing and cloud collaborative processing for power allocation. Compared with the traditional single-point computing mode, it has stronger real-time response capabilities and algorithm optimization capabilities, and can dynamically balance the relationship between the total power control target and the power constraints of each inverter, reduce the number of invalid iterations, and improve convergence efficiency and computing stability. Furthermore, by encoding the target power value into standardized control instructions and distributing them to each inverter using a time-sharing and partitioning strategy, network congestion and communication conflicts that may be caused by batch transmission are avoided. The controllability and distribution efficiency of the control instructions are also improved, enabling the system to maintain high reliability and synchronization when executing large-scale instruction scheduling. More importantly, this method introduces blockchain technology to securely store the control logic in the form of smart contracts. Each version of the control logic is cryptographically signed and timestamped, ensuring that the execution process of the control instructions is traceable and tamper-proof. This effectively addresses the issues of control logic being easily modified and execution results being difficult to verify in existing systems, significantly enhancing the system's reliability, security, and management transparency. Finally, by analyzing the power regulation response data and establishing an encrypted communication link to enable two-way data exchange between the power plant and the grid dispatch system, the system can meet the real-time dispatch requirements of the grid while providing feedback on the execution results. This truly implements a closed-loop response structure for dispatch control, improving the overall automation level, response speed, and grid adaptability of the photovoltaic power plant.

[0036] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0037] Establishing a multi-level communication topology structure, performing zoning management on the inverters in the photovoltaic power station according to geographical locations, and obtaining inverter regional groups;

[0038] Assigning differentiated communication frequencies to the inverter area groups, executing a collection strategy combining polling and event triggering, and obtaining an original parameter collection stream;

[0039] The original parameter acquisition stream is subjected to data format standardization processing, outliers are removed, and timestamps are aligned to obtain a preprocessed data set;

[0040] The preprocessed data set is fed into a data compression algorithm for dimensionality reduction processing to reduce redundant information and retain key features, thereby obtaining a compressed data stream;

[0041] Applying a feature extraction algorithm to the compressed data stream to calculate the input power, conversion efficiency, and health status index of each inverter to obtain a feature parameter matrix;

[0042] The characteristic parameter matrix is ​​stored in a distributed database according to a time series, and a parameter indexing mechanism is established to form the inverter information data set.

[0043] Specifically, the multi-level communication topology is planned based on the physical location of the PV power plant. All inverters are divided into several zones, each equipped with a local communication node or subnet controller to manage communication tasks for all inverters within that zone. These local communication nodes are then connected to a central control unit via optical fiber or industrial Ethernet, forming a bottom-up, multi-level tree-like structure. This structure effectively reduces communication conflicts, improves data collection manageability, and enhances concurrent processing capabilities.

[0044] Based on this structure, the system sets differentiated communication frequencies for each inverter regional grouping. Specifically, different communication time windows and data reporting cadences are configured for different partitions, thereby avoiding communication bandwidth bottlenecks caused by a large number of devices uploading data simultaneously. The system uses a combination of polling and event triggering for data collection and control. Polling involves the central control unit periodically sending status read requests to each inverter, while the event triggering mechanism immediately initiates data collection tasks upon detecting a sudden change in device power or abnormal status. This results in a complete raw parameter collection stream, including fields such as timestamp, device ID, and various physical parameters.

[0045] The collected raw parameters first need to be standardized. Because data formats and field naming differ among devices from different manufacturers, all data must be converted to a standard structure within the platform through protocol adaptation and field mapping rules. For example, different devices may label DC input power as "DC_Power," "InputPower," or "Pdc." The standardization process uniformly identifies it as "input_power" and ensures consistent units for all values ​​(for example, watts). The system then removes outliers from all collected data. Rules such as negative power values, zero voltage, and sudden increases in rate of change exceeding a certain limit are considered invalid and removed from the dataset. To ensure time consistency across data, all data is aligned using the device's local time or the control system's timestamp. Data from multiple inverters within the same time segment is assembled into a multidimensional vector, forming a preliminary preprocessed dataset. After preprocessing, the data volume remains large, requiring compression to improve subsequent computational efficiency and reduce transmission load. The dimensionality reduction and compression strategy employed here assesses the importance of each dimension of data and retains only the data dimensions most influential in sorting and adjustment decisions. Parameters with less influence on adjustment logic, such as input voltage, input current, and device temperature, are merged or removed, retaining only core metrics such as input power, conversion efficiency, and response time. This generates a compressed data stream. This compressed data is smaller in size but retains key features, making it suitable for rapid analysis and processing.

[0046] The system applies a feature extraction algorithm to the compressed data stream. The goal of this process is to calculate quantifiable and comparable performance indicators. The feature extraction algorithm generates typical performance parameters for each inverter by performing sliding window operations, statistical analysis, and normalization on the time series data. Input power is the average input power during the current time period; conversion efficiency is calculated as the ratio of output to input, taking into account a temperature correction factor; and health status indicators are derived from a weighted evaluation function of parameters such as historical operating time, number of failures, and operational stability. These characteristic data for all inverters are integrated in a unified format to form a characteristic parameter matrix, with each row corresponding to an inverter and each column corresponding to a characteristic parameter.

[0047] The characteristic parameter matrix is ​​written to a distributed database in a time series format. The database is managed using a key-value index structure, with the primary key consisting of the inverter ID and timestamp to ensure efficient data retrieval. For frequently queried fields such as device status and input power, the system establishes a multidimensional index structure to accelerate sorting and filtering. This completes the inverter information dataset, encompassing key operating indicators for all inverters within a specific time period, with excellent temporal continuity, data integrity, and analytical usability. This dataset serves as the fundamental data input source for the entire PV power automatic regulation process, providing comprehensive data support for subsequent sorting, power allocation, and regulation control.

[0048] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0049] Extracting current power data, historical operation records, and environmental parameters of each inverter from the inverter information data set to construct an evaluation basic data table;

[0050] Calculating the device efficiency index for each inverter in the evaluation basic data table, normalizing the output power to input power ratio in combination with the temperature correction coefficient to obtain an efficiency score;

[0051] Calculating the current load rate of each inverter based on the inverter information data set, performing piecewise function mapping on the ratio of real-time power to rated power to obtain a load score;

[0052] Performing a time series analysis on the historical operating data in the inverter information data set, calculating a weighted sum of the cumulative operating time and the failure rate, and obtaining a life balance coefficient;

[0053] Determine the weight coefficient of each scoring indicator according to a preset control mode selector, and perform weighted fusion on the efficiency score, load score and life balance coefficient to obtain a comprehensive score;

[0054] An improved quick sorting algorithm is executed on the comprehensive scores to determine ascending or descending order according to the current regulation direction to form the inverter power priority sequence.

[0055] Specifically, the logical correspondence between these parameters in the data structure is clearly defined. Current power data includes the inverter's input power and output power, representing the DC power received from the PV panels and the AC power output to the grid, respectively, and serves as the basis for energy efficiency evaluation. Historical operating records, including cumulative operating time, number of failures, and maintenance records, reflect the inverter's stability and service life. Environmental parameters, such as operating temperature and light intensity, directly impact its performance. The system extracts these parameters in batches by accessing structured data indexed by timestamp and device ID from a distributed database and integrates them into a horizontally scalable basic evaluation data table. Each row in the table corresponds to an inverter, and each column is a feature field, ensuring a complete structure and facilitating vectorized processing. To calculate the device efficiency index, the current output power and input power values ​​are first taken and divided to obtain the instantaneous energy conversion efficiency. Because inverter performance degrades at high temperatures or in extreme environments, this efficiency value must be corrected based on the ambient temperature. This correction is based on a known temperature-efficiency relationship model, normalizing the actual efficiency value to a reference state at a standard temperature. For example, if the temperature is too high, the system will extract the corresponding correction factor from the model based on the temperature deviation and adjust the efficiency value to eliminate non-equipment factors that interfere with the efficiency of the environment. After normalization, the efficiency score is obtained, and its value range is usually limited to between 0 and 1 to unify the subsequent scoring calculations.

[0056] The calculation of the current load rate depends on the ratio of the real-time power value to the rated power. The rated power is the maximum output capacity set by the inverter at the factory and is a fixed value. The current output power is divided by the rated power to obtain the actual load rate, which reflects the operating intensity of the inverter at the current moment. To avoid the impact of nonlinear scoring, the ratio is mapped through a piecewise function. For example, a load rate below 30% is considered light load, between 30%-70% is medium load, and above 70% is high load. Each section is assigned a different scoring interval to reflect the load balance status of the equipment operation. The piecewise function design takes into account the impact of the long-term operating load of the inverter on its life. The scoring results show a nonlinear distribution and can more realistically reflect the operating status.

[0057] The calculation of the life balance coefficient focuses more on analyzing long-term performance stability. The system extracts the cumulative operating hours and number of failures from the inverter's historical data records and constructs a time series analysis window, for example, covering the last 180 days. Within this window, the average daily operating time and monthly failure frequency of the device are calculated and weighted to form a lifespan indicator. In this weighted design, operating hours are given a higher weight, while failure rates are weighted more heavily based on their negative impact on device health. The result is a comprehensive lifespan score, with higher values ​​indicating healthier devices.

[0058] After calculating the three aforementioned metrics, the system activates the control mode selector based on the current control mode, selecting the weightings for the efficiency score, load score, and life factor. For example, in efficiency-first mode, the efficiency score might account for 60%, with the load score and life factor each accounting for 20%. In life-balanced mode, the life score's weighting increases to 50%. The three metrics are weighted and summed according to the selected weighting factors to generate a unified, comprehensive score, ensuring that the score conforms to the regulatory strategy. All scores are normalized to ensure consistency in the dimensions and magnitudes of the different metrics, facilitating integration.

[0059] Finally, the system applies an improved quick sort algorithm to the comprehensive scores. Traditional quick sort has good time complexity on average for large datasets, but performance degrades when scores are close or the data exhibits local order. Therefore, this implementation introduces an adaptive benchmark selection strategy, prioritizing points near the median or mean of the scores as sorting benchmarks to improve sorting efficiency. The sorting direction is determined by the regulation objective of the grid dispatch instruction. If the goal is to increase total power output, inverters with high comprehensive scores are prioritized; if the goal is to reduce output, inverters with low scores are prioritized. After the sorting is complete, a power priority sequence for each inverter is generated in order. Each inverter includes its device ID, current power status, and ranking position. This sequence serves as the basic input for subsequent iterative power calculations. The entire process is completed within a structured data framework, ensuring logical clarity and consistency between data, avoiding control logic anomalies caused by score fluctuations, and thus ensuring the stable operation of the entire automatic regulation mechanism.

[0060] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0061] Receive grid dispatch instructions or preset control targets, determine the total target power value of the photovoltaic power station, and calculate the difference between it and the current total power to obtain the power regulation amount;

[0062] Combining the power adjustment amount with the inverter power priority sequence, executing an initial power allocation strategy, allocating an initial target power value to each inverter, and obtaining an initial allocation plan;

[0063] Deploy a constraint checking algorithm on the edge computing unit to verify the upper and lower limit constraints and the rate of change constraint of the inverter power for the initial allocation scheme, mark the allocation values ​​that violate the constraints, and obtain the constraint verification result;

[0064] Based on the constraint verification results, an adaptive step-size gradient descent algorithm is started to iteratively optimize power allocation through a cloud-based high-performance computing cluster to obtain an optimized allocation solution;

[0065] Inputting the optimized allocation scheme into a response prediction model, wherein the response prediction model predicts the actual response behavior of each inverter based on historical response data to obtain a response prediction result;

[0066] The optimized allocation scheme is fine-tuned according to the response prediction result to ensure the total power control accuracy and load balance of each inverter, and finally form the target power value.

[0067] Specifically, three dimensions of data are extracted from the inverter information dataset to construct the basic evaluation data table: current power data, historical operating records, and environmental parameters. Current power data refers to the inverter's input power (typically DC input from photovoltaic modules) and output power (AC output after inverter conversion) at the sampling moment. Historical operating records include accumulated operating hours, fault records, and maintenance cycles. Environmental parameters include physical quantities such as the real-time temperature, irradiance, and humidity at the device's location that directly impact device performance. The system retrieves structured records from the dataset and, using the inverter ID as the primary key, assigns these three data types to fields in the data table, forming the basic evaluation data units. This ensures both horizontal and vertical temporal consistency of the data structure, resulting in the basic evaluation data table. This table serves as the direct carrier for multi-dimensional evaluation, and a data identification mechanism ensures the reliable source of each field and a clear update cycle. To calculate the device efficiency index, the system first selects the input power and output power fields from the basic evaluation data table and uses the ratio of the two to derive a preliminary energy conversion efficiency. Since the actual operating ambient temperature will affect the efficiency of the inverter, especially under high temperature conditions, the equipment often has the problem of derating operation. Therefore, it is necessary to correct the ratio in combination with the temperature correction coefficient. The correction coefficient is determined based on the temperature-efficiency relationship model provided in the inverter's technical specification, or mapped based on the empirical curve function obtained by fitting historical data. For example, after the ambient temperature exceeds a certain threshold, the efficiency will decrease by a certain percentage for every degree Celsius increase. The correction coefficient is applied to the preliminary ratio to obtain the temperature-corrected energy efficiency value. To facilitate the subsequent fusion of multiple indicators, the correction value must also be normalized so that it is expressed within a unified dimension range (such as 0 to 1), ultimately forming an efficiency score.

[0068] The load score is calculated based on the ratio of the current real-time power to the inverter's rated power, defined as the current load factor. Rated power is a fixed value, representing the device's maximum output capacity under design conditions. The load factor reflects the inverter's current workload. After obtaining the real-time power and rated power, the system uses the load factor to map a pre-set piecewise function to a range. For example, a load factor below 30% is considered light load, with a lower score assigned; between 30% and 70% is medium load, with an intermediate score assigned; and above 70% is high load, with a high score assigned. This piecewise function is based on long-term operating experimental data. Taking into account the effects of heat generation and aging within different load ranges, the score is not linearly assigned but rather a step-by-step approach to reflect the nonlinear impact of the load factor on operational management. The final load score is also normalized to a standard value for easy integration with other indicators. The calculation of the life balance coefficient relies on historical operating data. The system uses time series analysis to extract the inverter's cumulative operating time and fault records during that period to calculate its operational stability. Operating time reflects the intensity of equipment usage, while failure rate reveals its reliability. The system sets an analysis time window, such as the last 365 days, to count the daily operating hours of the inverter and the frequency of failures. A preset weighting function is used to combine operating time and failure rate into a life balance value. For example, stable operating equipment is given a higher score, while equipment with frequent failures is assigned a lower life score. This process proceeds in the form of a sliding window on the timeline, ensuring that the evaluation reflects the current stability trend of the equipment rather than static characteristics. All life assessment data is also normalized and standardized to output a life balance coefficient.

[0069] After completing the calculation of the three scoring items, the system determines the weight coefficient of each indicator under the current control strategy through the preset control mode selector. Control modes such as "efficiency priority", "life balance" or "load balance" have different preset weight profiles for each mode. For example, in efficiency priority mode, the efficiency score weight can be 0.6, the load score is 0.3, and the life score is 0.1; while in life balance mode, the life score may account for 0.5 weight. The weight coefficient is uniformly retrieved from the strategy configuration file, applied to the three scores of the current inverter, and the comprehensive score is calculated through the weighted fusion formula. Each inverter finally obtains a unified comprehensive score value, which is used to quantify its priority under the current regulation strategy.

[0070] After the comprehensive scoring is completed, the system passes the scoring vector to the improved quick sort algorithm for sorting. Although traditional quick sorting has good time efficiency in general scenarios, in this scenario, due to the large amount of data and the potentially small differences in the scores of each inverter, there are problems with sorting basis selection bias and uneven recursion depth. Therefore, a quick sort optimization algorithm that introduces the median sampling method as the partitioning basis is adopted. This method first selects the local median of the scoring data as the benchmark before sorting, effectively reducing the impact of extreme values ​​and improving partition balance. The choice of sorting direction is determined by the regulation objective. If the goal is to increase output, inverters with high comprehensive scores are ranked first and power is allocated first. If the goal is to reduce output, inverters with low comprehensive scores are ranked first and output is reduced first. After sorting, an inverter power priority sequence is formed. Each inverter corresponds to its unique identifier, current operating status, and sorting position information. This sequence will serve as the core input for the next power iteration calculation, ensuring optimal resource utilization and a reasonable allocation strategy during the regulation process. The entire sorting logic relies on the logical accuracy and data consistency of the previous scoring. Only by ensuring the rigorous processing of each scoring step can the final sorting result be interpretable and executable in terms of algorithmic adjustment priority.

[0071] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0072] Performing protocol adaptation processing on the target power value, converting it into a corresponding instruction format according to the communication protocol type of each inverter, and obtaining a protocol-compatible instruction set;

[0073] Adding a timestamp, a checksum, and instruction validity period information to the protocol-compatible instruction set to form a complete control instruction packet, and calculating a transmission priority to obtain an instruction sending queue;

[0074] Based on the network topology and load conditions, the instruction sending queue is batched according to geographical partitions and time windows to obtain a time-sharing and partitioned delivery plan;

[0075] Execute the time-sharing and partition-based distribution plan, transmit the control command to the target inverter through a secure communication channel, and request a return receipt confirmation message to obtain the command transmission status;

[0076] Deploy distributed monitoring points to collect data on the power changes of each inverter after receiving instructions, including response time, change curve, and stable value, to form a real-time response flow;

[0077] Data aggregation and anomaly detection are performed on the real-time response stream, the deviation between the actual power and the target power is calculated, and a power regulation response data report is generated.

[0078] Specifically, the determined target power value first requires protocol adaptation. Because inverters from different manufacturers and models support varying communication protocols, including MODBUS, IEC61850, CANopen, and some manufacturers' proprietary protocols, the target power value, in a unified format, must be converted into a specific command format recognized by each inverter. The system uses an internally maintained protocol mapping table and, based on the communication protocol identifier of each inverter, invokes the corresponding adapter module to parse the generic target power command into a specific protocol data frame. For example, in the MODBUS protocol, the target power value must be written to a specified register address and encapsulated in little-endian byte order, while in the IEC61850 protocol, the setting command must be sent via the MMS message format. This adaptation process creates a protocol-compatible command set tailored to each inverter. After obtaining the protocol-compatible command set, the system further adds a timestamp, checksum, and command validity period information to each command. The timestamp records the standard time when the instruction is generated, ensuring that subsequent devices can verify whether it is the latest command after receiving the instruction; the check code is calculated and generated based on the instruction content using the CRC16 or CRC32 algorithm, and is used to verify the data integrity after the inverter side receives the instruction; the instruction validity period field clearly indicates the maximum delay range within which the control command can be executed on the inverter side, preventing the instruction from being executed despite being invalid due to network delays or communication anomalies. After completing the encapsulation of these additional information, each instruction forms a complete control instruction package. The system calculates the instruction transmission priority based on the current operating status of the inverter, the network load status and the geographical location. For example, the inverter with high load and geographical location has a higher priority. In this way, the instruction sending queue is generated and the sending priority and sending window of each instruction are marked in the queue.

[0079] Based on the network topology and current load conditions, the system batches the instruction sending queue according to geographical partitions and time windows to form a time-sharing and partitioned delivery plan. Geographical partitioning refers to dividing the entire power station into several areas based on the physical layout of the inverters. Each area receives control instructions in batches within a specified time window to avoid excessive instantaneous network bandwidth usage or local communication congestion. The division of time windows is dynamically adjusted based on real-time network load monitoring data. For example, when the network is idle, the batch size within the window is increased, and when the load is high, the sending cycle is extended and the number of concurrent requests is reduced. The time-sharing and partitioned delivery plan defines the specific sending time, target inverter address, and confirmation mechanism parameters for each instruction.

[0080] When executing the time-sharing and partitioned distribution plan, all instructions are sent to each target inverter through a preset secure communication channel. The secure communication channel usually uses TLS (Transport Layer Security) encryption to ensure that the instructions are not tampered with or stolen during transmission. During the transmission process, each time an instruction is sent, the system requires the target inverter to return a receipt confirmation message. The confirmation message contains the instruction verification pass flag and the reception timestamp. If no confirmation is received within the set timeout period, the system automatically resends the instruction and retries a maximum of the specified number of times. If the number of retries fails, an exception is reported and subsequent compensation processing is performed. Through the reception confirmation mechanism, the system can grasp the transmission status of each instruction in real time, ensuring the consistency and reliability of the control instructions.

[0081] During the instruction transmission and execution process, the system deploys distributed monitoring points on-site. These monitoring points are responsible for collecting real-time data on the inverter's power change after receiving the instruction. The collected data includes the inverter response time (i.e., the time required from receiving the instruction to the start of the power change), the power change process curve (the trajectory of power changes over time), and the stable value (the stable output value after the power change is completed). All collected data is streamed in real time to the central control server, forming a continuous real-time response stream. Each response stream uses the inverter ID, timestamp, and power data as the basic unit to ensure the timing and integrity of the data stream.

[0082] Data aggregation and anomaly detection in real-time response streams are crucial steps in ensuring effective system regulation. The system aggregates all response data from the same inverter within a regulation cycle to form a response curve, which is then compared and analyzed against the issued target power value. Deviation calculation quantitatively measures the difference between the actual stable output power and the preset target power, using either absolute or relative deviation models. For example, if the target power is 4500W and the actual power is 4400W, the absolute deviation is 100W and the relative deviation is 2.22%. The system also sets anomaly detection threshold. If the deviation exceeds 5% or the response delay exceeds a set standard, the inverter is flagged as an abnormally responsive device, triggering compensation mechanisms or generating an early warning. All deviation calculation results, response curve analysis, and anomaly detection flags are integrated into a standardized data reporting format to generate a power regulation response data report. This report provides direct factual basis for subsequent optimization of regulation strategies, dynamic adjustment of inverter sequencing, and iterative power allocation. The entire process relies on four core steps: command adaptation, transmission confirmation, response monitoring and anomaly judgment. The data flow is coherent and the processing logic is clear, ensuring the system's stable control capability and efficient response performance in a dynamic environment.

[0083] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0084] The control logic is divided into data acquisition, sorting calculation, power distribution, command issuance and feedback processing according to function, forming a modular control structure;

[0085] Defining standardized interface specifications for each functional module in the modular control structure, establishing data flow relationships between modules, and obtaining a control logic flow graph;

[0086] Converting the control logic flow graph into a visual expression based on function blocks and mapping it into an intermediate layer language instruction set to obtain an executable control code;

[0087] Encrypting and signing the executable control code, combining it with the timestamp and operator identity information, and packaging it into a smart contract transaction unit to obtain a blockchain transaction package;

[0088] Submitting the blockchain transaction package to the distributed consensus network for verification, confirming the validity of the transaction through the proof-of-stake mechanism, and obtaining a confirmed control logic block;

[0089] The confirmed control logic block is linked to the main chain and synchronized between nodes to form an anti-tampering control policy library with version traceability and access permission control.

[0090] Specifically, the overall control logic is functionally divided to form a modular control structure. Specifically, according to the logical order of data flow and control steps, the control logic is divided into five independent modules: the data acquisition module, the sorting calculation module, the power allocation module, the instruction issuance module, and the feedback processing module. The data acquisition module is responsible for collecting necessary operating parameters from the inverter in real time; the sorting calculation module evaluates and sorts the inverter priorities based on the collected data; the power allocation module allocates target power based on the sorting results and scheduling objectives; the instruction issuance module sends the calculation results to each inverter via adapted protocol instructions; and the feedback processing module is responsible for collecting and analyzing response data. This division ensures that each module has a single internal function and clear external dependencies, laying the foundation for subsequent standardization and visual mapping. Subsequently, for each functional module in the above modular structure, the system formulates standardized interface specifications to clarify the input and output data formats, data types, trigger mechanisms, and exception handling protocols. For example, the data output from the data acquisition module is formatted as a timestamp, device ID, and parameter list. The sorting module requires a complete characteristic parameter matrix as input and outputs a sorted list. The power allocation module receives the sorted list and the target total power as input and outputs the target power value for each inverter. By defining these standardized interfaces, the system can clearly describe the data flow relationship between modules and graphically establish the overall control logic flow diagram. This control logic flow diagram clearly defines the path, sequence, and data dependencies of data flow between modules, thereby achieving seamless connection and collaborative operation between modules.

[0091] After establishing a complete control logic flow chart, the system further converts it into a visual expression based on function blocks. Function blocks refer to abstracting each module into standard graphic elements, such as rectangles, diamonds or connecting lines, and arranging them in a logical order and data flow to form a clear visual control process. Each function block encapsulates the processing logic, input and output interfaces and state transfer rules of the corresponding module. After completing the visual modeling, the system translates each function block based on the rules of the intermediate layer language and maps it into an intermediate layer instruction set. For example, the logical relationship is expressed in a process language or a block instruction language (such as FBD - Function Block Diagram Description Language) to obtain executable control code. The intermediate layer language has the characteristics of platform independence and standardization, which enables the control logic to be understood and executed by multiple execution environments, while retaining the ability to trace and debug the logic.

[0092] To prevent subsequent tampering and forgery, the system cryptographically signs the generated executable control code. This signing process involves first hashing the control code content to generate a unique digest value. This digest value is then encrypted using the authorized operator's private key to form a digital signature. This signature is then affixed with a timestamp and the operator's identity, encapsulating the complete smart contract transaction unit. This transaction unit contains not only the control code itself but also metadata such as version number, generation time, operator ID, and expiration date, ensuring that each copy of the control code has a unique and verifiable identity. Once the smart contract transaction unit is packaged, the system submits it to the distributed consensus network for verification. Nodes in the consensus network perform consensus verification based on a Proof of Stake (PoS) mechanism. Nodes holding sufficient stake verify the authenticity, legitimacy, and consistency of the transaction unit. Upon successful verification, the system generates a confirmed control logic block, which encapsulates the verified control instruction set and its signature information, preparing for subsequent block links.

[0093] The system links the confirmed control logic block to the existing blockchain main chain and synchronizes the blocks through all nodes in the network. During the synchronization process, each node verifies the connection between the hash value of the new block and the previous block to ensure the continuity and immutability of the chain structure. In this way, a complete tamper-proof control policy library is formed, and the generation, verification, release, and call of each version of the control logic can be fully recorded and traced. At the same time, based on the blockchain's access control mechanism, only authorized users or systems can read or update the control policy library, further improving the security, reliability, and compliance of the photovoltaic power automatic regulation system. This process ensures that every link of the control logic, from generation to application, is auditable, verifiable, and protected against malicious modification, greatly enhancing the credibility of the entire photovoltaic power plant automation operation system.

[0094] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0095] Marking the power regulation response data as sensitive information, applying a multi-level encryption strategy to perform differential encryption processing on data of different security levels to obtain a secure data packet;

[0096] Establish an encrypted communication tunnel based on TLS / SSL, adopt a two-way authentication mechanism to ensure the legitimacy of the identities of both ends of the communication, and form a secure communication link;

[0097] Regularly sending power plant status reports, including current total power, adjustable capacity, and response capability assessment, to the power grid dispatch system via the secure communication link to obtain power plant status feedback;

[0098] Receiving dispatch instructions issued by the power grid dispatch system through the secure communication link, performing legality verification and executability analysis on the instructions, and obtaining dispatch requirements;

[0099] Converting the scheduling requirements into internal control parameters of the photovoltaic power station, coordinating and integrating them with the target power value to form a final execution instruction;

[0100] The final execution instruction is transmitted to the IO measurement point and the execution terminal through the industrial field bus, and a data closed-loop verification mechanism is established to form a closed-loop control system.

[0101] Specifically, the overall control process is functionally divided to form a modular control structure. Specifically, the entire control logic is split into five functional modules based on data flow paths and processing sequences: the data acquisition module, the sorting and calculation module, the power allocation module, the command issuance module, and the feedback processing module. The data acquisition module is responsible for collecting real-time parameters such as input power, output power, voltage, current, and temperature from each inverter in the power station, generating a standardized real-time data stream. The sorting and calculation module extracts characteristic parameters such as equipment efficiency index, load factor, and lifespan coefficient from the collected data, performs comprehensive scoring and priority sorting, and generates a power adjustment order for the inverters. The power allocation module performs adaptive iterative calculations based on the sorting results and the target total power issued by the dispatcher to determine the target power value for each inverter. The command issuance module performs protocol adaptation processing on the target power values, encapsulating them into control commands according to the communication protocol standards supported by each inverter, and sending them to each inverter via a network time-sharing and partitioning strategy. The feedback processing module receives the response data from each inverter to the control commands, monitors the adjustment execution process in real time, and performs deviation detection and anomaly analysis. This modular design allows each module to focus on a single function. Modules are connected through clear data interfaces, ensuring clear logic, functional independence, and system scalability, laying the foundation for subsequent standardized descriptions and visual mapping. After establishing the modular structure, the system needs to define standardized interface specifications for each functional module to ensure consistent data exchange between modules. These interface specifications define the input and output data structures, data types, trigger conditions, and exception handling logic. For example, the output data of the data acquisition module is standardized as a combination of fields (such as device ID, timestamp, input power, output power, and temperature). The input of the sorting module requires a complete matrix of characteristic parameters, and the power allocation module receives a sorted list of inverter IDs and a total power regulation target value. Through these standardized interface descriptions, the system can logically establish the data flow relationships between modules, forming a control logic flow diagram. This flow diagram not only identifies the input and output relationships between modules but also clearly defines key control parameters such as the data flow sequence, synchronization mechanism, and data update frequency. This ensures unambiguous data flow throughout the entire process, avoiding logical loops or information loss.

[0102] After establishing the control logic flow graph, the system further converts it into a visual representation based on function blocks (FBs). Each FB represents a specific functional module or sub-processing unit, such as collecting power data, calculating load factors, allocating target power, and issuing commands. Function blocks are connected via standardized interfaces, forming a clear flow diagram. Each FB defines its processing logic in detail, including input conditions, data processing algorithms, state transition rules, and output formats. For example, the power allocation FB defines how to receive a sorted list and scheduling target values, perform power allocation calculations, and output the new target power values ​​for each inverter. After completing the visual modeling, the system maps each FB and its internal logic into an intermediate-layer instruction set based on the rules of an intermediate-layer language (such as FBD—Function Block Diagram Description Language). This generates platform-independent, clearly structured, and easily compilable executable control code. This intermediate-layer instruction set enables intuitive editing of control logic within the human-machine interface (HMI) while efficiently executing on different underlying hardware or software platforms, ensuring excellent portability and maintainability.

[0103] After generating the complete executable control code, the system requires cryptographic signing of each version of the control code to prevent subsequent code tampering, forgery, or untraceability. The signing process involves first performing a hash operation on the control code content to generate a unique, fixed-length digest value that sensitively reflects any minor changes to the code. This digest value is then encrypted using the private key of an authorized operator (such as a power plant engineer or automation system) to form a digital signature. During the packaging process, the generation timestamp, operator ID, version number, and expiration date are also appended, ultimately packaging the code into a complete smart contract transaction unit. A smart contract, in this context, refers to an immutable logical unit that encapsulates specific control logic content and execution conditions, ensuring that each version of the control code has a clear, verifiable identity and lifecycle management.

[0104] The generated smart contract transaction unit is then submitted to the distributed consensus network for verification. The consensus process utilizes a Proof of Stake (PoS) mechanism, where nodes holding system stake (such as certified power plant control nodes or certified cloud servers) are responsible for verifying the authenticity and consistency of the transaction unit. This verification process includes verifying the legitimacy of the digital signature, confirming that the code content complies with standards, and checking that the transaction unit is consistent with the state of the existing chain. If verification passes, the system generates a new control logic block containing the verified control code, intermediate instruction set, operator information, and digital signature. The system then links the verified control logic block to the existing blockchain main chain. Before linking, each new block verifies that its hash value matches the hash value of the previous block, ensuring the integrity and immutability of the entire blockchain structure. The decentralized nature of blockchain requires all nodes to synchronously update new blocks and independently verify their legitimacy. This mechanism effectively eliminates the risk of single points of failure and malicious tampering. This approach creates a complete, tamper-resistant control policy library, documenting the generation time, operator, version evolution history, and application status of each version of the control logic. The control policy library is protected by a blockchain-based access control system, ensuring that only users or systems with appropriate authorization keys can read, update, or roll back the control logic. This significantly enhances the security, reliability, and audit compliance capabilities of the photovoltaic power automatic regulation system. This management approach ensures that the entire lifecycle of the control logic, from design, generation, verification, deployment, and application, is transparently auditable, non-repudiable, and technically tamper-proof, providing a solid and reliable foundation for the intelligent and automated operation of photovoltaic power plants.

[0105] The above describes the photovoltaic power automatic adjustment method in the embodiment of the present application. The following describes the photovoltaic power automatic adjustment system in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the photovoltaic power automatic adjustment system includes:

[0106] The connection module 201 is used to establish a connection with all inverters in the photovoltaic power station through a distributed communication network, collect real-time operating parameters of each inverter, and obtain an inverter information data set;

[0107] a sorting module 202 for performing a multi-dimensional evaluation and sorting of the input power of each inverter according to the inverter information data set to obtain an inverter power priority sequence, wherein the multi-dimensional evaluation includes equipment efficiency index, load rate, and operating time analysis;

[0108] An allocation module 203 is configured to execute an adaptive iterative algorithm based on the inverter power priority sequence to perform power allocation calculations to obtain a target power value for each inverter, wherein the adaptive iterative algorithm integrates edge computing and cloud collaborative processing;

[0109] The sending module 204 is used to encode the target power value into a standardized control instruction, and send the control instruction to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data;

[0110] The conversion module 205 is used to construct a control logic security storage system using blockchain technology, convert the control logic into a smart contract format for storage, and obtain a tamper-proof control policy library;

[0111] The encryption module 206 is used to establish an encrypted communication link based on the power regulation response data, realize two-way data exchange between the power station and the power grid dispatching system, obtain a closed-loop control system, and transmit the target power value to the on-site execution terminal.

[0112] Through the collaborative efforts of these components, a distributed communication network connects all inverters within the PV power plant, ensuring the system's broad adaptability to inverters of varying manufacturers and models. This structurally avoids the bottlenecks and concentrated faults inherent in centralized communication architectures, resulting in greater concurrency, stability, and scalability in the data collection process. Secondly, the collected real-time operating parameters of the inverters are systematically organized into a complete data set. Further, through multi-dimensional evaluation and ranking of the inverter input power, including device efficiency index, load factor, and run time analysis, the inverter power regulation priority is no longer based on static configuration, but is dynamically adjusted based on the operating performance and health status of each device, significantly improving the rationality and targeted nature of the regulation. Furthermore, the system implements an adaptive iterative algorithm for power allocation that integrates edge computing and cloud-based collaborative processing. Compared to traditional single-point computing models, this algorithm offers stronger real-time responsiveness and algorithm optimization capabilities, dynamically balancing the relationship between the total power control target and the power constraints of each inverter, reducing the number of invalid iterations and improving convergence efficiency and computational stability. Furthermore, by encoding the target power value into standardized control instructions and distributing them to each inverter using a time-sharing and partitioning strategy, network congestion and communication conflicts that may be caused by batch transmission are avoided. The controllability and distribution efficiency of the control instructions are also improved, enabling the system to maintain high reliability and synchronization when executing large-scale instruction scheduling. More importantly, this method introduces blockchain technology to securely store the control logic in the form of smart contracts. Each version of the control logic is cryptographically signed and timestamped, ensuring that the execution process of the control instructions is traceable and tamper-proof. This effectively addresses the issues of control logic being easily modified and execution results being difficult to verify in existing systems, significantly enhancing the system's reliability, security, and management transparency. Finally, by analyzing the power regulation response data and establishing an encrypted communication link to enable two-way data exchange between the power plant and the grid dispatch system, the system can meet the real-time dispatch requirements of the grid while providing feedback on the execution results. This truly implements a closed-loop response structure for dispatch control, improving the overall automation level, response speed, and grid adaptability of the photovoltaic power plant.

[0113] above Figure 2 The photovoltaic power automatic regulation system in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The photovoltaic power automatic regulation device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0114] Figure 3: This is a schematic structural diagram of a photovoltaic power automatic regulation device provided by an embodiment of the present invention. The photovoltaic power automatic regulation device 300 may vary significantly due to different configurations or performances, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and the storage medium 330 may be temporary storage or permanent storage. The program stored in the storage medium 330 may include one or more modules (not shown in the figure), each of which may include a series of instruction operations in the photovoltaic power automatic regulation device 300. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, and execute a series of instruction operations in the storage medium 330 on the photovoltaic power automatic regulation device 300 to implement the steps of the above-mentioned photovoltaic power automatic regulation method.

[0115] The photovoltaic power automatic regulation device 300 may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the photovoltaic power automatic regulating device shown does not constitute a limitation on the photovoltaic power automatic regulating device provided by the present invention, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0116] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the photovoltaic power automatic adjustment method.

[0117] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0118] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a photovoltaic power automatic regulation device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc., various media that can store program code.

[0119] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A photovoltaic power automatic adjustment method, characterized in that: The method comprises: Establishing a connection with all inverters in the photovoltaic power station through a distributed communication network, collecting real-time operating parameters of each inverter, and obtaining an inverter information data set; Performing a multi-dimensional evaluation and sorting of the input power of each inverter according to the inverter information data set to obtain an inverter power priority sequence, wherein the multi-dimensional evaluation includes equipment efficiency index, load rate, and operating time analysis; Executing an adaptive iterative algorithm based on the inverter power priority sequence to perform power allocation calculation to obtain a target power value for each inverter, wherein the adaptive iterative algorithm integrates edge computing and cloud collaborative processing; Encoding the target power value into a standardized control instruction, and issuing the control instruction to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data; Build a control logic security storage system through blockchain technology, convert the control logic into smart contract form and store it, thus obtaining a tamper-proof control strategy library; An encrypted communication link is established based on the power regulation response data to achieve two-way data exchange between the power station and the power grid dispatching system, obtain a closed-loop control system, and transmit the target power value to the on-site execution terminal.

2. The photovoltaic power automatic adjustment method according to claim 1, characterized in that: The method of establishing a connection with all inverters in the photovoltaic power station through the distributed communication network, collecting real-time operating parameters of each inverter, and obtaining an inverter information data set includes: Establishing a multi-level communication topology structure, performing zoning management on the inverters in the photovoltaic power station according to geographical locations, and obtaining inverter regional groups; Assigning differentiated communication frequencies to the inverter area groups, executing a collection strategy combining polling and event triggering, and obtaining an original parameter collection stream; The original parameter acquisition stream is subjected to data format standardization processing, outliers are removed, and timestamps are aligned to obtain a preprocessed data set; The preprocessed data set is fed into a data compression algorithm for dimensionality reduction processing to reduce redundant information and retain key features, thereby obtaining a compressed data stream; Applying a feature extraction algorithm to the compressed data stream to calculate the input power, conversion efficiency, and health status index of each inverter to obtain a feature parameter matrix; The characteristic parameter matrix is ​​stored in a distributed database according to a time series, and a parameter indexing mechanism is established to form the inverter information data set.

3. The photovoltaic power automatic adjustment method according to claim 1, characterized in that: The input power of each inverter is evaluated and sorted in a multi-dimensional manner according to the inverter information data set to obtain an inverter power priority sequence, wherein the multi-dimensional evaluation includes equipment efficiency index, load rate and operating time analysis, including: Extracting current power data, historical operation records, and environmental parameters of each inverter from the inverter information data set to construct an evaluation basic data table; Calculating the device efficiency index for each inverter in the evaluation basic data table, normalizing the output power to input power ratio in combination with the temperature correction coefficient to obtain an efficiency score; Calculating the current load rate of each inverter based on the inverter information data set, performing piecewise function mapping on the ratio of real-time power to rated power to obtain a load score; Performing a time series analysis on the historical operating data in the inverter information data set, calculating a weighted sum of the cumulative operating time and the failure rate, and obtaining a life balance coefficient; Determine the weight coefficient of each scoring indicator according to a preset control mode selector, and perform weighted fusion on the efficiency score, load score and life balance coefficient to obtain a comprehensive score; An improved quick sorting algorithm is executed on the comprehensive scores to determine ascending or descending order according to the current regulation direction to form the inverter power priority sequence.

4. The photovoltaic power automatic adjustment method according to claim 1, characterized in that: The adaptive iterative algorithm is executed based on the inverter power priority sequence to perform power allocation calculation to obtain the target power value of each inverter. The adaptive iterative algorithm integrates edge computing and cloud collaborative processing, including: Receive grid dispatch instructions or preset control targets, determine the total target power value of the photovoltaic power station, and calculate the difference between it and the current total power to obtain the power regulation amount; Combining the power adjustment amount with the inverter power priority sequence, executing an initial power allocation strategy, allocating an initial target power value to each inverter, and obtaining an initial allocation plan; Deploy a constraint checking algorithm on the edge computing unit to verify the upper and lower limit constraints and the rate of change constraint of the inverter power for the initial allocation scheme, mark the allocation values ​​that violate the constraints, and obtain the constraint verification result; Based on the constraint verification results, an adaptive step-size gradient descent algorithm is started to iteratively optimize power allocation through a cloud-based high-performance computing cluster to obtain an optimized allocation solution; Inputting the optimized allocation scheme into a response prediction model, wherein the response prediction model predicts the actual response behavior of each inverter based on historical response data to obtain a response prediction result; The optimized allocation scheme is fine-tuned according to the response prediction result to ensure the total power control accuracy and load balance of each inverter, and finally form the target power value.

5. The photovoltaic power automatic adjustment method according to claim 1, characterized in that: The target power value is encoded into a standardized control instruction, and the control instruction is sent to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data, including: Performing protocol adaptation processing on the target power value, converting it into a corresponding instruction format according to the communication protocol type of each inverter, and obtaining a protocol-compatible instruction set; Adding a timestamp, a checksum, and instruction validity period information to the protocol-compatible instruction set to form a complete control instruction packet, and calculating a transmission priority to obtain an instruction sending queue; Based on the network topology and load conditions, the instruction sending queue is batched according to geographical partitions and time windows to obtain a time-sharing and partitioned delivery plan; Execute the time-sharing and partition-based distribution plan, transmit the control command to the target inverter through a secure communication channel, and request a return receipt confirmation message to obtain the command transmission status; Deploy distributed monitoring points to collect data on the power changes of each inverter after receiving instructions, including response time, change curve, and stable value, to form a real-time response flow; Data aggregation and anomaly detection are performed on the real-time response stream, the deviation between the actual power and the target power is calculated, and a power regulation response data report is generated.

6. The photovoltaic power automatic adjustment method according to claim 1, characterized in that: The control logic security storage system is constructed by using blockchain technology, and the control logic is converted into a smart contract form for storage to obtain a tamper-proof control policy library, including: The control logic is divided into data acquisition, sorting calculation, power distribution, command issuance and feedback processing according to function, forming a modular control structure; Defining standardized interface specifications for each functional module in the modular control structure, establishing data flow relationships between modules, and obtaining a control logic flow graph; Converting the control logic flow graph into a visual expression based on function blocks and mapping it into an intermediate layer language instruction set to obtain an executable control code; Encrypting and signing the executable control code, combining it with the timestamp and operator identity information, and packaging it into a smart contract transaction unit to obtain a blockchain transaction package; Submitting the blockchain transaction package to the distributed consensus network for verification, confirming the validity of the transaction through the proof-of-stake mechanism, and obtaining a confirmed control logic block; The confirmed control logic block is linked to the main chain and synchronized between nodes to form an anti-tampering control policy library with version traceability and access permission control.

7. The photovoltaic power automatic adjustment method according to claim 1, characterized in that: The method of establishing an encrypted communication link based on the power regulation response data, realizing bidirectional data exchange between the power station and the power grid dispatching system, obtaining a closed-loop control system, and transmitting the target power value to the on-site execution terminal includes: Marking the power regulation response data as sensitive information, applying a multi-level encryption strategy to perform differential encryption processing on data of different security levels to obtain a secure data packet; Establish an encrypted communication tunnel based on TLS / SSL, adopt a two-way authentication mechanism to ensure the legitimacy of the identities of both ends of the communication, and form a secure communication link; Regularly sending power plant status reports, including current total power, adjustable capacity, and response capability assessment, to the power grid dispatch system via the secure communication link to obtain power plant status feedback; Receiving dispatch instructions issued by the power grid dispatch system through the secure communication link, performing legality verification and executability analysis on the instructions, and obtaining dispatch requirements; Converting the scheduling requirements into internal control parameters of the photovoltaic power station, coordinating and integrating them with the target power value to form a final execution instruction; The final execution instruction is transmitted to the IO measurement point and the execution terminal through the industrial field bus, and a data closed-loop verification mechanism is established to form a closed-loop control system.

8. A photovoltaic power automatic regulation system, characterized in that: For implementing the photovoltaic power automatic regulation method according to any one of claims 1 to 7, the photovoltaic power automatic regulation system comprises: A connection module is used to establish a connection with all inverters in the photovoltaic power station through a distributed communication network, collect real-time operating parameters of each inverter, and obtain an inverter information data set; a sorting module, configured to perform a multi-dimensional evaluation and sorting of the input power of each inverter according to the inverter information data set to obtain an inverter power priority sequence, wherein the multi-dimensional evaluation includes equipment efficiency index, load rate and operating time analysis; An allocation module, configured to execute an adaptive iterative algorithm based on the inverter power priority sequence to perform power allocation calculations to obtain a target power value for each inverter, wherein the adaptive iterative algorithm integrates edge computing and cloud collaborative processing; a sending module, configured to encode the target power value into a standardized control instruction, and send the control instruction to each inverter through a time-sharing and partitioning strategy to obtain power regulation response data; A conversion module is used to build a control logic security storage system using blockchain technology, convert the control logic into a smart contract format for storage, and obtain a tamper-proof control policy library; The encryption module is used to establish an encrypted communication link based on the power regulation response data, realize two-way data exchange between the power station and the power grid dispatching system, obtain a closed-loop control system, and transmit the target power value to the on-site execution terminal.

9. A photovoltaic power automatic adjustment device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the photovoltaic power automatic adjustment method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the processor is enabled to perform the photovoltaic power automatic adjustment method according to any one of claims 1 to 7.

Images