Zero-carbon factory intelligent operation and maintenance central control system and electric energy quality optimization method and device

Abstract

The invention relates to the technical field of photovoltaic power generation, and discloses a zero-carbon factory intelligent operation and maintenance central control system and an electric energy quality optimization method and device. According to the method, a central control platform is arranged in a preset zero-carbon factory and comprises a photovoltaic monitoring module and an electric energy quality analysis module; constructing a factory energy digital twinborn model according to the characteristics of the photovoltaic module, the parameters of the load equipment and the characteristics of the power grid access point; acquiring real-time meteorological data and historical power generation data corresponding to the zero-carbon factory; inputting the power generation data, the electric energy quality parameters, the real-time meteorological data and the historical power generation data into a factory energy digital twinborn model to generate a future power generation power curve; matching an optimal treatment scheme in a preset fault mode self-learning database according to the power generation data of the zero-carbon factory and a future power generation power curve; and adjusting the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the optimal treatment scheme, and completing the power quality optimization of the zero-carbon factory. And the problem of power quality control of grid connection of the distributed photovoltaic system is effectively solved.

Description

Technical Field

[0001] The present application relates to the field of photovoltaic power generation technology, and in particular to a zero-carbon factory intelligent operation and maintenance central control system and a power quality optimization method and device. Background Art

[0002] Driven by the "dual carbon" goal, zero-carbon factories generally adopt distributed photovoltaic power generation systems. However, the intermittent photovoltaic output and nonlinear load operation lead to excessive total harmonic distortion (THD) of the power grid and increased three-phase imbalance. Traditional methods have the following defects:

[0003] 1. Disconnection between prediction and control: The existing photovoltaic monitoring system only collects power generation data, and does not dynamically associate meteorological prediction with power quality parameters, resulting in the configuration of harmonic control equipment (such as APF) lagging behind actual needs;

[0004] 2. Insufficient model accuracy: The factory energy system is constructed by using a single physical model, but the attenuation characteristics of photovoltaic modules and grid impedance parameters are not integrated, resulting in a power prediction error of more than 15% when the load changes suddenly;

[0005] 3. Rigid governance scheme: The fixed threshold trigger compensation strategy does not consider the impact of PV output fluctuations on the harmonic characteristics of the grid, which can easily lead to over-compensation or under-compensation;

[0006] 4. Insufficient data utilization: Existing methods do not use machine learning to match historical fault patterns with real-time power generation curves, resulting in a three-phase unbalanced regulation response time of more than 200ms.

[0007] Therefore, a method is urgently needed to solve at least one of the above problems. Summary of the invention

[0008] The present application provides a zero-carbon factory intelligent operation and maintenance central control system and a power quality optimization method and device, aiming to solve the problems of existing zero-carbon factories generally using distributed photovoltaic power generation systems, such as disconnection between prediction and management, insufficient model accuracy, rigid management solutions and insufficient data utilization.

[0009] In a first aspect, the present application provides a method for optimizing power quality of a zero-carbon factory intelligent operation and maintenance, comprising:

[0010] A central control platform is set up in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the roof of the zero-carbon factory area, which is used to collect power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory, which is used to collect power quality parameters;

[0011] Construct a factory energy digital twin model based on the characteristics of the photovoltaic modules, the parameters of the load equipment, and the characteristics of the grid connection point corresponding to the zero-carbon factory;

[0012] Obtain the real-time meteorological data and historical power generation data corresponding to the zero-carbon factory;

[0013] Input the power generation data, power quality parameters, real-time meteorological data, and historical power generation data into the factory energy digital twin model to generate a future power generation power curve;

[0014] When the total harmonic distortion rate detected by the power quality analysis module exceeds the preset threshold, match the best governance plan in the preset fault mode self-learning database according to the power generation data and future power generation power curve of the zero-carbon factory; adjust the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the best governance plan to complete the optimization of the power quality of the zero-carbon factory; the best governance plan includes the switching strategy of the compensation capacitor bank of the zero-carbon factory.

[0015] In some embodiments, the constructing a factory energy digital twin model based on the characteristics of the photovoltaic modules, the parameters of the load equipment, and the characteristics of the grid connection point corresponding to the zero-carbon factory includes: constructing a three-dimensional simulation model of the photovoltaic power generation unit of the zero-carbon factory according to the characteristics of the photovoltaic modules; constructing a load dynamic model according to the parameters of the load equipment; constructing a grid equivalent model according to the characteristics of the grid connection point; based on the hybrid modeling method of the mechanism model and the data-driven model, coupling the three-dimensional simulation model of the photovoltaic power generation unit, the load dynamic model, and the grid equivalent model to form a digital twin model with real-time simulation capabilities.

[0016] Exemplarily, the constructing a three-dimensional simulation model of the photovoltaic power generation unit of the zero-carbon factory according to the characteristics of the photovoltaic modules; constructing a load dynamic model according to the parameters of the load equipment; constructing a grid equivalent model according to the characteristics of the grid connection point includes: analyzing the characteristics of the photovoltaic modules to obtain the temperature coefficient and photoelectric conversion efficiency curve of the photovoltaic modules for constructing the three-dimensional simulation model of the photovoltaic power generation unit; analyzing the parameters of the load equipment to obtain the power consumption characteristics curve and harmonic spectrum distribution of the load equipment of the zero-carbon factory for constructing the load dynamic model; obtaining the short-circuit capacity and system impedance ratio of the grid connection point according to the characteristics of the grid connection point for constructing the grid equivalent model.

[0017] In some embodiments, inputting the power generation data, power quality parameters, real-time meteorological data, and historical power generation data into the factory energy digital twin model to generate a future power generation power curve includes: performing normalization preprocessing on the power generation data, power quality parameters, real-time meteorological data, and historical power generation data to establish three-dimensional feature space information including irradiance, temperature, and power; constructing a time series prediction model corresponding to the factory energy digital twin model using a long short-term memory neural network; performing rolling prediction by the time series prediction model through a sliding time window mechanism; and inputting the three-dimensional feature space information into the time series prediction model to output the future power generation power curve.

[0018] Exemplarily, after inputting the three-dimensional feature space information into the time series prediction model to output the future power generation power curve, it further includes: parsing the real-time meteorological data to obtain a cloud movement vector; and correcting the photovoltaic output prediction value of the future power generation power curve based on the cloud movement vector according to a particle filter algorithm.

[0019] In some embodiments, matching the best governance plan in a preset fault mode self-learning database according to the power generation data of the zero-carbon factory and the future power generation power curve includes: obtaining a harmonic spectrum feature vector and transient process recording data according to the power generation data and the future power generation power curve; calculating the waveform similarity between the harmonic spectrum feature vector and the transient process recording data and the fault mode self-learning database using a dynamic time warping algorithm to obtain multiple candidate governance plans in the fault mode self-learning database; establishing a load demand response constraint condition according to the predicted slope change rate of the future power generation power curve; and performing multi-objective optimization on the candidate governance plans based on the load demand response constraint condition according to a genetic algorithm to determine the best governance plan among the multiple candidate governance plans.

[0020] In some embodiments, adjusting the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the best governance plan to complete the power quality optimization of the zero-carbon factory further includes: controlling a static var generator of the power supply network of the zero-carbon factory for dynamic compensation; adjusting the grid-connected power factor of the energy storage system of the power supply network of the zero-carbon factory to compensate for the remaining unbalance amount of a preset SVG device; and during the compensation process, real-time monitoring the change of the neutral line current of the power supply network of the zero-carbon factory to dynamically optimize the switching strategy of the compensation capacitor bank.

[0021] In some embodiments, after completing the power quality optimization of the zero-carbon factory, it further includes: generating encrypted electronic vouchers with timestamps for the photovoltaic power generation amount, reactive power compensation amount, and equivalent carbon emission reduction amount data of the zero-carbon factory; and the electronic vouchers conform to the verification standards of the carbon trading market.

[0022] Second aspect, the present application provides a zero-carbon factory intelligent operation and maintenance central control system, which includes:

[0023] A central control platform; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the ceiling of the zero-carbon factory plant area for collecting power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory for collecting power quality parameters.

[0024] A control device, including a memory and a processor; the memory is used to store computer programs; the processor is used to execute the computer programs and implement the methods provided in any embodiment of the present application when executing the computer programs.

[0025] Third aspect, the present application provides a zero-carbon factory intelligent operation and maintenance power quality optimization device, including:

[0026] A platform setting unit for setting a central control platform in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the ceiling of the zero-carbon factory plant area for collecting power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory for collecting power quality parameters.

[0027] A model construction unit for constructing a factory energy digital twin model according to the photovoltaic module characteristics, load device parameters and grid access point characteristics corresponding to the zero-carbon factory.

[0028] A data acquisition unit for acquiring real-time meteorological data and historical power generation data corresponding to the zero-carbon factory.

[0029] A curve generation unit for inputting the power generation data, power quality parameters, real-time meteorological data and historical power generation data into the factory energy digital twin model to generate a future power generation power curve.

[0030] An optimization completion unit for, when the total harmonic distortion rate detected by the power quality analysis module exceeds a preset threshold, matching the best governance plan in a preset fault mode self-learning database according to the power generation data of the zero-carbon factory and the future power generation power curve; adjusting the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the best governance plan to complete the power quality optimization of the zero-carbon factory; the best governance plan includes a switching strategy for the compensation capacitor bank of the zero-carbon factory.

[0031] This application provides a zero-carbon factory intelligent operation and maintenance central control system, an electric power quality optimization method and device, constructs a central control platform integrating photovoltaic monitoring and electric power quality analysis, collects power generation data in real time through a photovoltaic array sensor group, monitors the parameters of the distribution bus through a multi-parameter electric power quality detector, and realizes the closed-loop linkage between the data acquisition layer and the governance execution layer. Innovatively adopts digital twin modeling technology, integrates photovoltaic module characteristics (such as conversion efficiency, attenuation coefficient), load equipment dynamic parameters (such as non-linear load characteristics), and grid connection point characteristics (such as short-circuit capacity, impedance characteristics) to construct a three-dimensional energy model, breaking through the limitations of traditional single physical modeling. By establishing a four-dimensional data fusion mechanism (real-time power generation data + historical operation data + weather forecast data + electric power quality parameters), multi-variable time series analysis is carried out through the digital twin model to generate a power generation power prediction curve with a minute-level accuracy, and the accuracy is improved by more than 15% compared with the traditional LSTM prediction model. Initiates a fault mode self-learning database, uses transfer learning algorithms to extract features from historical governance cases, and establishes a mapping relationship library between THD exceeding standard events and governance strategies. When it is detected that the total harmonic distortion rate (THD) exceeds the IEC 61000-3-6 standard threshold, the optimal switching strategy of the compensation capacitor bank is dynamically generated through a similarity matching algorithm. Proposes a three-phase unbalance optimization model based on genetic algorithm, synchronously optimizes the phase current distribution during reactive power compensation, and controls the unbalance within the range of ≤2% specified in GB / T 15543.

[0032] By integrating multi-dimensional information such as meteorological data (irradiance, temperature, wind speed), equipment status data (inverter efficiency, module temperature), and grid parameters (voltage fluctuation, harmonic spectrum), a 500ms-level fast simulation is realized through the digital twin model, and the response speed is increased by 3 times compared with the traditional SCADA system. Adopts a two-layer optimization strategy, the upper layer performs 24-hour rolling optimization based on model predictive control (MPC), and the lower layer realizes millisecond-level dynamic compensation through a fuzzy PID controller to solve the problem of power quality deterioration caused by the intermittency of photovoltaic output and the volatility of load.

[0033] The voltage fluctuation rate can be controlled within ±2% (better than the national standard of ±7%). The THD is reduced from the conventional 8 - 12% to below 4%. The harmonic loss is reduced by about 18 - 25% annually, and the service life of sensitive equipment is extended by more than 30%. Through the precise switching strategy of capacitor banks, the power factor is maintained in the range of 0.95 - 1.0, reducing the risk of reactive power fines. Combining photovoltaic prediction and optimized scheduling, the configuration capacity of the energy storage system is reduced by 15 - 20%, saving about 2 million yuan of initial investment per MW. The photovoltaic accommodation rate is increased to 98.5%, which is 12 percentage points higher than the traditional control method. The annual standby time of diesel generators can be reduced by more than 600 hours, equivalent to a carbon emission reduction of 1500 tons per year. The constructed equipment health assessment model can predict the deterioration trend of power quality 72 hours in advance. The operation and maintenance response time is shortened to within 10 minutes, the fault location accuracy is increased to 95%, and the loss of unplanned downtime is reduced by about 40%.

[0034] In summary, through the collaborative innovation of digital twin modeling, multi-source data fusion, and intelligent optimization algorithms, this technical solution effectively solves the power quality control problem brought about by the grid connection of distributed photovoltaic systems, provides a technical implementation path with both real-time performance and economy for the carbon neutrality goal in the industrial field, and has outstanding industrial application value.

[0035] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and do not limit this application. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of this application, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of this application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0037] Figure 1 It is a schematic block diagram of the structure of the intelligent operation and maintenance central control system for a zero-carbon factory provided by an embodiment of this application;

[0038] Figure 2 It is a schematic flowchart of the steps of the power quality optimization method for intelligent operation and maintenance of a zero-carbon factory provided by an embodiment of this application;

[0039] Figure 3 It is a schematic block diagram of the structure of the control device provided by an embodiment of this application.

[0040] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and do not limit this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0042] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0043] It should be understood that, in order to facilitate the clear description of the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, the words "first", "second", etc. are used to distinguish the same items or similar items with substantially the same functions and effects. Those skilled in the art can understand that the words "first", "second", etc. do not limit the quantity and execution order, and the words "first", "second", etc. do not necessarily limit the difference.

[0044] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0045] It should also be understood that the term “and / or” used in the specification and appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0046] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0047] Driven by the "dual carbon" goal, zero-carbon factories generally adopt distributed photovoltaic power generation systems. However, the intermittent photovoltaic output and nonlinear load operation lead to excessive total harmonic distortion (THD) of the power grid and increased three-phase imbalance. Traditional methods have the following defects:

[0048] 1. Disconnection between prediction and control: The existing photovoltaic monitoring system only collects power generation data, and does not dynamically associate meteorological prediction with power quality parameters, resulting in the configuration of harmonic control equipment (such as APF) lagging behind actual needs;

[0049] 2. Insufficient model accuracy: By adopting a single physical model to construct the factory energy system, the attenuation characteristics of photovoltaic components and the grid impedance parameters are not integrated, resulting in a power prediction error of more than 15% when the load suddenly changes;

[0050] 3. Rigid governance solution: Through a fixed-threshold trigger compensation strategy, the impact of photovoltaic output fluctuations on the grid harmonic characteristics is not considered, which is likely to cause over-compensation or under-compensation;

[0051] 4. Insufficient data utilization: Existing methods do not perform machine learning matching between historical fault modes and real-time power generation curves, resulting in a three-phase unbalance adjustment response time exceeding 200 ms.

[0052] Therefore, there is an urgent need for a method to solve at least one of the above problems.

[0053] To solve the above problems, please refer to Figure 1 This application provides a zero-carbon factory intelligent operation and maintenance central control system, which is characterized by including: a central control platform; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the roof of the zero-carbon factory plant area for collecting power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory for collecting power quality parameters; a control device, including a memory and a processor; the memory is used to store computer programs; the processor is used to execute the computer programs and implement the methods provided in any embodiment of this application when executing the computer programs.

[0054] Exemplarily, the control device is configured to execute the following method: Set up a central control platform in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the roof of the zero-carbon factory plant area for collecting power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory for collecting power quality parameters; construct a factory energy digital twin model according to the photovoltaic component characteristics, load device parameters and grid access point characteristics corresponding to the zero-carbon factory; obtain the real-time meteorological data and historical power generation data corresponding to the zero-carbon factory; input the power generation data, power quality parameters, real-time meteorological data and historical power generation data into the factory energy digital twin model to generate a future power generation power curve; when the power quality analysis module detects that the total harmonic distortion rate exceeds a preset threshold, match the best governance solution in a preset fault mode self-learning database according to the power generation data and future power generation power curve of the zero-carbon factory; adjust the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the best governance solution to complete the optimization of the power quality of the zero-carbon factory; the best governance solution includes the switching strategy of the compensation capacitor bank of the zero-carbon factory.

[0055] Specifically, the present invention provides an intelligent operation and maintenance central control system for a zero-carbon factory, which realizes the dynamic coordination of power generation prediction and power quality governance by constructing a digital twin model and an adaptive learning mechanism. The central control platform integrates a photovoltaic monitoring module and a power quality analysis module; Photovoltaic monitoring module: Deploy a photovoltaic array sensor group (including irradiance sensors, temperature sensors, current / voltage sensors) on the roof of the factory area to collect power generation data in real time (such as DC side voltage, current, power); The power quality analysis module installs a multi-parameter power quality detector (supporting the IEC 61000-4-30 standard) on the distribution bus side to collect THD, three-phase unbalance degree, harmonic spectrum and voltage fluctuation data; The control device is configured with a memory and a processor to perform digital twin modeling, dynamic prediction and governance strategy optimization; The factory energy digital twin model integrates a photovoltaic component attenuation model (simulating the aging rate based on the Arrhenius equation), load device parameters (such as non-linear load harmonic emission coefficient) and grid impedance parameters (online identified by the frequency sweep method) to construct a multi-physical field coupling model;

[0056] The fault mode self-learning database stores historical fault modes (such as APF overload events, capacitor bank switching failure records) and corresponding governance solutions, and dynamically updates the weights through the reinforcement learning algorithm.

[0057] For example, deploy a distributed sensor group in the photovoltaic array and upload the power generation data to the central control platform at a sampling period of 1 second; Real-time monitor THD and three-phase unbalance degree through the power quality detector on the distribution bus, and synchronize the data to the digital twin model; Based on the attenuation curve provided by the photovoltaic component manufacturer (such as an annual attenuation rate of 0.5%) and real-time meteorological data (wind speed, cloud cover rate), dynamically correct the photovoltaic output prediction model.

[0058] Input the real-time meteorological data (such as the predicted irradiance in the next 15 minutes) and the historical power generation curve into the digital twin model to generate a high-precision power generation power curve (error ≤ 5%); When it is detected that the THD exceeds the preset threshold (such as 5%), start the fault mode matching: extract the current harmonic spectrum characteristics (such as the 5th harmonic ratio exceeding 40%), and combine with the future power generation power curve to match the best governance solution in the self-learning database;

[0059] If it is predicted that the PV output will decrease by 30% within 10 minutes, the APF dynamic capacity increase mode is preferentially enabled instead of switching the capacitor bank to avoid overcompensation. Based on the real-time power generation curve, calculate the load deviation rate of each phase; dynamically adjust the switching strategy of the compensation capacitor bank (such as switching 2 groups of capacitors in phase C) through the particle swarm optimization algorithm, so that the three-phase unbalance degree is reduced from 8% to within 2%, and the response time is shortened to 50 ms. After each treatment is completed, record the deviation between the actual THD change rate and the predicted value, and update the weight coefficient of the digital twin model through the backpropagation algorithm; when it is detected that the efficiency of the PV module decays (such as the output power decreases by 3% year-on-year), automatically adjust the model parameters to match the current state.

[0060] By real-time associating meteorological data with power quality parameters, the response time of APF configuration is shortened by 60%; integrating the multi-source data model of PV attenuation characteristics and grid impedance, reducing the power prediction error from 15% to within 5%; based on the dynamic threshold adjustment of power generation power fluctuation prediction, reducing the overcompensation probability by 40%; matching the historical fault mode with real-time data, shortening the three-phase unbalance adjustment response time from 200 ms to 50 ms; by optimizing the switching strategy of the capacitor bank, reducing the number of switch operations by 30% and extending the service life of the compensation device.

[0061] Taking a certain zero-carbon factory of an automobile as an example: by installing a 500 kW PV array on the factory roof, configuring temperature and humidity composite sensors and string current monitoring units; deploying a power quality detector on the distribution bus side to monitor the harmonic content below the 19th order in real time; when the irradiance drops suddenly due to clouds at noon, the system enables the APF capacity expansion mode in advance according to the prediction curve, and the THD drops from 6.2% to 4.1%, avoiding the THD exceeding the standard (7.5%) caused by the response delay of the traditional scheme.

[0062] Through the deep integration of digital twin and machine learning technologies, the present invention solves the collaborative problem of PV power generation and power quality governance in zero-carbon factories, and provides a reliable solution for intelligent factories with high-proportion renewable energy access.

[0063] Please refer to Figure 2 , Figure 2 which is a schematic flow chart of the method for optimizing the power quality of intelligent operation and maintenance of a zero-carbon factory provided by an embodiment of the present application. The execution device of the method is the control device of the central control system for intelligent operation and maintenance of a zero-carbon factory provided by any embodiment of the present application.

[0064] As Figure 2 shown, the provided method includes steps S101 to S105. Among them, the control device can be a handheld terminal, a laptop computer, a wearable device or a robot, etc. It is used to implement steps S101 to S105 and their corresponding embodiments.

[0065] Step S101. Set up a central control platform in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the roof of the zero-carbon factory, which is used to collect power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory, which is used to collect power quality parameters.

[0066] Specifically, this step aims to establish a core control architecture in the zero-carbon factory, and to form a dynamic energy perception network for the entire factory by integrating the photovoltaic monitoring module and the power quality analysis module. The photovoltaic monitoring module is deployed on the photovoltaic array on the roof of the factory area. It consists of a high-precision irradiance sensor, a photovoltaic panel temperature sensor, and a current / voltage transformer. It collects the DC side power generation power, component surface temperature and ambient light intensity data of the photovoltaic module in real time. The power quality analysis module uses a multi-parameter power quality detector (integrated with a high-speed sampling chip and an FFT analysis unit) installed on the distribution bus to synchronously monitor key parameters such as the three-phase voltage / current harmonic content, phase offset, power factor and imbalance. The two modules are interconnected with the central control platform through industrial Ethernet to achieve millisecond-level synchronous transmission of data between the power generation side and the power consumption side.

[0067] The deployment of the photovoltaic monitoring module includes configuring a group of sensor nodes for every 100 square meters of photovoltaic array, including: Irradiance sensor: an amorphous silicon sensor with a spectral response range of 300-1100nm, installed vertically on the surface of the photovoltaic panel, with a sampling frequency of 1Hz. Temperature sensor: a patch PT1000 platinum resistor, embedded in the photovoltaic backplane, to monitor the operating temperature of the component. Current / voltage acquisition unit: a Hall effect sensor (accuracy ±0.5%) is connected in parallel to the string combiner box to collect the DC side output. Communication architecture: The sensor node is connected to the edge gateway through the RS-485 bus. The gateway has a built-in Modbus TCP protocol stack to aggregate data to the central control platform time series database.

[0068] The deployment of the power quality analysis module includes: Detection point layout: Install a multi-parameter power quality detector (such as the Fluke 1750 series) on the outgoing side of the low-voltage busbar in the distribution room, configure an independent sampling channel for each phase (sampling rate 256 points / cycle), and calculate THD (total harmonic distortion), interharmonic content and negative sequence components in real time. Data processing: The detector has a built-in DSP chip to perform real-time harmonic analysis (based on the IEC 61000-4-30 standard) and upload data to the central control platform through the OPC UA protocol.

[0069] By synchronously collecting the power output on the power generation side and the power quality parameters on the power consumption side, the traditional system data island is broken, providing a data basis for dynamic governance. Millisecond-level sampling and standardized algorithms ensure that the THD detection error ≤ 0.2%, and the temperature monitoring accuracy is ±0.5°C, providing reliable data for model input. Industrial Ethernet communication enables the data delay < 10ms, solving the problem of response delay of governance equipment caused by data lag in traditional systems.

[0070] Step S102. Construct a factory energy digital twin model according to the characteristics of the photovoltaic modules, the parameters of the load equipment, and the characteristics of the grid connection point corresponding to the zero-carbon factory.

[0071] Specifically, in this step, a digital twin model with multi-physical field coupling is constructed by integrating the physical characteristics of photovoltaic modules, the dynamic response characteristics of loads, and the grid impedance parameters. The model includes a photovoltaic power generation prediction sub-model, a grid harmonic propagation sub-model, and a load disturbance response sub-model, and modular modeling technology is used to realize the interactive simulation of each subsystem. The photovoltaic power generation sub-model is based on the I-V curve data provided by the photovoltaic module manufacturer (including the attenuation coefficient β = 0.5% / year), and a single-diode equivalent circuit model is constructed. The formula is:

[0072] ;

[0073] Among them, is the photocurrent, is the reverse saturation current, and the dynamic correction value is updated according to the real-time temperature and irradiance. The attenuation compensation introduces the service life factor α = 1 - β×T to compensate for the annual attenuation of the component output power (T is the number of operating years). Among them, for monocrystalline silicon components, ≈ 5 to 8 A (standard test conditions: irradiance 1000 W / m², temperature 25°C). The of polycrystalline silicon components is slightly lower by about 5%. is the internal resistance of the battery, including the electrode contact resistance and the material bulk resistance. It is calculated by the slope of the I-V curve near the open circuit point under illumination. n is the diode ideality factor, which is a correction coefficient reflecting the non-ideal characteristics of the PN junction. n = 1 for an ideal diode. is the thermal voltage, the thermodynamic voltage related to temperature. characterizes the equivalent resistance of the battery edge leakage current and surface defects.

[0074] The power grid harmonic propagation sub - model includes calculating the impedance - frequency characteristics of the power grid based on the distribution network topology map using the nodal admittance matrix method. The formula is:

[0075] ;

[0076] Among them, R, L, and C are the equivalent parameters of the line, which are obtained through on - site measurement by the frequency - sweeping method. R is the equivalent AC resistance of the power grid line, including the skin effect and proximity effect. The system resonance points are identified using the impedance - frequency curve to predict the amplification risk of specific harmonics (such as the 5th and 7th harmonics). For example, for low - voltage cables (copper cores): R = 0.1 to 0.3 Ω / km. For medium - voltage lines (aluminum stranded wires): R = 0.2 to 0.5 Ω / km.

[0077] The load disturbance response sub - model establishes a time - domain model based on the V - I characteristics for devices such as frequency converters and electric arc furnaces, and uses state - space equations to describe their harmonic emission characteristics:

[0078] ;

[0079] Among them, u is the input voltage disturbance, and y is the harmonic spectrum of the output current.

[0080] Step S103. Obtain the real - time meteorological data and historical power generation data corresponding to the zero - carbon factory.

[0081] Specifically, in this step, through multi - source data fusion technology, meteorological satellite data, local meteorological station observations, and historical power generation records are integrated to construct a meteorological - power generation correlation database covering short - term (0 - 4 hours) and medium - to long - term (4 - 72 hours). The satellite data accesses the short - wave radiation forecast data of the FY - 4 satellite of the China Meteorological Administration (spatial resolution 1 km, time resolution 15 minutes), and extracts the predicted value of GHI (global horizontal irradiance) of the grid where the factory is located. For local observations, a total - sky imager (TSI - 880) is deployed at the highest point of the factory area to monitor the cloud movement speed and direction in real - time and correct the local errors of the satellite data.

[0082] Data cleaning performs outlier removal (3σ principle) and missing value imputation (KNN algorithm) on the power generation data of the past 5 years. Feature engineering constructs a feature vector library by extracting the peak - valley difference, volatility, and Pearson correlation coefficient with temperature / irradiance of the daily power generation curve.

[0083] The multi - source meteorological data fusion reduces the irradiance prediction error from 20% to 8%. The historical data feature library supports similar - day matching and improves the generalization ability of the model under extreme weather conditions.

[0084] Step S104. Input the power generation data, power quality parameters, real - time meteorological data, and historical power generation data into the factory energy digital twin model to generate the future power generation power curve.

[0085] Input real-time data into the digital twin model, adopt a hybrid prediction method combining physical models and data-driven approaches to generate a minute-by-minute power generation curve for the next 24 hours, and quantify the uncertainty of the prediction results.

[0086] Based on the photovoltaic sub-model in step S102, calculate the theoretical power generation Pphy. Use an LSTM neural network, input historical power sequences, meteorological prediction values, and component temperatures, and output a correction factor α. The final power is Pphy * α. Generate 1000 sets of perturbation samples through Monte Carlo simulation and output the 90% confidence interval of the power prediction values. Receive the latest meteorological data every 15 minutes and trigger model recalibration to ensure that the prediction curve is dynamically adjusted according to weather changes.

[0087] The hybrid model enables the short-term prediction error within 15 minutes to be ≤ 3% and the 4-hour prediction error to be ≤ 7%. The confidence interval output can identify the risk of sudden power generation drops (such as being blocked by dark clouds) in advance and support the pre-start of governance equipment.

[0088] Step S105. When the power quality analysis module detects that the total harmonic distortion rate exceeds the preset threshold, match the best governance plan in the preset fault mode self-learning database based on the power generation data of the zero-carbon factory and the future power generation curve; adjust the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the best governance plan to complete the optimization of the power quality of the zero-carbon factory; the best governance plan includes the switching strategy of the compensation capacitor bank of the zero-carbon factory.

[0089] Specifically, when it is detected that the THD exceeds the standard, combine the real-time power generation curve and the prediction results, match the optimal governance strategy from the fault mode self-learning database, and dynamically adjust the compensation current of the APF (active power filter) and the switching scheme of the capacitor bank.

[0090] The knowledge base construction is based on historical fault records (such as THD exceeding the limit, voltage sudden rise), extract feature vectors (harmonic spectrum, power volatility), and use hierarchical clustering to generate typical fault modes. Calculate the Euclidean distance between the current working condition and each clustering center in real time, and select the mode with the closest distance as the basis for governance. According to the predicted fluctuation of photovoltaic output, adopt the model predictive control (MPC) algorithm to roll-optimize the harmonic current injection amount of the APF. The objective function is:

[0091] ;

[0092] where is the change amount of the APF output current, and λ is the weight coefficient. Solve the optimal switching combination based on the genetic algorithm, with the goal of minimizing the three-phase unbalance degree and network loss. The constraint conditions include the capacitor bank life (the switching times ≤ 5 times / hour).

[0093] In some embodiments, constructing a factory energy digital twin model according to the characteristics of photovoltaic components, load device parameters, and grid connection point characteristics corresponding to the zero-carbon factory includes: constructing a three-dimensional simulation model of the photovoltaic power generation unit of the zero-carbon factory according to the characteristics of the photovoltaic components; constructing a load dynamic model according to the load device parameters; constructing a grid equivalent model according to the characteristics of the grid connection point; and performing multi-physical field coupling on the three-dimensional simulation model of the photovoltaic power generation unit, the load dynamic model, and the grid equivalent model based on a hybrid modeling method combining a mechanism model and a data-driven model to form a digital twin model with real-time simulation capabilities.

[0094] The construction of the three-dimensional simulation model of the photovoltaic power generation unit is based on the electrical characteristics (such as open-circuit voltage and short-circuit current) and physical layout parameters (such as tilt angle and azimuth angle) of the photovoltaic components. Using three-dimensional modeling software (such as ANSYS Twin Builder) to establish the geometric model of the photovoltaic array and integrating the equivalent circuit model of the photovoltaic cell to achieve the dynamic simulation of the influence of light intensity and temperature change on the power generation power.

[0095] The construction of the load dynamic model is to analyze the nameplate parameters (such as rated power and power factor) and operation log data of the load device, establish a time-domain simulation model including the starting impact of the motor and the harmonic emission characteristics of the nonlinear load, and use the state-space equation to describe the dynamic response characteristics of the load.

[0096] The construction of the grid equivalent model is to establish a Thevenin equivalent circuit model according to the short-circuit capacity and impedance ratio of the grid connection point and embed the grid harmonic background distortion rate parameter to simulate the influence of grid voltage fluctuation on the power quality of the factory.

[0097] The multi-physical field coupling is to couple the above models through a joint simulation platform (such as COMSOL Multiphysics), use the finite element method to solve the interaction of the electromagnetic field-thermal field-mechanical field, and at the same time introduce historical operation data to calibrate the parameters of the mechanism model online to enhance the real-time simulation ability of the model.

[0098] Through multi-physical field coupling modeling, accurately reflect the interaction between photovoltaic power generation, load dynamics and the grid environment, and improve the adaptability of the digital twin model to complex working conditions;

[0099] The hybrid modeling method combines the theoretical rigor of the mechanism model and the flexibility of the data-driven model, significantly improves the simulation accuracy, and provides a reliable decision-making basis for power quality optimization.

[0100] Exemplarily, a three-dimensional simulation model of a photovoltaic power generation unit for constructing a zero-carbon factory is built according to the characteristics of the photovoltaic module; a load dynamic model is built according to the parameters of the load equipment; a grid equivalent model is built according to the characteristics of the grid connection point, including: analyzing the characteristics of the photovoltaic module to obtain the temperature coefficient and the photovoltaic conversion efficiency curve of the photovoltaic module for constructing the three-dimensional simulation model of the photovoltaic power generation unit; analyzing the parameters of the load equipment to obtain the power consumption characteristics curve and the harmonic spectrum distribution of the load equipment in the zero-carbon factory for constructing the load dynamic model; obtaining the short-circuit capacity and the system impedance ratio of the grid connection point according to the characteristics of the grid connection point for constructing the grid equivalent model.

[0101] The backplane temperature data of the components is collected in real time through the photovoltaic monitoring module, and the photovoltaic conversion efficiency curve is corrected by combining the temperature coefficient (such as -0.45% / °C) to construct a three-dimensional simulation model considering the hot spot effect. The historical power consumption data is analyzed by Fourier transform to extract the characteristic harmonic spectrum of the load equipment (such as the proportion of the 5th and 7th harmonics), and a load dynamic model including the harmonic current injection capacity is established. The three-phase voltage unbalance degree of the grid connection point is measured through the power quality analysis module, and the system impedance ratio (X / R ratio) is calculated by combining the short-circuit capacity to construct a grid equivalent model that can simulate voltage sags and harmonic resonances.

[0102] The photovoltaic model is dynamically corrected based on the temperature coefficient and irradiance to accurately predict the sudden power drop caused by shadow occlusion; the harmonic spectrum characteristics are integrated into the load model to provide a data basis for subsequent harmonic control scheme matching; the grid equivalent model reflects the actual impedance characteristics and effectively predicts the harmonic amplification risk.

[0103] Exemplarily, distributed collaborative training is performed on the energy digital twin models of multiple zero-carbon factories through the federated learning framework, specifically including: each factory locally trains the LSTM prediction network of the digital twin model, encrypts and uploads the model gradient to the central server; aggregates the global gradient and distributes it to each node to achieve model parameter update and privacy protection.

[0104] Federated learning architecture: Participants: Multiple zero-carbon factories act as clients, and the central server uses the Paillier homomorphic encryption protocol to aggregate model parameters. Each factory only shares the LSTM network weight ΔW and does not transmit the original power generation data.

[0105] The training process includes: Local training: Each factory trains the LSTM prediction model based on its own historical data (such as 100-day photovoltaic power generation sequence), calculates the gradient gi = L(W); Gradient Encryption: Add noise to the gradient using differential privacy technology (noise standard deviation σ = 0.1), and upload it after encryption by the SM9 algorithm; Global Aggregation: The server performs weighted averaging on the encrypted gradients (weights are the proportion of the data volume of each factory), and updates the global model Wglobal = W - η∑gi. Federated learning iteration is performed every 24 hours to dynamically adapt to the meteorological and load differences in different regions. Cross-factory data collaboration improves the generalization ability of the prediction model, and the prediction error in the multi-cloud region is reduced by 12%; Encryption and differential privacy technology ensure data security and meet the GDPR compliance requirements.

[0106] In some embodiments, the inputting the power generation data, power quality parameters, real-time meteorological data, and historical power generation data into the factory energy digital twin model to generate a future power generation power curve includes: performing normalization preprocessing on the power generation data, power quality parameters, real-time meteorological data, and historical power generation data, and establishing three-dimensional feature space information including irradiance, temperature, and power; constructing a time series prediction model corresponding to the factory energy digital twin model using a long short-term memory neural network; the time series prediction model performs rolling prediction through a sliding time window mechanism; inputting the three-dimensional feature space information into the time series prediction model and outputting the future power generation power curve.

[0107] Perform Min-Max normalization on the power generation data (such as DC side voltage, inverter efficiency) and meteorological data (such as irradiance, wind speed) to eliminate the dimension difference; The construction of the time series prediction model uses the LSTM neural network architecture, sets the number of input layer nodes as the feature dimension (irradiance, temperature, power), the number of hidden layer neurons as 64, and the output layer as the power generation power sequence for the next 1 hour; The rolling prediction mechanism uses a 15-minute time window step, and updates the input data after each prediction to generate a continuous 72-hour power curve.

[0108] Perform Min-Max normalization processing on the power generation data such as DC side current / voltage and inverter AC output power collected by the photovoltaic monitoring module, and the irradiance (unit: W / m²) and ambient temperature (unit: °C) data obtained by the weather station respectively, and map each parameter to the [0,1] interval. The formula is:

[0109] ;

[0110] Align the historical power generation data and real-time data according to the time stamp, and construct a three-dimensional data set containing time series features, with dimensions of time step (such as 15-minute interval), number of features (irradiance, temperature, power), and sample batch.

[0111] The construction of the time series prediction model uses a two-layer LSTM neural network. The number of nodes in the input layer is 3 (corresponding to irradiance, temperature, and power), the number of neurons in the hidden layer is set to 128, and the Dropout rate is 0.2 to prevent overfitting. The output layer is a fully connected layer that generates a predicted power generation sequence for the next 24 hours (96 time points). When training the model, the Adam optimizer is used, the loss function is the mean squared error (MSE), the initial learning rate is set to 0.001, and the early stopping method (patience = 10) is adopted to dynamically adjust the number of training epochs.

[0112] The rolling prediction mechanism sets the sliding time window length to 24 hours and the step size to 1 hour. Each time a prediction is made, the latest 1-hour real-time data (4 time points) is concatenated with the historical 23-hour data to form an input sequence, and the power value for the next 1 hour is predicted. The input window is updated cyclically to achieve continuous 72-hour rolling prediction.

[0113] Normalization processing eliminates the dimensional difference and improves the model convergence speed. The two-layer structure of the LSTM network can effectively capture long-term and short-term dependencies, especially suitable for predicting power fluctuations caused by sudden changes in irradiance under cloudy weather.

[0114] The sliding window mechanism dynamically fuses real-time data, and the prediction error is reduced by more than 15% compared with traditional static models.

[0115] Exemplarily, after inputting the three-dimensional feature space information into the time series prediction model and outputting the future power generation curve, it further includes: parsing the real-time meteorological data to obtain the cloud motion vector; based on the particle filter algorithm, correcting the predicted photovoltaic output value of the future power generation curve according to the cloud motion vector.

[0116] The cloud motion vector is extracted through meteorological radar data or a sky imager to obtain the cloud movement speed and direction, and a heat map of cloud shading distribution for the next 30 minutes is constructed. The particle filter correction uses the photovoltaic array partition as a particle swarm, adjusts the weights of each particle according to the cloud occlusion probability, performs Monte Carlo simulation on the predicted power curve, and outputs the corrected confidence interval power value.

[0117] The sky image is obtained through a sky imager, the optical flow method is used to calculate the cloud movement speed and direction, and a two-dimensional motion vector field with the center of the photovoltaic array as the origin is generated. The cloud position for the next 15 minutes is extrapolated according to the vector field, the photovoltaic array is divided into a 10×10 grid, and the probability (in the range of 0 - 1) that each grid cell is blocked by clouds is calculated.

[0118] Define each string of the photovoltaic array as a particle, and a total of 1000 particles are generated. The state of each particle includes position, occlusion probability, and output power correction coefficient; adjust the particle weights according to the cloud movement vector. For example, when the cloud moves westward, the particle weights of the western array increase; use the residual between the actual power and the predicted power as the observation value, and use Monte Carlo sampling to redistribute the particles to generate the corrected power prediction value and its confidence interval (such as 95% confidence level).

[0119] The cloud movement vector quantifies the short-term occlusion effect and solves the problem of prediction lag caused by the rapid movement of clouds in the traditional model; the particle filter combines probability prediction and physical models, reducing the power prediction error within 15 minutes from 8% to less than 3%.

[0120] In some embodiments, the matching of the best governance plan in the preset fault mode self-learning database according to the power generation data of the zero-carbon factory and the future power generation power curve includes: obtaining the harmonic spectrum feature vector and transient process recorded wave data according to the power generation data and the future power generation power curve; using the dynamic time warping algorithm to calculate the waveform similarity between the harmonic spectrum feature vector and the transient process recorded wave data and the fault mode self-learning database to obtain multiple candidate governance plans in the fault mode self-learning database; establishing a load demand response constraint condition according to the predicted slope change rate of the future power generation power curve; based on the load demand response constraint condition, performing multi-objective optimization on the candidate governance plans according to the genetic algorithm to determine the best governance plan among the multiple candidate governance plans.

[0121] The waveform similarity calculation performs dynamic time warping (DTW) on the transient recorded wave data (such as voltage swell waveform), calculates the morphological distance between it and the historical waveform in the database, and screens the candidate plans with a similarity higher than 90%; the multi-objective optimization takes the lowest governance cost, the largest reduction in harmonic distortion rate, and the longest service life of the compensation equipment as the objective function, uses the NSGA-II genetic algorithm to perform Pareto front search on the candidate plans, and finally selects the best governance plan through fuzzy decision-making.

[0122] For the transient recorded wave data (such as voltage swell waveform) collected by the power quality analysis module, extract its harmonic spectrum feature vectors H3, H5,..., H50 (unit: %); use the dynamic time warping (DTW) algorithm to calculate the morphological distance between the current harmonic spectrum and the historical waveform in the database. The expression includes:

[0123] ;

[0124] where π is the alignment path, and screen the candidate plans with a DTW distance less than the threshold (such as 0.1).

[0125] Set the objective functions: governance cost f1(x), THD reduction rate f2(x), and capacitor bank life loss f3(x); adopt the NSGA-II genetic algorithm: set the population size to 200, the crossover probability to 0.8, and the mutation probability to 0.05; perform non-dominated sorting to generate the Pareto front solution set; select the comprehensive optimal solution through the fuzzy membership function, for example, the weight allocation is f1: 0.4, f2: 0.4, f3: 0.2.

[0126] The DTW algorithm overcomes the problem of waveform time-axis stretching and deformation, and the similarity matching accuracy is improved to 92%; multi-objective optimization balances economy and governance effect, avoiding over-compensation caused by a single solution (for example, THD is reduced from 8% to 2%, and the cost is controlled within ￥5000).

[0127] In some embodiments, adjusting the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the best governance plan to complete the power quality optimization of the zero-carbon factory further includes: controlling the static var generator of the power supply network of the zero-carbon factory for dynamic compensation; adjusting the grid-connected power factor of the energy storage system of the power supply network of the zero-carbon factory to compensate for the remaining unbalance amount of the preset SVG device; wherein, during the compensation process, the change of the neutral line current of the power supply network of the zero-carbon factory is monitored in real time to dynamically optimize the switching strategy of the compensation capacitor bank.

[0128] The SVG dynamic compensation controls the IGBT bridge arm of the SVG to generate reverse harmonic current according to the harmonic current command value in the best governance plan, and compensates the 5th and 7th characteristic harmonics in real time; when the SVG capacity is insufficient, the energy storage system coordinates and adjusts the reactive power output mode of the energy storage converter to compensate for the remaining unbalance amount, and at the same time monitors the zero-sequence component of the neutral line current to dynamically adjust the switching sequence of the capacitor bank to avoid resonance.

[0129] According to the harmonic current command generated by the best governance plan (such as the 5th harmonic compensation amount = 20A), control the IGBT module of the SVG to generate reverse harmonic current and track the compensation in real time; adopt the dq rotating coordinate system decoupling control and realize the zero-static error tracking of the harmonic current through the PI regulator (proportional coefficient Kp = 0.5, integral coefficient Ki = 0.1).

[0130] When the SVG capacity is insufficient (such as the compensation demand exceeds 80% of its rated capacity), switch the energy storage converter to the reactive power priority mode and output the remaining compensation amount ΔQ = Q demand - QSVG; monitor the zero-sequence component I0 of the neutral line current in real time. If I0 > 10% IN (IN is the rated current), trigger the optimization of the capacitor bank switching strategy: use the binary search method to dynamically adjust the capacitor bank switching sequence to avoid resonance frequency falling into the harmonic frequency band (such as near the 11th harmonic).

[0131] The three-phase unbalance degree is reduced from 15% to less than 2% through the collaborative control of SVG and energy storage; the neutral line current feedback mechanism prevents the compensation device from causing resonance, and the system stability is improved by 30%.

[0132] In some embodiments, after completing the power quality optimization of the zero-carbon factory, it further includes: generating encrypted electronic vouchers with timestamps for the photovoltaic power generation, reactive power compensation, and equivalent carbon emission reduction data of the zero-carbon factory; the electronic vouchers comply with the verification standards of the carbon trading market.

[0133] The data such as photovoltaic power generation and reactive power compensation are stored distributedly through IPFS, and a unique hash value is generated; based on the smart contract, the hash value is bound to the timestamp, and the national cryptographic SM2 algorithm is used for signature to generate carbon emission reduction vouchers that comply with the VCS (Verified Carbon Standard), supporting atomic transactions in the exchange.

[0134] For example, extract the daily photovoltaic power generation EPV and reactive power compensation from the power quality analysis module to calculate the equivalent carbon emission reduction (assuming the grid carbon emission factor is 0.8); upload the data to the Hyperledger Fabric blockchain network, generate a unique data fingerprint using the SHA-256 algorithm, and perform attribute encryption through the national cryptographic SM9 algorithm to ensure that only authorized parties can decrypt. Based on the smart contract, a Verified Emission Reduction (VER) voucher that complies with the VCS standard is automatically generated, including fields: Project ID: CN-2023-ZCFACTORY-001; Timestamp: ISO 8601 format (such as 2023-07-20T08:30:00Z); Emission reduction; Digital signature: Sign the voucher hash value using the SM2 elliptic curve algorithm.

[0135] The blockchain technology ensures the immutability of data and meets the requirements of the carbon trading market for data traceability; the encrypted electronic vouchers support one-key order trading, and the carbon asset monetization cycle is shortened from 30 days to within 24 hours.

[0136] Exemplarily, when the carbon emission reduction electronic certificate is traded in the blockchain trading market, the reactive power compensation equipment of the zero-carbon factory is automatically expanded. Specifically, it includes: calculating the SVG capacity that can be purchased additionally according to the carbon trading income amount, and calling the central control platform through Chaincode to execute equipment upgrades. The contract logic designed by the smart contract: If the carbon certificate transaction amount exceeds the threshold (such as ￥10,000), an expansion instruction is triggered; after receiving the expansion instruction, the central control platform automatically sends a capacity increase order to the SVG manufacturer (such as adding a 100 kVar module); when the new equipment is connected, it is automatically registered to the power supply network control system through the IEC 61850 protocol. 80% of the carbon trading income is earmarked for equipment procurement, and 20% is deposited into the smart contract reserve pool. It realizes a positive feedback loop of "carbon asset monetization - equipment upgrade - emission reduction capacity improvement", promoting the sustainable development of zero-carbon factories; the automatic execution of the smart contract reduces manual intervention, and the equipment expansion cycle is shortened from 30 days to 7 days.

[0137] In some embodiments, the coordinated compensation strategy of SVG and the energy storage system is dynamically optimized through a deep reinforcement learning algorithm (DRL). Specifically, it includes: constructing a state space with the three-phase unbalance degree, harmonic distortion rate, and equipment loss rate, an action space with the SVG output current instruction and the energy storage power adjustment amount, and a reward function with the maximization of the comprehensive power quality index improvement; using the proximal policy optimization (PPO) algorithm to train the agent to generate optimal compensation instructions in real time. The state space includes parameters such as the real-time three-phase current unbalance degree (ϵ), total harmonic distortion rate (THD), SVG module temperature (TSVG), and capacitor bank switching times (Nswitch), which are normalized into a 10-dimensional vector. Action space: Define the d-axis / q-axis harmonic compensation current (Id, Iq) of SVG and the reactive power adjustment amount (ΔQ) of the energy storage system, a total of 3 continuous actions. Reward function:

[0138] ;

[0139] where the weight coefficients are w1 = 0.4, w2 = 0.3, w3 = 0.2, and w4 = 0.1.

[0140] The PPO algorithm is trained using the Actor-Critic network structure. The Actor network is a 3-layer fully connected network (256 - 128 - 64 nodes), and the Critic network is a 2-layer fully connected network (128 - 64 nodes); set the discount factor γ = 0.99, the Clip range ϵ = 0.2, and update the policy every 1000 steps; inject random harmonic disturbances (such as 3% - 15% THD change) into the simulation environment and train for 500,000 iterations until convergence. Embed the trained policy network into the central control platform to receive power quality parameters in real time, output compensation instructions, and control equipment actions.

[0141] Dynamically adapt to complex harmonic scenarios, with a 40% improvement in response speed compared to traditional PI control strategies; balance power quality and equipment life through a multi-objective reward function, extending the service life of capacitor banks by more than 30%.

[0142] In some embodiments, edge computing nodes are deployed on the distribution bus side of a zero-carbon factory. Abnormal power quality is detected in real time through a lightweight convolutional neural network (CNN), specifically including: performing wavelet transform on the voltage / current waveforms collected by a multi-parameter power quality detector to generate a time-frequency diagram; using the pruned MobileNetV3 model to classify abnormal types (such as voltage sags and harmonic resonances).

[0143] An edge computing device is adopted, configured with 128 CUDA cores, and the power consumption is <15W; a lightweight CNN model (model size <5MB) is deployed, supporting inference to be completed within 20ms. Perform 6-layer Db4 wavelet decomposition on the voltage signal to generate a time-frequency energy distribution diagram (size 64×64); perform channel pruning on MobileNetV3 (removing 50% of redundant convolutional kernels), and use knowledge distillation technology (the teacher model is ResNet50) to improve the accuracy. Define 7 types of abnormalities: voltage sags, swells, harmonics, flicker, interruptions, oscillations, and resonances; trigger an alarm on the central control platform when the output confidence level >90%.

[0144] Localized processing of edge nodes reduces data transmission latency, and the detection response time <50ms; while maintaining 95% accuracy, the memory occupancy of the pruned model is reduced by 70%.

[0145] The embodiments of the present application also provide a power quality optimization device for intelligent operation and maintenance of a zero-carbon factory. The power quality optimization device for intelligent operation and maintenance of a zero-carbon factory is used to execute the steps of the power quality optimization method for intelligent operation and maintenance of a zero-carbon factory shown in the above embodiments. The power quality optimization device for intelligent operation and maintenance of a zero-carbon factory can be a single server or a server cluster, or the power quality optimization device for intelligent operation and maintenance of a zero-carbon factory can be a terminal, and the terminal can be a handheld terminal, a laptop computer, a wearable device, or a robot, etc.

[0146] The power quality optimization device for intelligent operation and maintenance of a zero-carbon factory includes:

[0147] A platform setting unit for setting a central control platform in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the ceiling of the zero-carbon factory plant area for collecting power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory for collecting power quality parameters;

[0148] A model construction unit for constructing a factory energy digital twin model based on the characteristics of photovoltaic components, load device parameters, and grid connection point features corresponding to the zero-carbon factory;

[0149] A data acquisition unit for acquiring real-time meteorological data and historical power generation data corresponding to the zero-carbon factory;

[0150] A curve generation unit for inputting the power generation data, power quality parameters, real-time meteorological data, and historical power generation data into the factory energy digital twin model to generate a future power generation power curve;

[0151] An optimization completion unit for, when the total harmonic distortion rate detected by the power quality analysis module exceeds a preset threshold, matching the best governance plan in a preset fault mode self-learning database according to the power generation data of the zero-carbon factory and the future power generation power curve; adjusting the three-phase unbalance degree of the power supply network of the zero-carbon factory according to the best governance plan to complete the power quality optimization of the zero-carbon factory; the best governance plan includes the switching strategy of the compensation capacitor bank of the zero-carbon factory.

[0152] It should be noted that those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working processes of the above-described zero-carbon factory intelligent operation and maintenance power quality optimization device and each unit can refer to the corresponding processes in the embodiments of the zero-carbon factory intelligent operation and maintenance power quality optimization method described in the above embodiments, and will not be repeated here.

[0153] The above-described zero-carbon factory intelligent operation and maintenance power quality optimization method is implemented in the form of a computer program, and this computer program can run on the above device.

[0154] Please refer to Figure 3 , Figure 3 which is a schematic structural block diagram of a control device provided by an embodiment of the present application. The control device includes a processor, a memory, and a network interface connected through a device bus. Among them, the memory may include a storage medium and an internal memory.

[0155] The storage medium can store an operating device and a computer program. This computer program includes program instructions, and when these program instructions are executed, the processor can execute any embodiment of the zero-carbon factory intelligent operation and maintenance power quality optimization method.

[0156] The processor is used to provide computing and control capabilities to support the operation of the entire control device.

[0157] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When this computer program is executed by the processor, the processor can execute any zero-carbon factory intelligent operation and maintenance central control system method.

[0158] The network interface is used for network communication, such as sending the assigned tasks, etc. Those skilled in the art can understand that Figure 3 the structure shown in Figure 3 is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the terminals to which the solution of this application is applied. The specific control device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0159] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0160] Among them, in one embodiment, the processor is used to run the computer program stored in the memory to implement the following steps:

[0161] Set up a central control platform in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the ceiling of the zero-carbon factory plant area for collecting power generation data; the power quality analysis module includes a multi-parameter power quality detector on the distribution bus of the zero-carbon factory for collecting power quality parameters;

[0162] Construct a factory energy digital twin model according to the characteristics of photovoltaic components, load device parameters and grid connection point characteristics corresponding to the zero-carbon factory;

[0163] Obtain the real-time meteorological data and historical power generation data corresponding to the zero-carbon factory;

[0164] Input the power generation data, power quality parameters, real-time meteorological data and historical power generation data into the factory energy digital twin model to generate a future power generation power curve;

[0165] When the power quality analysis module detects that the total harmonic distortion rate exceeds the preset threshold, the best governance solution is matched in the self-learning database of preset fault modes according to the power generation data of the zero-carbon factory and the future power generation power curve; the three-phase unbalance degree of the power supply network of the zero-carbon factory is adjusted according to the best governance solution to complete the optimization of the power quality of the zero-carbon factory; the best governance solution includes the switching strategy of the compensation capacitor bank of the zero-carbon factory.

[0166] It should be noted that those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working process of the above-described processor can refer to the corresponding process in the method embodiments described in the above various embodiments, and will not be repeated here.

[0167] An embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and the computer program includes program instructions. The processor executes the program instructions to implement the steps of the zero-carbon factory intelligent operation and maintenance power quality optimization method provided in the above various embodiments of the present application.

[0168] Among them, the computer-readable storage medium may be an internal storage unit of the control device described in the foregoing embodiment, such as the hard disk or memory of the control device. The computer-readable storage medium may also be an external storage device of the control device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the control device.

[0169] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or substitutions within the technical scope disclosed by the present application, and these modifications or substitutions should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for optimizing power quality for intelligent operation and maintenance of a zero-carbon factory, characterized in that: include: A central control platform is set up in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; The photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the roof of the zero-carbon factory area, which is used to collect power generation data; the power quality analysis module includes a multi-parameter power quality detector on the zero-carbon factory distribution bus, which is used to collect power quality parameters; Building a factory energy digital twin model according to the photovoltaic module characteristics, load equipment parameters and grid access point characteristics corresponding to the zero-carbon factory; Obtaining real-time meteorological data and historical power generation data corresponding to the zero-carbon plant; Inputting the power generation data, power quality parameters, real-time meteorological data and historical power generation data into the plant energy digital twin model to generate a future power generation curve; When the power quality analysis module detects that the total harmonic distortion rate exceeds a preset threshold, the optimal control plan is matched in the preset fault mode self-learning database according to the power generation data of the zero-carbon factory and the future power generation power curve; the three-phase imbalance of the power supply network of the zero-carbon factory is adjusted according to the optimal control plan to complete the power quality optimization of the zero-carbon factory; the optimal control plan includes the compensation capacitor group switching strategy of the zero-carbon factory.

2. The method according to claim 1, characterized in that The factory energy digital twin model is constructed according to the photovoltaic component characteristics, load equipment parameters and grid access point characteristics corresponding to the zero-carbon factory, including: Constructing a three-dimensional simulation model of a photovoltaic power generation unit of a zero-carbon factory according to the characteristics of the photovoltaic module; Constructing a load dynamic model according to the load equipment parameters; Constructing a power grid equivalent model according to the characteristics of the power grid access point; Based on a hybrid modeling method of a mechanism model and a data-driven model, the three-dimensional simulation model of the photovoltaic power generation unit, the load dynamic model and the grid equivalent model are coupled in multiple physical fields to form a digital twin model with real-time simulation capabilities.

3. The method according to claim 2, characterized in that The three-dimensional simulation model of the photovoltaic power generation unit of the zero-carbon factory is constructed according to the characteristics of the photovoltaic components; the load dynamic model is constructed according to the load equipment parameters; and the grid equivalent model is constructed according to the characteristics of the grid access point, including: Analyze the characteristics of the photovoltaic module, obtain the temperature coefficient and photoelectric conversion efficiency curve of the photovoltaic module, and use them to build a three-dimensional simulation model of the photovoltaic power generation unit; Analyze the load equipment parameters to obtain the load equipment power consumption characteristic curve and harmonic spectrum distribution of the zero-carbon factory, which are used to construct the load dynamic model; The short-circuit capacity and system impedance ratio of the grid access point are obtained according to the characteristics of the grid access point, and are used to construct the grid equivalent model.

4. The method according to claim 1, characterized in that: The power generation data, power quality parameters, real-time meteorological data and historical power generation data are input into the plant energy digital twin model to generate a future power generation curve, including: Normalizing and preprocessing the power generation data, power quality parameters, real-time meteorological data, and historical power generation data to establish three-dimensional feature space information including irradiance, temperature, and power; A time series prediction model corresponding to the plant energy digital twin model is constructed by using a long short-term memory neural network; the time series prediction model performs rolling prediction through a sliding time window mechanism; The three-dimensional feature space information is input into the time series prediction model, and the future power generation curve is output.

5. The method according to claim 4, characterized in that After inputting the three-dimensional feature space information into the time series prediction model and outputting the future power generation curve, the method further includes: Analyzing the real-time meteorological data to obtain cloud motion vectors; The photovoltaic output prediction value of the future power generation curve is corrected according to the cloud motion vector based on a particle filtering algorithm.

6. The method according to claim 1, characterized in that The method of matching the best governance solution in the preset fault mode self-learning database according to the power generation data of the zero-carbon plant and the future power generation curve includes: Acquire harmonic spectrum feature vectors and transient process recording data according to the power generation data and future power generation curve; A dynamic time warping algorithm is used to calculate the harmonic spectrum feature vector and the waveform similarity between the transient process recording data and the fault mode self-learning database, so as to obtain multiple candidate treatment solutions in the fault mode self-learning database; Establishing load demand response constraint conditions according to the predicted slope change rate of the future power generation curve; Based on the load demand response constraint condition, the candidate governance schemes are optimized with multiple objectives according to a genetic algorithm to determine the best governance scheme among the multiple candidate governance schemes.

7. The method according to claim 1, characterized in that The method of adjusting the three-phase imbalance of the zero-carbon factory power supply network according to the optimal governance scheme to optimize the power quality of the zero-carbon factory also includes: Controlling the static VAR generators of the zero-carbon plant power supply network for dynamic compensation; The grid-connected power factor of the energy storage system of the zero-carbon factory power supply network is adjusted to compensate for the residual unbalance of the preset SVG equipment; wherein, during the compensation process, the changes in the neutral line current of the zero-carbon factory power supply network are monitored in real time to dynamically optimize the switching strategy of the compensation capacitor group.

8. The method according to claim 1, characterized in that After the power quality optimization of the zero-carbon plant is completed, the method further includes: The photovoltaic power generation, reactive power compensation and equivalent carbon emission reduction data of the zero-carbon factory are used to generate an encrypted electronic certificate with a timestamp; the electronic certificate complies with the certification standards of the carbon trading market.

9. A zero-carbon factory intelligent operation and maintenance central control system, characterized in that: include: Central control platform; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; the photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the roof of the zero-carbon factory area, which is used to collect power generation data; the power quality analysis module includes a multi-parameter power quality detector on the zero-carbon factory distribution bus, which is used to collect power quality parameters; A control device, comprising a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and implement the method according to any one of claims 1 to 8 when executing the computer program.

10. A zero-carbon factory intelligent operation and maintenance power quality optimization device, characterized in that: include: A platform setting unit, used to set up a central control platform in a preset zero-carbon factory; the central control platform includes a photovoltaic monitoring module and a power quality analysis module; The photovoltaic monitoring module includes a photovoltaic array sensor group deployed on the roof of the zero-carbon factory area, which is used to collect power generation data; the power quality analysis module includes a multi-parameter power quality detector on the zero-carbon factory distribution bus, which is used to collect power quality parameters; A model building unit, used to build a factory energy digital twin model according to the photovoltaic component characteristics, load equipment parameters and grid access point characteristics corresponding to the zero-carbon factory; A data acquisition unit, used to acquire real-time meteorological data and historical power generation data corresponding to the zero-carbon plant; A curve generating unit, used for inputting the power generation data, power quality parameters, real-time meteorological data and historical power generation data into the plant energy digital twin model to generate a future power generation curve; The optimization completion unit is used to match the best control plan in the preset fault mode self-learning database according to the power generation data of the zero-carbon factory and the future power generation power curve when the power quality analysis module detects that the total harmonic distortion rate exceeds a preset threshold; adjust the three-phase imbalance of the power supply network of the zero-carbon factory according to the best control plan to complete the power quality optimization of the zero-carbon factory; the best control plan includes the compensation capacitor group switching strategy of the zero-carbon factory.

Images