Smart energy system based on power electronic transformer and monitoring method

Abstract

The invention belongs to the technical field of power equipment, and discloses a smart energy system based on a power electronic transformer and a monitoring method, and the method comprises the steps: installing a sensor; equipment operation data is collected in real time, and environment monitoring equipment collects environment data; the data processing module preprocesses the collected data; the abnormity monitoring module judges whether an abnormal condition exists or not; the fault detection module judges whether the system has a fault hidden danger or has a fault; the energy scheduling control module formulates an energy scheduling strategy; the electric energy quality collaborative optimization module ensures that the electric energy quality of the whole system meets the standard; and the load management unit manages different types of loads according to the hierarchical load control strategy. According to the invention, abnormal conditions and potential fault hidden dangers can be found; intelligent energy scheduling can be realized, energy utilization efficiency can be improved, electric energy quality can be cooperatively optimized, hierarchical load management can be realized, and key load power supply can be guaranteed; and a hierarchical load control strategy is adopted to carry out differential management on loads.

Description

Technical Field

[0001] This application relates to the field of power equipment technology, and more specifically, to a smart energy system and monitoring method based on power electronic transformers. Background Technology

[0002] Smart energy systems require a range of energy infrastructure and advanced information and communication technologies. This necessitates innovation in key technologies to enhance technical feasibility, while also striving to reduce costs and improve economic viability. Achieving these goals requires technological innovation, with key technologies for smart energy systems including power electronic transformers, energy routers, distributed power supply equipment, and energy storage technologies.

[0003] The prior art publication CN110535137A discloses a smart energy system based on a power electronic transformer, including a power electronic transformer, an AC / DC converter, an AC / AC converter, a DC / AC converter, a DC / DC converter, distributed power sources such as photovoltaics, electric vehicles, energy storage systems, DC loads, AC loads, and various AC generators. This smart energy system based on a power electronic transformer uses an AC / DC converter, a bidirectional DC / DC converter, and an AC / DC converter to form a power electronic transformer module. This allows the power electronic transformer to switch between four states: low DC voltage, low AC voltage, high DC voltage, and high AC voltage. Compared to traditional transformers that can only transmit voltage in a single frequency and direction, the power electronic transformer can transmit voltage bidirectionally in almost all situations and is very flexible in achieving frequency and voltage conversion. It can also effectively block the transmission of faults across the transformer.

[0004] While the existing technical solutions described above can achieve the relevant beneficial effects through their structure, they still suffer from the following drawbacks: 1. Lack of efficient data processing and analysis mechanisms, making it impossible to deeply mine the collected data, accurately identify potential faults, resulting in a high equipment failure rate and poor system stability. 2. Inefficient and unintelligent energy dispatching: Existing technologies typically employ traditional dispatching methods, relying on manual experience or simple rules for energy allocation, failing to dynamically adjust based on real-time changes in power output, load demand, and energy storage status. 3. Limited power quality assurance methods: Existing technologies lack collaborative mechanisms between devices and modules in power quality optimization, failing to guarantee power quality from multiple levels. 4. Inefficient load management strategies: Existing load management methods are mostly unified, lacking detailed classification and differentiated control of loads, failing to provide precise power supply guarantees based on the importance and power consumption characteristics of different types of loads.

[0005] In view of this, we propose a smart energy system and monitoring method based on power electronic transformers. Summary of the Invention

[0006] 1. Technical problems to be solved The purpose of this application is to provide a smart energy system and monitoring method based on power electronic transformers, which solves the technical problems mentioned in the background art. It realizes the analysis of collected data through anomaly monitoring module and fault detection module, which can promptly detect abnormal situations and potential fault hazards such as abnormal temperature rise and current change; it can intelligently dispatch energy, improve energy utilization efficiency, coordinately optimize power quality to meet high standard requirements, and implement hierarchical load management to ensure power supply to critical loads. The load management unit adopts a hierarchical load control strategy to achieve the technical effect of differentiated management of different types of loads.

[0007] 2. Technical Solution This application provides a smart energy system based on a power electronic transformer, comprising: The Power Electronic Transformer (PET) cluster adopts a combined architecture of modular multilevel power electronic transformers (MMC-PET) and isolated bidirectional full-bridge (DAB) PETs. ​​The MMC-PET is deployed on the high-voltage side, converting 10kV-35kV AC mains input to medium-voltage DC (±375V) output. The DAB-PET is used for medium- and low-voltage DC (±375V-48V) conversion, supporting bidirectional energy flow and meeting the needs of distributed power grid connection and load power supply. Each PET module is equipped with an intelligent monitoring unit to collect real-time data such as voltage, current, and temperature, and achieves fault diagnosis and autonomous protection through edge computing nodes. DC converter module: Constructs a hierarchical DC converter network, including boost, buck, and buck-boost converters. Each converter integrates an adaptive control chip to dynamically adjust output parameters according to load changes, ensuring power supply stability. Power modules include main power systems, distributed power sources, and energy storage power sources; Main power supply system: connected to the city power grid, with electrical isolation between the power grid and the internal DC system achieved through a power electronic transformer, and equipped with power quality regulation devices to suppress the impact of grid harmonics and voltage fluctuations on the system. Distributed power generation: Supports the connection of photovoltaic arrays and wind turbine generators. The photovoltaic system uses a maximum power point tracking (MPPT) controller and a DC / DC converter to boost the photovoltaic DC power before connecting it to the medium-voltage DC bus; the wind power generation achieves AC-DC conversion through a converter and works in conjunction with PET for efficient grid connection of new energy sources.

[0008] Energy storage power supply: Deploy lithium battery energy storage systems and supercapacitors. Lithium batteries are used for long-term energy storage, while supercapacitors respond quickly to power fluctuations. A bidirectional DC / DC converter connects the energy storage device to the DC bus, allowing charging during off-peak hours and discharging during peak hours, thus smoothing out power fluctuations and improving system stability. Data processing module: preprocesses various collected power equipment and environmental data, including data cleaning and standardization; and synchronizes multi-source data through timestamps and other methods.

[0009] Load Management Unit: Employing a hierarchical load control strategy, loads are categorized into critical loads (such as hospital life support systems), important loads (data center servers), and adjustable loads (industrial equipment, electric vehicle charging stations). Critical loads are dually protected by uninterruptible power supplies (UPS) and direct PET power lines; important loads achieve dynamic power allocation through intelligent distribution boxes; adjustable loads are connected to a demand response platform, which adjusts power consumption via DC converters based on real-time grid status and electricity price signals to achieve peak shaving and valley filling. Energy dispatch and control module: Based on power output, load demand, and energy storage status, it formulates energy dispatch strategies using model predictive control algorithms. Distributed power sources are prioritized for supply, with any shortfall supplemented by the main grid; the energy storage system charges when there is a surplus of new energy sources and discharges when the system power is insufficient. Power quality co-optimization module: Optimization enables the power electronic transformer, DC converter and energy storage system to work together. The PET is responsible for macro voltage regulation and harmonic suppression, the DC converter eliminates local power supply harmonics, and the energy storage system compensates for reactive power, ensuring that the power quality of the whole system meets the IEEE 519 standard. Fault Detection Module: Performs multi-source data monitoring and intelligent fault analysis and diagnosis; collects data such as voltage, current, frequency, temperature, and vibration in real time, and uses edge computing nodes and cloud-based intelligent algorithms to analyze the collected data in real time to identify abnormal patterns. It determines whether the system has potential faults or whether a fault has already occurred.

[0010] Fault emergency handling module: When the system detects a fault, the PET quickly disconnects the faulty line, the DC converter switches to the backup power supply path, and the energy storage system ensures power supply to critical loads.

[0011] Alarm module: Sound and light alarms are installed at key equipment locations in the system. Once a fault is detected, the local sound and light alarm will be activated immediately, emitting flashing lights and a high-decibel alarm sound to alert on-site personnel.

[0012] Environmental monitoring equipment includes temperature and humidity sensors, smoke sensors, and harmful gas detectors, which monitor the temperature, humidity, fire hazards, and concentration of harmful gases in the system's operating environment in real time to prevent environmental factors from affecting the normal operation of the equipment or even causing safety accidents.

[0013] Anomaly monitoring module: Monitors abnormal situations through multimodal collaborative monitoring; Data storage devices: used to store data; employing data storage devices such as hard disk arrays (RAID) and distributed storage systems (such as Ceph) for long-term data storage and management.

[0014] Control Center: Connected to the network of each module.

[0015] As an optional embodiment of the present invention, an AC / DC converter, an AC / AC converter, a DC / AC converter, or a DC / DC converter may be provided as needed.

[0016] AC / DC Converters: While power electronic transformers (PETs) provide electrical isolation between the power grid and the internal DC system, in some scenarios, additional AC / DC converters may be needed for more flexible AC-to-DC conversion to adapt to the DC power requirements of different devices, or to supplement or adjust the DC power supplied by the PET. For example, when the AC input characteristics of some distributed power sources are not well-matched with the PET, pre-processing can be performed using an AC / DC converter.

[0017] AC / AC converters are essential if a system contains AC loads with different voltage levels or frequency requirements. They enable voltage step-up / step-down conversion and frequency regulation to meet the power supply needs of various AC loads. For example, some specific industrial equipment has special requirements for AC voltage and frequency; an AC / AC converter allows the system to better power these loads.

[0018] DC / AC converters: In the energy storage power supply section, lithium batteries and supercapacitors store direct current (DC). When AC loads need to be powered, DC / AC converters are required to convert DC to AC. Furthermore, DC / AC converters can also function when the power electronic transformer (PET) outputs DC, but some loads require AC power, enabling DC-to-AC conversion and ensuring the versatility of the system's power supply.

[0019] As an optional embodiment of the present invention, the energy dispatch control module formulates an energy dispatch strategy based on a model predictive control algorithm according to power output, load demand, and energy storage status, including the following steps: 1. Real-time data acquisition: Real-time data acquisition, including power output data acquisition, load demand data acquisition, and energy storage status data acquisition; 2. System Status Analysis: 2.1 Data Integration and Processing: The collected power output, load demand, and energy storage status data are integrated, and abnormal data and noise interference are removed to ensure the accuracy and validity of the data. Data cleaning algorithms are used to interpolate missing data and correct abnormal fluctuations. 2.2 Energy Supply and Demand Assessment: Analyze the balance between power output and load demand to determine whether the current system is in a state of power surplus, power balance, or power shortage. Based on the status of the energy storage system, assess the adjustment capabilities of the energy storage equipment under different operating conditions. 3. Scheduling strategy formulation: 3.1 Prioritize the use of distributed power sources: When the system performs energy dispatch, it first determines whether the output power of the distributed power sources meets the load requirements. 3.2 Supplementing the main grid's power: If the output power of distributed generation cannot meet the load demand, the power shortfall is calculated, and the main grid provides supplementary power to ensure the normal operation of the load. During the process of drawing power from the main grid, the power supply quality and electricity price information of the grid are monitored in real time to optimize the power drawing strategy and reduce electricity costs. 3.3 Energy Storage System Charging and Discharging Control: When there is a surplus of new energy power generation, i.e., the power output exceeds the load demand, the energy storage system is controlled to charge, storing the excess electrical energy. When the system power is insufficient, the energy storage system is scheduled to discharge, releasing the stored energy to make up for the power gap, smooth power fluctuations, and improve the stability and reliability of the system. 4. Execution of dispatching instructions: The established energy dispatching strategy is transformed into specific control instructions, which are then sent to the power electronic transformer (PET) cluster, DC converter module, energy storage system and load management unit through the communication network. 5. Dynamic Adjustment and Optimization: Continuously monitor the system's operating status in real time, compare actual operating data with the expected goals of the scheduling strategy, and evaluate the effectiveness of the scheduling strategy. By analyzing the dynamic changes in power output, load power, and energy storage status, determine whether adjustments to the current scheduling strategy are necessary. If the system operating status deviates from the expected goals, such as a sudden drop in distributed power generation or a surge in load demand, the energy dispatch control module quickly activates the dynamic adjustment mechanism. Based on the latest system status data, it reapplies model predictive control algorithms to formulate new energy dispatch strategies, and promptly adjusts power allocation, energy storage charging and discharging, and load power regulation schemes to ensure the system always operates in a highly efficient and stable state.

[0020] This invention provides a smart energy monitoring method based on power electronic transformers, comprising the following steps: S1. Deploy temperature and current sensors at key components such as power devices in the power electronic transformer (PET) cluster; install voltage fluctuation monitoring devices at the output of the DC converter module; deploy frequency detection equipment on the distributed power inverter side of the power module; and install corresponding monitoring sensors at key nodes of the load management unit. Install environmental monitoring equipment consisting of temperature and humidity sensors, smoke sensors, and hazardous gas detectors in various areas of the system.

[0021] S2 collects real-time operational data from the intelligent monitoring unit of the power electronic transformer (PET) cluster, the adaptive control chip of the DC converter module, various devices in the power supply module, and the load management unit, including voltage, current, frequency, temperature, and vibration data. Environmental monitoring equipment collects environmental data such as temperature, humidity, and harmful gas concentrations. The collected data is transmitted to the control center via the network, while the PET cluster's data is simultaneously transmitted to edge computing nodes.

[0022] S3, the data processing module, preprocesses the various types of data collected, including data cleaning, standardization, and data synchronization. S4. The anomaly monitoring module analyzes data collected from various components such as the power electronic transformer (PET) cluster and DC converter module through multimodal collaborative analysis, extracts abnormal features such as abnormal temperature rise and sudden current change, and determines whether there is an abnormal situation.

[0023] The S5 fault detection module combines edge computing nodes and cloud-based intelligent algorithms, using deep learning models, data comparison, and trend analysis to conduct in-depth analysis of the data and determine whether the system has potential faults or has already experienced a fault. S6, the energy dispatch control module analyzes the system's energy supply and demand based on the collected power module output, load management unit requirements, and energy storage power status data, and formulates energy dispatch strategies to achieve rational energy allocation. S7, the power quality collaborative optimization module coordinates the power electronic transformer (PET) cluster, DC converter module and energy storage power supply to work together. The PET cluster is responsible for macro voltage regulation and harmonic suppression, the DC converter module eliminates local power supply harmonics, and the energy storage power supply compensates for reactive power, ensuring that the power quality of the whole system meets the IEEE 519 standard. S8. The load management unit manages different types of loads according to the hierarchical load control strategy.

[0024] S9. When the fault detection module determines that a system fault has occurred, the fault emergency handling module is immediately activated. The power electronic transformer (PET) cluster quickly disconnects the faulty line, the DC converter module switches to the backup power supply path, and the energy storage power supply prioritizes power supply to critical loads. The energy dispatch control module uses blockchain technology to achieve network-wide sharing of fault information, accelerating fault location. The alarm module is activated and issues an alarm. S10, Data storage devices are used for long-term storage and management of various types of data.

[0025] 3. Beneficial effects One or more technical solutions provided in this application have at least the following technical effects or advantages: 1. This invention analyzes the collected data through an anomaly monitoring module and a fault detection module, which can promptly detect abnormal situations and potential fault hazards such as abnormal temperature rise and sudden current changes, providing strong support for equipment maintenance and fault prevention, reducing equipment failure rate, and ensuring stable system operation. 2. Intelligent energy dispatching to improve energy efficiency: The energy dispatching and control module, based on model predictive control algorithms, prioritizes distributed power sources and rationally allocates power between the main grid and energy storage based on power output, load demand, and energy storage status data. This achieves efficient utilization of new energy sources and rational energy distribution. It charges when there is a surplus of new energy and discharges when there is a shortage, avoiding energy waste, smoothing power fluctuations, improving energy efficiency, reducing energy costs, and simultaneously reducing dependence on traditional energy sources, thus promoting sustainable energy development. 3. Collaborative optimization of power quality to meet high standards: The power quality collaborative optimization module coordinates the PET cluster, DC converter module, and energy storage power supply to work together. The PET cluster is responsible for macroscopic voltage regulation and harmonic suppression, the DC converter module eliminates local power supply harmonics, and the energy storage power supply compensates for reactive power, ensuring power quality from multiple levels. This ensures that the power quality of the entire system meets the IEEE 519 standard, improving the overall system performance and reliability. 4. Hierarchical load management to ensure power supply to critical loads: The load management unit adopts a hierarchical load control strategy to manage different types of loads differently. 5. Rapid fault emergency handling to reduce the impact of faults: When the fault detection module determines that a system fault has occurred, the fault emergency handling module is quickly activated, the PET cluster quickly disconnects the faulty line, the DC converter module switches to the backup power supply path, and the energy storage power supply prioritizes the power supply to critical loads, minimizing the impact of the fault on the system. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating a preferred embodiment of the smart energy monitoring method based on power electronic transformers disclosed in this application. Detailed Implementation

[0027] The present application will be further described in detail below with reference to the accompanying drawings.

[0028] Reference Figure 1 This application provides a smart energy system based on a power electronic transformer, comprising: The Power Electronic Transformer (PET) cluster adopts a combined architecture of modular multilevel power electronic transformers (MMC-PET) and isolated bidirectional full-bridge (DAB) PETs. ​​The MMC-PET is deployed on the high-voltage side, converting 10kV-35kV AC mains input to medium-voltage DC (±375V) output, reducing harmonics to THD <2% through carrier phase-shift modulation technology. The DAB-PET is used for medium- and low-voltage DC (±375V-48V) conversion, supporting bidirectional energy flow to meet the needs of distributed power grid connection and load power supply. Each PET module is equipped with an intelligent monitoring unit to collect real-time data such as voltage, current, and temperature, achieving fault diagnosis and autonomous protection through edge computing nodes. DC Converter Module: Constructs a hierarchical DC converter network, including boost, buck, and buck-boost converters. The first-level DC converter converts the ±375V DC output from the PET circuit to a 48V bus voltage, powering low-voltage equipment such as building lighting and communication base stations. The second-level DC converter, designed for precision electronic equipment such as servers and medical instruments, converts 48V to specific voltages such as 12V and 5V, employing soft-switching technology to reduce conversion losses to over 97%. Each converter integrates an adaptive control chip, dynamically adjusting output parameters according to load changes to ensure power supply stability. Power modules include main power systems, distributed power sources, and energy storage power sources; Main power supply system: connected to the city power grid, with electrical isolation between the power grid and the internal DC system achieved through a power electronic transformer, and equipped with power quality regulation devices to suppress the impact of grid harmonics and voltage fluctuations on the system. Distributed power generation: Supports the connection of photovoltaic arrays and wind turbine generators. The photovoltaic system uses a maximum power point tracking (MPPT) controller and a DC / DC converter to boost the photovoltaic DC power before connecting it to the medium-voltage DC bus; the wind power generation achieves AC-DC conversion through a converter and works in conjunction with PET for efficient grid connection of new energy sources.

[0029] Energy storage power supply: Deploy lithium battery energy storage systems and supercapacitors. Lithium batteries are used for long-term energy storage, while supercapacitors respond quickly to power fluctuations. A bidirectional DC / DC converter connects the energy storage device to the DC bus, allowing charging during off-peak hours and discharging during peak hours, thus smoothing out power fluctuations and improving system stability. Data processing module: Preprocesses various collected power equipment and environmental data, including data cleaning and standardization; cleans the data to remove noise, outliers and missing values; performs standardization to unify the range of different types of data; and synchronizes multi-source data through timestamps and other methods to improve data quality and prepare for in-depth analysis.

[0030] Load Management Unit: Employing a hierarchical load control strategy, loads are categorized into critical loads (such as hospital life support systems), important loads (data center servers), and adjustable loads (industrial equipment, electric vehicle charging stations). Critical loads are dually protected by uninterruptible power supplies (UPS) and direct PET power lines; important loads achieve dynamic power allocation through intelligent distribution boxes; adjustable loads are connected to a demand response platform, which adjusts power consumption via DC converters based on real-time grid status and electricity price signals to achieve peak shaving and valley filling. Energy dispatch and control module: This serves as the intelligent control hub, formulating energy dispatch strategies based on model predictive control algorithms according to power output, load demand, and energy storage status. It prioritizes distributed power sources, supplementing any shortfall with the main grid; the energy storage system charges when renewable energy sources are in surplus and discharges when system power is insufficient. Power quality co-optimization module: Optimization enables the power electronic transformer, DC converter and energy storage system to work together. The PET is responsible for macro voltage regulation and harmonic suppression, the DC converter eliminates local power supply harmonics, and the energy storage system compensates for reactive power, ensuring that the power quality of the whole system meets the IEEE 519 standard. Fault Detection Module: This module performs multi-source data monitoring and intelligent fault analysis and diagnosis. High-precision sensors are deployed at key nodes in the power electronic transformer (PET) cluster, DC converter module, power supply module, and load management unit to collect real-time data on voltage, current, frequency, temperature, vibration, etc. For example, temperature and current sensors are installed at the power devices of the PET, voltage fluctuation monitoring devices are set at the output of the DC converter, and frequency detection equipment is deployed on the inverter side of the distributed power supply. Intelligent analysis and diagnosis are performed using edge computing nodes and cloud-based intelligent algorithms to analyze the collected data in real time. A deep learning-based fault diagnosis model is adopted, trained on a large amount of normal and fault state data to identify abnormal patterns. Simultaneously, data comparison and trend analysis methods are used to compare real-time data with historical data and standard parameters to determine whether there are potential faults or if a fault has already occurred in the system. For example, if a sudden surge in the current of a phase of the PET exceeds a threshold, or if the output voltage of the DC converter continuously deviates from the normal range, the system automatically identifies it as a fault signal.

[0031] Emergency fault handling module: When the system detects a fault, the PET quickly disconnects the faulty line, the DC converter switches to the backup power supply path, and the energy storage system ensures power supply to critical loads. Simultaneously, the intelligent control center uses blockchain technology to achieve network-wide sharing of fault information, accelerating fault location and repair.

[0032] Alarm Module: Audible and visual alarms are installed at key equipment locations throughout the system, such as PET control cabinets, DC converter cabinets, and energy storage battery cabinets. Upon detecting a fault, the local audible and visual alarms immediately activate, emitting flashing lights and a high-decibel alarm to alert on-site personnel. A mobile terminal alarm push system is established to push fault information in real-time to maintenance personnel's mobile phones, tablets, and other devices via SMS and app notifications. Alarm information includes detailed information such as fault type, location, and severity, facilitating rapid response from maintenance personnel. A large-screen display system is installed in the control center to display fault information in real-time using visual charts, enabling management personnel to have a comprehensive understanding of system fault conditions.

[0033] Environmental monitoring equipment includes temperature and humidity sensors, smoke sensors, and harmful gas detectors, which monitor the temperature, humidity, fire hazards, and concentration of harmful gases in the system's operating environment in real time to prevent environmental factors from affecting the normal operation of the equipment or even causing safety accidents.

[0034] Anomaly monitoring module: Monitors abnormal situations through multimodal collaborative monitoring; Data storage devices: These devices store data. As the system operates, it generates a large amount of data, including equipment operation data, fault data, and energy scheduling data. Data storage devices such as RAID arrays and distributed storage systems (like Ceph) are used for long-term data storage and management, providing support for data analysis, system optimization, and fault tracing.

[0035] Control Center: Connected to the network of each module.

[0036] Furthermore, the load management unit adopts a hierarchical load control strategy, including the following steps: 1. Load classification access: including critical load access, important load access, and adjustable load access; Critical Load Connection: For critical loads such as hospital life support systems, a dual power supply line is established. First, it is connected to the input of the uninterruptible power supply (UPS), and simultaneously connected to the direct power supply line of the power electronic transformer (PET). The UPS integrates a UPS monitoring module, mainly composed of a data acquisition unit, a signal processing unit, and a communication unit. The data acquisition unit collects real-time data on the UPS battery voltage, current, remaining charge, and temperature, as well as UPS output voltage, current, frequency, and power parameters through various sensors. The signal processing unit filters and amplifies the collected raw signals, converting them into digital signals that can be recognized by the system. The communication unit is responsible for uploading the processed data to the load management system via communication protocols such as Modbus and RS485, ensuring that emergency power supply is readily available.

[0037] Critical load access: Critical loads such as data center servers are connected to the system through intelligent distribution boxes. These boxes contain high-precision power monitoring instruments that can collect data such as voltage, current, and power for each load. Adjustable load access: Adjustable loads such as industrial equipment and electric vehicle charging stations are connected to the demand response platform via a dedicated communication interface. Simultaneously, a DC converter is deployed at the front end of the adjustable load. The DC converter establishes a communication connection with the demand response platform, receiving and executing power adjustment commands. 2. Power supply guarantee for critical loads: 2.1 Normal Power Supply Mode: During normal system operation, critical loads are prioritized for power supply by the PET direct power line, and the UPS is in hot standby mode. The UPS monitoring module periodically triggers the UPS to perform self-tests, including battery charge / discharge tests and inverter function tests, to ensure sufficient battery power and normal function of components such as the inverter. Once the UPS monitoring module detects a potential fault in the UPS itself, such as abnormal battery voltage or inverter output waveform distortion, it will immediately send an early warning message to the load management system. 2.2 Abnormal Power Supply Handling: When an abnormal situation such as a sudden voltage drop or power outage occurs in the PET direct power supply line, the UPS monitoring module quickly detects the abnormal signal and switches to battery inverter power supply mode within milliseconds to maintain uninterrupted operation of critical loads. Simultaneously, the load management system sends alarm information to maintenance personnel, indicating a PET line fault. 3. Dynamic power distribution for critical loads: 3.1 Data Acquisition and Analysis: The intelligent distribution box collects real-time power consumption data from critical loads and transmits it to the load management system. The system utilizes big data analytics algorithms to analyze historical power consumption data and current operating status of the loads, predicting future power demands for each load. 3.2 Power Allocation Strategy Formulation: Based on the forecast results and the real-time power supply capacity of the power grid, the load management system formulates a dynamic power allocation strategy. If the power grid supply is sufficient, power is allocated according to load priority and demand; if the power grid supply is tight, the power of non-critical loads is limited, and priority is given to ensuring the operation of core business loads. 3.3 Power Regulation Execution: The load management system sends power allocation commands to the intelligent distribution box. The intelligent distribution box, through its internal controllable switches and adjustment modules, adjusts the current in each load branch to achieve dynamic power allocation. During the process, the actual load power is continuously monitored to ensure accurate allocation. Dynamic power allocation is performed according to the following formula: U(t) = K p (t)e(t)+K i (t)ƒ t 0e(τ)dτ+K d (t)[de(t) / (dt)]; e(t) = P i opt -P i act In the formula, U(t) represents the control quantity output by the intelligent distribution box at time t. This control quantity is used to adjust the current or voltage of each load branch, thereby realizing the dynamic distribution and adjustment of the load power to achieve the desired power value. K p P(t) represents the proportionality coefficient at time t, which is multiplied by the power error e(t) to provide a fast response to changes in the power error. i opt P is the target power of the i-th load (i.e., the expected power value allocated to the load by the load management system). i act K represents the actual power of the i-th load. e(t) reflects the degree of deviation between the current load's actual power and the target power. i (t) represents the integral coefficient at time t. e(τ) represents the power error of the integral variable at time τ, where τ is a variable K that varies in the interval [0, t]. d (t) is the differential coefficient at time t. de(t) / (dt) represents the rate of change of the power error e(t) at time t, reflecting how fast the error changes. By calculating the rate of change of the error, the differential element can be adjusted when the error just begins to show a trend of change, preventing the error from expanding further and enhancing the system's response speed and stability.

[0038] 4. Adjustable load demand response: 4.1 Grid Information Reception: The demand response platform receives real-time grid status information from the grid dispatch center, including the current grid load factor and electricity price signals. Simultaneously, it analyzes and processes this information to determine whether power regulation of adjustable loads is necessary. 4.2 Regulation command generation: When the grid load factor is too high or the electricity price is at its peak, the demand response platform generates power regulation commands according to the preset regulation strategy to determine the power value that each adjustable load needs to reduce or increase. 4.3 Command Execution and Feedback: The demand response platform sends adjustment commands to the DC converters at the front end of each adjustable load. After receiving the command, the DC converter changes the load power consumption by adjusting the output voltage or current. After execution, the DC converter feeds back the actual adjusted power data to the demand response platform. The platform evaluates the adjustment effect and feeds the results back to the power grid dispatch center.

[0039] Furthermore, the energy dispatch control module formulates an energy dispatch strategy based on model predictive control algorithms according to power output, load demand, and energy storage status, including the following steps: 1. Real-time data acquisition: Real-time data acquisition, including power output data acquisition, load demand data acquisition, and energy storage status data acquisition; Power Output Data Acquisition: The energy dispatch and control module acquires output data from various power sources in real time via the communication network. For the main power system, it collects parameters such as voltage, current, and power connected to the urban power grid; for distributed power sources, such as photovoltaic arrays, it collects information such as real-time power generation, DC voltage, and current of the photovoltaic panels; for wind turbine generators, it collects data such as wind speed, generator speed, and output power; for energy storage power sources, it acquires data such as the remaining capacity, charging and discharging current, and voltage of lithium batteries, as well as the state of charge and charging and discharging power of supercapacitors. Load demand data acquisition: Establishes real-time communication with the load management unit to receive power demand information from critical loads, important loads, and adjustable loads. This includes real-time power consumption from hospital life support systems, data center servers, etc., as well as the current operating power and potential power adjustment needs of adjustable loads such as industrial equipment and electric vehicle charging stations. Energy storage status data acquisition: Continuously monitor the operating status of the energy storage system. In addition to the basic data of lithium batteries and supercapacitors mentioned above, it also includes information such as the number of charge and discharge cycles of energy storage devices and health status assessment indicators, so as to fully understand the available capacity and performance status of the energy storage system. 2. System Status Analysis: 2.1 Data Integration and Processing: The collected power output, load demand, and energy storage status data are integrated, and abnormal data and noise interference are removed to ensure the accuracy and validity of the data. Data cleaning algorithms are used to interpolate missing data and correct abnormal fluctuations. 2.2 Energy Supply and Demand Assessment: Analyze the balance between power output and load demand to determine whether the current system is in a state of power surplus, power balance, or power shortage. Simultaneously, considering the status of the energy storage system, assess the adjustment capabilities of the energy storage equipment under different operating conditions, such as the amount of electricity it can absorb during power surplus and the energy it can release during power shortage. 3. Scheduling strategy formulation: 3.1 Prioritize Distributed Power Generation: When the system performs energy dispatch, it first determines whether the output power of distributed power sources meets the load demand. If the generating capacity of distributed power sources such as photovoltaic and wind power is sufficient, they should be allocated to various loads first to maximize the use of renewable energy and reduce dependence on the main power grid. 3.2 Supplementing the main grid's power: If the output power of distributed generation cannot meet the load demand, the power shortfall is calculated, and the main grid provides supplementary power to ensure the normal operation of the load. During the process of drawing power from the main grid, the power supply quality and electricity price information of the grid are monitored in real time to optimize the power drawing strategy and reduce electricity costs. 3.3 Energy Storage System Charging and Discharging Control: When there is a surplus of new energy power generation, i.e., the power output exceeds the load demand, the energy storage system is controlled to charge, storing the excess electrical energy. When the system power is insufficient, the energy storage system is scheduled to discharge, releasing the stored energy to make up for the power gap, smooth power fluctuations, and improve the stability and reliability of the system. 4. Dispatch Command Execution: The established energy dispatch strategy is translated into specific control commands, which are then sent via the communication network to the power electronic transformer (PET) cluster, DC converter module, energy storage system, and load management unit. For example, voltage regulation and power transmission commands are sent to the PET cluster, power allocation adjustment commands to the DC converter module, charge / discharge control commands to the energy storage system, and load power adjustment commands to the load management unit. Upon receiving the dispatch commands, each device and module immediately executes the corresponding operation and feeds back the execution results to the energy dispatch control module in real time. For instance, the PET cluster provides feedback on the actual voltage regulation and power transmission status, the energy storage system provides information on changes in charge / discharge current, voltage, and power consumption, and the load management unit provides feedback on the actual power consumption after load power adjustment. 5. Dynamic Adjustment and Optimization: Continuously monitor the system's operating status in real time, compare actual operating data with the expected goals of the scheduling strategy, and evaluate the effectiveness of the scheduling strategy. By analyzing the dynamic changes in power output, load power, and energy storage status, determine whether adjustments to the current scheduling strategy are necessary. The optimal control sequence (grid interaction power, energy storage charging and discharging power, and SOC trajectory) is obtained through dynamic adjustment and optimization according to the following formula: minΣ Np k=0 {c g (t+k)P g (t+k)+λ ba (t+k)|P bat (t+k)|+γ[SOC(t+k)-SOC ref ]}; λ ba (t+k)=λ0×e -a1×SOH(t) ;P bat (t+k)=P li-ion (t+k)+P sc (t+k); where Np is the prediction time-domain step size of Model Predictive Control (MPC); t is time; t+k is the time of the kth future time step; c g (t+k) represents the time-of-use electricity price (which varies over time), linked to the real-time electricity price of the power grid, and dynamically optimizes economic efficiency; λ ba (t+k) is the dynamic weighting coefficient for energy storage loss; λ0 is the initial loss coefficient benchmark value, calibrated according to the battery type; a1 is the health degradation sensitivity coefficient, controlling the intensity of the influence of SOH on the loss coefficient; Pbat (t+k) represents the charging and discharging power of the energy storage system at the k-th time step (t+k). A positive value indicates energy storage discharging (supplying power to the system); a negative value indicates energy storage charging (absorbing electrical energy from the system); SOH(t) is the battery health state, ranging from [0,1]; it is updated in real time, incorporating historical charge-discharge cycle counts, internal resistance changes, and other data; γ is the SOC tracking penalty coefficient, forcing the energy storage SOC to be maintained at a reference value (e.g., 60%) to avoid deep charging and discharging and extend its lifespan; SOC ref It is the optimal state of charge reference value for the energy storage system (usually set to 60%); it is dynamically adjusted according to the battery chemistry characteristics (such as lithium iron phosphate / ternary lithium); P li-ion (t+k) is the output / input power of the lithium battery pack (positive for discharging, negative for charging); P sc (t+k) is the output / input power of the supercapacitor.

[0040] If the system's operating status deviates from the expected target, such as a sudden drop in distributed power generation or a surge in load demand, the energy dispatch control module quickly activates the dynamic adjustment mechanism. Based on the latest system status data, it re-applies model predictive control algorithms to formulate new energy dispatch strategies, and promptly adjusts power distribution, energy storage charging and discharging, and load power regulation schemes to ensure that the system always operates in a highly efficient and stable state.

[0041] Furthermore, the power quality collaborative optimization module optimizes power quality, enabling the power electronic transformer, DC converter, and energy storage system to work collaboratively. The PET is responsible for macroscopic voltage regulation and harmonic suppression, the DC converter eliminates local power supply harmonics, and the energy storage system compensates for reactive power, ensuring that the power quality of the entire system meets the standards. This includes the following steps: 1. Real-time Data Acquisition and Monitoring: At key nodes in the power system, such as the input and output terminals of power electronic transformers (PETs), the connection lines of DC converters, and the access points of energy storage systems, monitoring equipment such as voltage sensors, current sensors, and harmonic analyzers is deployed to collect power quality-related data in real time, including voltage, current, harmonic content, and reactive power. A stable communication network, such as Ethernet or industrial fieldbus (e.g., Modbus, PROFIBUS), is established to transmit the collected data to the central control unit of the power quality collaborative optimization module in real time, ensuring the timeliness and accuracy of the data. 2. Power quality problem diagnosis: including data preprocessing, standard comparison analysis, and problem localization; Data preprocessing: The collected data is preprocessed by filtering, denoising, normalizing and other operations to remove interference components and improve data quality. Standard Comparative Analysis: The preprocessed data is compared and analyzed with the IEEE 519 standard to determine the current power quality status of the system. Specifically, the analysis includes whether the voltage deviation is within the allowable range (e.g., ±5% of rated voltage), whether the harmonic content exceeds the standard (e.g., different current distortion rate limits are specified for different harmonics), and whether the system's reactive power is balanced. Problem Localization: If power quality is found to be substandard, the source of the problem is accurately located through data analysis and fault diagnosis algorithms. This involves determining whether the issue stems from abnormal voltage regulation or harmonic suppression in the PET (Power Supply Device), ineffective elimination of local harmonics by the DC converter, or insufficient reactive power compensation in the energy storage system. 3. Collaborative Optimization Strategy Development: 3.1 Determine optimization objectives: Based on the diagnostic results, clarify the key areas for optimization. If the voltage deviation is large, the primary objective is to stabilize the voltage; if harmonics are severely excessive, the focus is on reducing harmonic content; if there is a reactive power imbalance, the objective is to compensate for reactive power to ensure that the overall power quality meets the IEEE 519 standard. 3.2 PET Strategy Formulation: A regulation strategy is formulated for the macroscopic voltage regulation and harmonic suppression functions of the PET. If the voltage is too high, the PET reduces the output voltage by adjusting the turns ratio; if the voltage is too low, the turns ratio is increased. For harmonic suppression, the PET generates a corresponding compensation current command based on the detected harmonic components and content, and injects it into the system to cancel the harmonic current. 3.3 DC Converter Strategy Formulation: Based on the harmonic detection results of the local power supply line, the DC converter adjusts its own control parameters, such as switching frequency and duty cycle, to generate a compensation current that is equal in magnitude and opposite in direction to the harmonic current, thereby eliminating local harmonics and ensuring the power quality of the local power supply. 3.4 Energy Storage System Strategy Formulation: Based on the system's reactive power demand, calculate the amount of reactive power that the energy storage system needs to compensate. When the system's reactive power is insufficient, the energy storage system outputs reactive power to compensate; when reactive power is excessive, the energy storage system absorbs reactive power to maintain the system's reactive power balance. 4. Optimized Command Execution: The power quality collaborative optimization module translates the formulated optimization strategy into specific control commands, which are then sent to the PET (Power Filter), DC converter, and energy storage system via the communication network. For example, it sends voltage regulation parameters and harmonic compensation current commands to the PET, switching control parameters to the DC converter, and reactive power output or absorption commands to the energy storage system. Upon receiving the commands, the PET, DC converter, and energy storage system immediately adjust their operating states and parameters according to the command requirements, executing the corresponding optimization operations. The PET adjusts the voltage and injects compensation current, the DC converter generates harmonic compensation current, and the energy storage system outputs or absorbs reactive power. 5. Real-time Performance Monitoring: After the optimization command is executed, the power quality parameters of the power system are continuously monitored in real time, observing changes in indicators such as voltage, harmonic content, and reactive power to determine the effectiveness of the optimization strategy. The real-time monitored data is fed back to the power quality collaborative optimization module for comparison and analysis with the optimization target. If the power quality indicators do not meet the requirements of the IEEE 519 standard, the module re-evaluates the problem based on the feedback results, adjusts the optimization strategy, and issues the command again for optimization until the power quality meets the standard. 6. System Operation Status Recording and Analysis: Detailed records are kept of all data generated during the power quality collaborative optimization process, including raw monitoring data, optimization strategies, executed commands, and adjusted power quality indicators, forming a historical data archive. The recorded data is analyzed periodically to summarize the experience and patterns of power quality optimization, evaluate the collaborative working effect of PET, DC converters, and energy storage systems, and provide a reference for subsequent system optimization, upgrades, and fault diagnosis.

[0042] Furthermore, the fault detection module performs multi-source data monitoring and intelligent fault analysis and diagnosis; including the following steps: 1. Sensor Deployment and Data Acquisition: Precisely locate key nodes within the power electronic transformer (PET) cluster, DC converter module, power module, and load management unit. These include power devices in the PET, the output of the DC converter, the inverter side of the distributed power supply, the input / output interfaces of the power module, and the load connection points of the load management unit. Deploy high-precision sensors, installing temperature and current sensors at the power devices in the PET to monitor temperature changes and current magnitudes in real time. Install voltage fluctuation monitoring devices at the output of the DC converter to quickly capture instantaneous voltage fluctuations. Deploy frequency detection equipment on the inverter side of the distributed power supply, with a frequency measurement error not exceeding ±0.01Hz. Simultaneously, install vibration sensors at key locations in each module to monitor vibration during equipment operation. Collect data such as voltage, current, frequency, temperature, and vibration in real time through sensors; the sampling frequency is set according to the equipment characteristics. 2. Data Transmission and Preprocessing: A stable and reliable communication network is constructed, employing industrial Ethernet, fiber optic communication, and other methods to transmit data collected by various sensors to edge computing nodes. To ensure the real-time performance and accuracy of data transmission, data transmission priorities are set, with data related to important faults transmitted first. At the edge computing nodes, the received data undergoes preprocessing, including data filtering to remove noise interference; data normalization to unify data of different types and ranges to a specific range; and missing data is filled using linear interpolation or machine learning-based prediction methods to improve data quality. 3. Intelligent analysis and diagnosis: 3.1 Preliminary Analysis of Edge Computing: Edge computing nodes use simple algorithms deployed locally to perform preliminary analysis on preprocessed data, such as calculating statistical characteristics of the data, such as mean, variance, maximum and minimum values, to quickly identify obvious abnormal data and reduce the amount of data uploaded to the cloud. 3.2 Cloud-based Deep Learning Diagnosis: The preliminary analyzed data is uploaded to the cloud for in-depth analysis using a deep learning-based fault diagnosis model. This model is pre-trained with a large amount of normal and fault state data, such as data collected on PET under various fault states like overcurrent and overheating, and DC converters under abnormal voltage and excessive ripple, to train a convolutional neural network (CNN) or long short-term memory network (LSTM) model, enabling it to accurately identify abnormal patterns in equipment operation. 3.3 Data Comparison and Trend Analysis: Data comparison and trend analysis methods are used to compare real-time data with historical data and standard parameters. Normal threshold ranges are set for each parameter, such as the normal range for PET phase current being 80%-120% of the rated current, and the normal range for DC converter output voltage being ±5% of the rated voltage. When real-time data exceeds the threshold range, or when the data trend shows abnormal changes (such as a continuous and rapid rise in temperature or a gradual increase in current fluctuation), it is determined that the system has potential faults or a fault has already occurred. Trend analysis is performed according to the following formula: P=[Σ n i-1 (w i x i y i )+x - Σ n i-1 (w i y i )] / {Σ n i-1 (w i x i 2 )-x - Σ n i-1 (w i x i )+ζΣ n-1 i-1 [r i,i+ 1x i w i (x i -x i+1 )]}; x - =[Σ n i-1 (w i x i )] / [Σ n i-1 (wi )];Σ n i-1 (w i )=1; where P is the slope of the trend line and n is the number of data points; w i This is a weighting coefficient used to reflect the importance of the i-th data point in trend judgment. i The independent variable data is the value of the independent variable at the i-th data point; y i This is the dependent variable data, specifically the value of the dependent variable at the i-th data point; x - It is the average value; r i,i+1 ζ is the autocorrelation coefficient between adjacent data points, representing the correlation between the i-th data point and the (i+1)-th data point. It measures the dependence of data on a time series, with a value ranging from -1 to 1. The closer the absolute value is to 1, the stronger the correlation. ζ is a regularization parameter used to balance fitting error and autocorrelation constraints. It improves the model's ability to fit data trends while preventing the model from overfitting noise in the data. By adjusting the value of ζ, the strength of the autocorrelation constraint can be controlled.

[0043] 4. Fault Determination: Fault determination is performed based on a combination of preliminary edge computing analysis results, cloud-based deep learning diagnostic results, and data comparison and trend analysis results. When a sudden surge in the current of a single phase of the PET circuit exceeds a set threshold of 150%, or when the DC converter output voltage deviates from the normal range for more than 10 seconds, the system automatically identifies this as a fault signal. Once a fault is determined, the fault detection module immediately outputs a fault signal and sends information such as the fault type, fault location, and fault severity to the system's central control unit via the communication network. Simultaneously, an alarm notification is sent to relevant equipment management personnel to facilitate timely action. 5. Fault Response and Handling: Upon receiving a fault signal, the system automatically triggers corresponding protection measures based on the fault type and severity. For example, for a PET overcurrent fault, it automatically cuts off the power supply to the faulty phase; for a DC converter voltage abnormality fault, it adjusts the output voltage or switches to the backup power supply. After receiving the alarm notification, equipment management personnel will go to the site for inspection and handling based on the fault information. The fault detection module provides management personnel with a detailed fault analysis report, including data changes before and after the fault occurred, possible causes of the fault, etc., to assist them in quickly locating and resolving the fault. 6. Fault Record Analysis: Detailed records are kept of all faults, including the time of occurrence, type, location, pre- and post-fault operational data, and corrective actions taken, forming a complete fault file. The fault record data is analyzed periodically to summarize patterns and characteristics of fault occurrences and evaluate the performance and accuracy of the fault detection module. Based on the analysis results, the fault diagnosis model and algorithm are optimized, and monitoring parameters and thresholds are adjusted to improve the reliability and effectiveness of the fault detection module. Furthermore, the anomaly monitoring module monitors anomalies through multimodal collaborative monitoring, including the following steps: 1. Sensing Equipment Deployment: Identify key components within equipment such as power electronic transformers (PETs) requiring monitoring, including power devices and windings, as well as the external environment. Based on monitoring needs, select multimodal sensing equipment such as temperature and humidity sensors, fiber optic sensors, millimeter-wave radar, smoke sensors, hazardous gas detectors, and infrared thermal imagers. Install fiber optic sensors in key internal components of the PET for high-precision current and temperature measurement. Install millimeter-wave radar at suitable locations externally to monitor vibration and deformation. Deploy infrared thermal imagers to comprehensively detect hotspot distribution on the equipment surface. Install temperature and humidity sensors around and in suitable internal areas of the equipment to acquire real-time environmental temperature and humidity information. Connect all deployed sensing equipment via wired or wireless communication to form a collaborative sensing network covering key internal components and the external environment.

[0044] 2. Data Acquisition: Appropriate data acquisition frequency and accuracy parameters are set for each sensing device to ensure accurate and timely acquisition of the required data. Each sensing device begins real-time data acquisition according to the set parameters, including current and temperature data collected by fiber optic sensors, equipment vibration and deformation data collected by millimeter-wave radar, hotspot data collected by infrared thermal imagers, and temperature and humidity data collected by temperature and humidity sensors. Simultaneously, other relevant data such as current and voltage are collected.

[0045] 3. Data preprocessing: Preprocessing the collected data, including data cleaning, data standardization, and data synchronization; Data cleaning: Cleaning the collected data to remove noise, outliers, and missing values.

[0046] Data standardization: Standardizing data of different types and ranges.

[0047] Data synchronization: Synchronize data from multiple sources to ensure data consistency over time.

[0048] 4. Multi-source data fusion analysis: Based on data characteristics and monitoring needs, select fusion methods such as Deep Belief Networks (DBNs) based on machine learning. Fusion of pre-processed multi-source data comprehensively considers information from each data source to extract more comprehensive and accurate features. For example, fusing data such as current, temperature, vibration, and hotspot data collected by infrared thermal imagers to analyze the correlations between them.

[0049] Anomaly feature extraction: Extract features related to anomalies from the fused data, such as abnormal temperature rise, sudden current change, and increased equipment vibration.

[0050] 5. Anomaly Diagnosis: Based on historical data and expert experience, an anomaly diagnosis model is established. Machine learning algorithms, such as decision trees and neural networks, are used to classify normal and abnormal states. Extracted anomaly features are input into the diagnostic model to determine if an anomaly exists. If the model identifies an anomaly, the type and severity of the anomaly are further determined. Once an anomaly is detected, an early warning is immediately issued to relevant personnel via audible and visual alarms, SMS, email, etc., while detailed information about the anomaly is displayed in the monitoring system, including the time, location, type, and severity of the anomaly, so that timely measures can be taken to address it.

[0051] 6. Result Verification: Upon receiving the early warning information, relevant personnel will go to the site to verify the anomaly, confirming its existence and details. The verification results will be fed back to the anomaly monitoring module for correction and optimization of the diagnostic model. If false alarms or missed alarms are found, the causes will be analyzed, and model parameters or fusion methods will be adjusted to improve the accuracy and reliability of anomaly monitoring.

[0052] This invention provides a smart energy monitoring method based on power electronic transformers, comprising the following steps: S1. Deploy temperature and current sensors at key components such as power devices in the power electronic transformer (PET) cluster; install voltage fluctuation monitoring devices at the output of the DC converter module; deploy frequency detection equipment on the distributed power inverter side of the power module; and install corresponding monitoring sensors at key nodes of the load management unit. Simultaneously, install environmental monitoring equipment consisting of temperature and humidity sensors, smoke sensors, and hazardous gas detectors in various areas of the system.

[0053] S2 collects real-time and accurate operating data from the intelligent monitoring units of the power electronic transformer (PET) cluster, the adaptive control chips of the DC converter modules, various devices in the power supply modules, and the load management units, including voltage, current, frequency, temperature, and vibration data. Environmental monitoring equipment collects environmental data such as temperature, humidity, and harmful gas concentrations. The collected data is transmitted to the control center via the network, while the PET cluster's data is simultaneously transmitted to edge computing nodes. S3, the data processing module cleans the data collected from power electronic transformer (PET) clusters, DC converter modules, power supply modules, load management units, etc., removing noise, outliers and missing values; it performs standardization processing to unify the range of different types of data; and it synchronizes multi-source data through timestamps and other methods to improve data quality and prepare for in-depth analysis. S4. The anomaly monitoring module analyzes data collected from various components such as the power electronic transformer (PET) cluster and DC converter module through multimodal collaborative analysis, extracts abnormal features such as abnormal temperature rise and sudden current change, and determines whether there is an abnormal situation.

[0054] The S5 fault detection module combines edge computing nodes and cloud-based intelligent algorithms, using deep learning models, data comparison, and trend analysis to conduct in-depth analysis of the data and determine whether the system has potential faults or has already experienced a fault. S6. The energy dispatch control module analyzes the system's energy supply and demand based on the collected data on power module output, load management unit requirements, and energy storage power status, and formulates energy dispatch strategies. Based on model predictive control algorithms, the module prioritizes distributed power supply according to the power module output, load management unit requirements, and energy storage power status, supplementing any shortfall with the main grid. When there is a surplus of renewable energy, the module controls the charging of energy storage power; when the system power is insufficient, it allows the energy storage power to discharge, thus achieving rational energy allocation. S7, the power quality collaborative optimization module coordinates the power electronic transformer (PET) cluster, DC converter module and energy storage power supply to work together. The PET cluster is responsible for macro voltage regulation and harmonic suppression, the DC converter module eliminates local power supply harmonics, and the energy storage power supply compensates for reactive power, ensuring that the power quality of the whole system meets the IEEE 519 standard. S8, the load management unit manages different types of loads according to a hierarchical load control strategy. Critical loads are powered by both uninterruptible power supplies (UPS) and direct power lines from the PET cluster; important loads achieve dynamic power allocation through intelligent distribution boxes; adjustable loads are connected to the demand response platform, and adjust power consumption through DC converter modules based on real-time grid status and electricity price signals to achieve peak shaving and valley filling. S9. When the fault detection module determines that a system fault has occurred, the fault emergency handling module is immediately activated. The power electronic transformer (PET) cluster quickly disconnects the faulty line, the DC converter module switches to the backup power supply path, and the energy storage power supply prioritizes power supply to critical loads. The energy dispatch control module uses blockchain technology to achieve network-wide sharing of fault information, accelerating fault location. The alarm module is activated, sound and light alarms are issued at key equipment such as PET control cabinets, DC converter cabinets, and energy storage battery cabinets, fault information is pushed to maintenance personnel via mobile terminals, and the fault situation is visualized on the control center's large screen, facilitating rapid response and fault handling by maintenance personnel. S10. Data storage devices employ hard disk arrays (RAID), distributed storage systems (such as Ceph), etc., to store and manage equipment operation data generated by power electronic transformer (PET) clusters, DC converter modules, power supply modules, load management units, etc., as well as fault data detected by fault detection modules and energy dispatch data formulated by energy dispatch control modules, providing data support for subsequent data analysis, system optimization, and fault tracing.

[0055] The working principle of this invention's smart energy system based on power electronic transformers is as follows: Temperature and current sensors are deployed at key components such as power devices in the power electronic transformer (PET) cluster; a voltage fluctuation monitoring device is installed at the output of the DC converter module; a frequency detection device is deployed on the distributed power inverter side of the power module; and corresponding monitoring sensors are installed at key nodes of the load management unit. Environmental monitoring equipment consisting of temperature and humidity sensors, smoke sensors, and hazardous gas detectors is installed in various areas of the system. Real-time and accurate data collection is performed on the voltage, current, frequency, temperature, vibration, and other operating data of the intelligent monitoring unit of the PET cluster, the adaptive control chip of the DC converter module, various devices of the power module, and the load management unit. Environmental monitoring equipment also collects environmental data such as temperature and humidity, and hazardous gas concentrations. The data processing module preprocesses the collected data; the anomaly monitoring module performs multimodal collaborative analysis of the data collected from the PET cluster, DC converter module, and other components to extract abnormal features such as abnormal temperature increases and sudden current changes, and to determine whether any abnormalities exist. The fault detection module combines edge computing nodes and cloud-based intelligent algorithms, employing deep learning models, data comparison, and trend analysis to conduct in-depth data analysis and determine whether the system has potential faults or has already experienced faults. The energy dispatch and control module analyzes the system's energy supply and demand based on collected power module output, load management unit requirements, and energy storage power status data, and formulates energy dispatch strategies. The power quality collaborative optimization module coordinates the collaborative work of the power electronic transformer (PET) cluster, DC converter module, and energy storage power. The PET cluster is responsible for macroscopic voltage regulation and harmonic suppression, the DC converter module eliminates local power supply harmonics, and the energy storage power compensates for reactive power, ensuring that the power quality of the entire system meets the IEEE 519 standard. The load management unit manages different types of loads according to a hierarchical load control strategy. Critical loads are dually powered by uninterruptible power supplies (UPS) and direct power lines from the PET cluster; important loads achieve dynamic power allocation through intelligent distribution boxes; adjustable loads are connected to the demand response platform, and based on real-time grid status and electricity price signals, the DC converter module adjusts power consumption to achieve peak shaving and valley filling. When the fault detection module determines that a system fault has occurred, the fault emergency handling module immediately activates. The power electronic transformer (PET) cluster quickly disconnects the faulty line, the DC converter module switches to the backup power supply path, and the energy storage power supply prioritizes power to critical loads. The energy dispatch control module uses blockchain technology to achieve network-wide sharing of fault information, accelerating fault location. The alarm module activates, sounding and visual alarms at key equipment such as the PET control cabinet, DC converter cabinet, and energy storage battery cabinet. Data storage devices provide long-term storage and management of various types of data.

[0056] This invention analyzes collected data through an anomaly monitoring module and a fault detection module, enabling timely detection of abnormal conditions such as abnormal temperature increases and sudden current changes, as well as potential fault hazards. This provides strong support for equipment maintenance and fault prevention, reduces equipment failure rates, and ensures stable system operation. The energy dispatch control module, based on model predictive control algorithms, prioritizes distributed power supply according to power output, load demand, and energy storage status data. It rationally allocates power between the main grid and energy storage, achieving efficient utilization of new energy and rational energy distribution. It charges when new energy is in surplus and discharges when power is insufficient, avoiding energy waste, smoothing power fluctuations, improving energy utilization efficiency, reducing energy costs, and simultaneously reducing dependence on traditional energy sources, thus promoting sustainable energy development. The power quality collaborative optimization module coordinates the collaborative work of the PET cluster, DC converter module, and energy storage. The PET cluster is responsible for macroscopic voltage regulation and harmonic suppression, the DC converter module eliminates local power supply harmonics, and the energy storage compensates for reactive power, ensuring power quality from multiple levels. This ensures that the power quality of the entire system meets the IEEE 519 standard, improving overall system performance and reliability. The load management unit adopts a hierarchical load control strategy, providing differentiated management for different types of loads.

[0057] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart energy monitoring method based on power electronic transformers, characterized in that, Includes the following steps: S1. Install sensors at key nodes of the power electronic transformer (PET) cluster, DC converter module, power supply module, and load management unit; install environmental monitoring equipment in various areas of the system; S2: Real-time acquisition of equipment operation data; environmental monitoring equipment acquires environmental data; data from the PET cluster is transmitted to edge computing nodes. S3, the data processing module preprocesses the various types of data collected; S4. The anomaly monitoring module analyzes the data collected from each part through multimodal collaborative analysis, extracts anomaly features, and determines whether there are any anomalies. S5. The fault detection module performs in-depth analysis of the data to determine whether there are potential faults in the system or whether a fault has already occurred. S6, the energy dispatch and control module analyzes the system's energy supply and demand and formulates energy dispatch strategies; S7, the power quality collaborative optimization module ensures that the power quality of the entire system meets the standards; S8. The load management unit manages different types of loads according to the hierarchical load control strategy; S9. When the fault detection module determines that a system fault has occurred, the fault emergency handling module is immediately activated and the alarm module issues an alarm. S10, Data storage device, stores and manages various types of data.

2. The intelligent energy monitoring method based on power electronic transformers according to claim 1, characterized in that: Step S8 includes the following steps: S81. Load Classification Access: Includes critical load access, important load access, and adjustable load access; S82. Critical load power supply guarantee: During normal system operation, critical loads are prioritized to be powered by the PET direct supply line; when the PET direct supply line malfunctions, it switches to battery inverter power supply mode to maintain uninterrupted operation of critical loads. S83, Dynamic power distribution for critical loads; S84 Adjustable Load Demand Response: Receives real-time grid status information from the grid dispatch center, analyzes and processes the information to determine whether power regulation of the adjustable load is required; generates power regulation commands based on preset regulation strategies; and sends the regulation commands to the DC converters at the front end of each adjustable load.

3. The intelligent energy monitoring method based on power electronic transformers according to claim 2, characterized in that: Step S83 includes the following steps: S83.1 Data Acquisition and Analysis: Real-time acquisition of power consumption data of important loads, analysis of historical power consumption data and current operating status of loads, and prediction of power demand of each load in future periods; S83.2 Power Allocation Strategy Formulation: Based on the forecast results and the real-time power supply capacity of the power grid, formulate a dynamic power allocation strategy; S83.3 Power Regulation Execution: Sends power distribution instructions to the intelligent distribution box to adjust the current of each load branch and realize dynamic power distribution.

4. The intelligent energy monitoring method based on power electronic transformers according to claim 1, characterized in that: Step S6 includes the following steps: S61. Real-time data acquisition: Real-time data acquisition, including power output data acquisition, load demand data acquisition, and energy storage status data acquisition; S62. System Status Analysis: S62.1 Data Integration and Processing: Integrate the collected power output, load demand and energy storage status data; S62.2 Energy Supply and Demand Assessment: Analyze the balance between power output and load demand to determine the current system status; assess the adjustment capability of energy storage equipment under different operating conditions; S63, Scheduling strategy formulation; S64. Execution of dispatching instructions: The established energy dispatching strategy is transformed into specific control instructions. After receiving the dispatching instructions, each device and module immediately executes the corresponding operation. S65. Dynamic Adjustment and Optimization: Continuously monitor the system's operating status in real time and evaluate the effectiveness of the scheduling strategy; adjust the current scheduling strategy as needed.

5. The intelligent energy monitoring method based on power electronic transformers according to claim 4, characterized in that: Distributed power sources are given priority; if the output power of distributed power sources cannot meet the load demand, the power gap is calculated and supplementary power is provided by the main grid; when there is a surplus of new energy power generation, the energy storage system is controlled to perform charging operations; when the system power is insufficient, the energy storage system is scheduled to discharge.

6. The intelligent energy monitoring method based on power electronic transformers according to claim 1, characterized in that: Step S7 includes the following steps: S71. Real-time data acquisition and monitoring: Real-time acquisition of data related to voltage, current, harmonic content, and reactive power; S72. Power quality problem diagnosis: Compare and analyze the preprocessed data with the IEEE 519 standard to determine the current power quality status of the system; if the power quality is found to be non-compliant with the standard, accurately locate the source of the problem. S73. Collaborative optimization strategy formulation: Based on the diagnostic results, clarify the key directions for optimization; formulate adjustment strategies for the macroscopic voltage regulation and harmonic suppression functions of PET; adjust the control parameters of the DC converter according to the harmonic detection results of the local power supply line; calculate the amount of reactive power that the energy storage system needs to compensate based on the reactive power requirements of the system. S74, Optimization Instruction Execution: The formulated optimization strategy is converted into specific control instructions and sent to the PET, DC converter and energy storage system respectively; S75. Real-time effect monitoring: Continuously monitor the power quality parameters of the power system in real time, judge the effect of the optimization strategy; re-evaluate the problem based on the feedback results, and adjust the optimization strategy accordingly. S76. System Operation Status Record Analysis: Generate historical data archives for various data during the power quality collaborative optimization process, and analyze the recorded data periodically.

7. The intelligent energy monitoring method based on power electronic transformers according to claim 1, characterized in that: Step S5 includes the following steps: S51. Sensor Deployment and Data Acquisition: Real-time acquisition of voltage, current, frequency, temperature, and vibration data via sensors; S52. Data transmission and preprocessing: Transmit the data collected by each sensor to the edge computing node; preprocess the received data; S53, Intelligent Analysis and Diagnosis; S54. Fault Determination: Based on the preliminary analysis results of edge computing, data comparison, and trend analysis results, fault determination is performed. S55. Fault Response and Handling: Upon receiving a fault signal, the corresponding protection measures are automatically triggered according to the fault type and severity. S56. Fault Record Analysis: Record the faults that occur in detail and analyze the fault record data regularly.

8. The intelligent energy monitoring method based on power electronic transformers according to claim 7, characterized in that: Step S53 includes the following steps: S53.1 Preliminary analysis of edge computing: Perform preliminary analysis on the preprocessed data to quickly identify obvious abnormal data; S53.2, Cloud-based Deep Learning Diagnosis: Upload the data after preliminary analysis to the cloud and use a deep learning-based fault diagnosis model for in-depth analysis; S53.3 Data Comparison and Trend Analysis: Using data comparison and trend analysis methods, real-time data is compared with historical data and standard parameters to conduct trend analysis.

9. The intelligent energy monitoring method based on power electronic transformers according to claim 1, characterized in that: Step S4 includes the following steps: S41. Deployment of sensing devices: For power electronic transformer-related equipment, identify key internal parts and external environmental areas, select multimodal sensing devices, and install them in the key parts and external areas respectively; S42. Data Acquisition: Collect data from each sensor in real time; S43. Data preprocessing: Preprocessing the collected data, including data cleaning, data standardization, and data synchronization; S44. Multi-source data fusion analysis: The Deep Belief Network (DBN) fusion method is selected to fuse the preprocessed multi-source data and extract features related to anomalies from the fused data. S45. Abnormal Situation Diagnosis: Establish an abnormal situation diagnosis model; input the extracted abnormal features into the diagnosis model to determine whether there is an abnormal situation; S46. Result Verification: Verify abnormal situations, feed back the results of on-site verification to the anomaly monitoring module, and correct and optimize the diagnostic model.

10. A smart energy system based on a power electronic transformer, comprising: The system comprises a power electronic transformer (PET) cluster, a DC converter module, a power supply module, a data processing module, a load management unit, an energy dispatch control module, a power quality collaborative optimization module, a fault detection module, a fault emergency handling module, environmental monitoring equipment, an anomaly monitoring module, data storage equipment, and a control center; its features are: Power electronic transformer PET cluster: adopts a combined architecture of modular multilevel power electronic transformer (MMC-PET) and isolated bidirectional full-bridge (DAB) PET; DC converter module: Constructs a hierarchical DC converter network and dynamically adjusts output parameters according to load changes; Power modules include main power systems, distributed power sources, and energy storage power sources; Data processing module: preprocesses the collected power equipment and environmental data, including data cleaning, standardization, and data synchronization; Load Management Unit: Employs a hierarchical load control strategy, dividing the load into critical loads, important loads, and adjustable loads; Energy dispatch control module: Formulates energy dispatch strategies based on power output, load demand, and energy storage status; Power quality collaborative optimization module: Ensures that the power quality of the entire system meets the standards; Fault detection module: performs multi-source data monitoring and intelligent fault analysis and diagnosis; Fault emergency handling module: When the system detects a fault, it performs emergency handling for the fault; Alarm module: Install audible and visual alarms at key equipment locations in the system; Environmental monitoring equipment: Real-time monitoring of system operating environment data; Anomaly monitoring module: Monitors abnormal situations through multimodal collaborative monitoring; Data storage devices: devices used to store data; Control Center: Connected to the network of each module.

Images