Modularized redundancy design method of power electronic transformer (PET) and electronic transformer

Abstract

The invention belongs to the technical field of power electronics, particularly relates to a modular redundancy design method of a power electronic transformer (PET), aims to improve the reliability, fault-tolerant capability and maintainability of the power electronic transformer, and is suitable for application scenes needing efficient electric energy conversion, such as an intelligent power grid, new energy access and electric vehicle charging. Comprising a module division step, a redundancy configuration step, a fault detection step, a fault isolation step, a redundancy switching step and a health management step, through a reasonable modular structure design and a redundancy strategy, the reliability and the fault-tolerant capability of the PET system are improved, when part of modules break down, the system can still keep stable operation, and the reliability of the system is improved. And fault isolation and system reconstruction are rapidly realized, the utilization efficiency of a redundant module is optimized, and the cost and the size of the system are reduced.

Description

Technical Field

[0001] The present invention belongs to the field of power electronics technology, and specifically relates to a modular redundant design method for power electronic transformers (PETs). The method aims to improve the reliability, fault tolerance, and maintainability of power electronic transformers, and is suitable for application scenarios requiring efficient power conversion, such as smart grids, new energy access, and electric vehicle charging. Background Art

[0002] Power electronic transformer (PET) is a new type of electric energy conversion device based on power electronics technology and high-frequency transformer technology. Compared with traditional electromagnetic transformers, PET has the advantages of small size, light weight, strong power quality regulation ability, and the ability to achieve AC / DC hybrid power distribution. It has broad application prospects in modern power systems. However, since PET contains a large number of power electronic devices, such as power semiconductor devices and control chips, these devices are easily affected by factors such as overvoltage, overcurrent, and temperature changes during long-term operation, causing failures.

[0003] Once a key component fails, it may cause the entire PET system to shut down, seriously affecting the stable operation and power supply reliability of the power system.

[0004] Currently, some PET designs use a simple redundancy backup strategy, which is to simply connect the same modules in parallel. Although this approach improves the reliability of the system to a certain extent, it has problems such as low utilization of redundant modules, low efficiency of fault diagnosis and switching, increased system cost and volume. Some designs use complex centralized control to manage redundant modules, but centralized control is too dependent on the main controller. Once the main controller fails, the entire redundancy management mechanism will fail, and it will be difficult to quickly achieve fault isolation and system reconstruction. Therefore,

[0005] A new modular redundancy design approach is urgently needed to achieve efficient fault tolerance, flexible system reconfiguration, and optimal resource utilization. Summary of the Invention

[0006] The purpose of the present invention is to provide a modular redundant design method for a power electronic transformer (PET) to address the problems existing in the prior art. Through reasonable modular structural design and redundancy strategy, the reliability and fault tolerance of the PET system are improved. When some modules fail, the system can still maintain stable operation and quickly achieve fault isolation and system reconstruction. At the same time, the utilization efficiency of redundant modules is optimized, and the system cost and volume are reduced.

[0007] The technical solution of the present invention is:

[0008] A modular redundant design method for a power electronic transformer (PET) comprises the following steps:

[0009] Module division: The overall function of the power electronic transformer is divided into multiple functionally independent power conversion modules. Each power conversion module includes an input filter unit, a full-bridge or multi-level conversion circuit unit, a high-frequency transformer unit, and an output rectifier unit. The input filter unit is used to filter out high-frequency harmonics in the input power supply. The full-bridge or multi-level conversion circuit unit converts the input DC or AC power into high-frequency AC power, achieving electrical isolation and voltage level conversion through a high-frequency transformer. The high-frequency transformer also converts the high-frequency AC power into DC power. The output filter unit filters the rectified DC voltage and outputs stable DC or AC power. Each power conversion module is connected through a standardized electrical interface and communication interface. The electrical interface is used to achieve power transmission, and the communication interface is used for information exchange and coordinated control between modules. Each module is electrically connected and communicates through standardized interfaces.

[0010] Redundancy: Redundant modules are installed at each functional level in the PET system. The number of redundant modules is determined based on system reliability requirements and cost budget. Specifically, redundant input filter modules are installed on the input side, a certain ratio (N+M, where N is the number of normal operating modules and M is the number of redundant modules, 1≤M≤N / 2) of redundant power conversion modules are installed in the power conversion link, and redundant output rectifier modules are installed on the output side. The redundant modules have the same hardware structure and functions as the working modules. Under normal operating conditions, the redundant modules are in hot standby mode, monitoring their own status in real time and exchanging information with the working modules to obtain system operating parameters.

[0011] Fault Detection: Sensors and monitoring circuits are built into each power conversion module and key components to collect voltage, current, temperature, and switching device status parameters in real time. This collected data is analyzed by a data processing module based on a Kalman filter algorithm, and a module fault is determined based on preset fault thresholds. Each power conversion module incorporates an independent fault detection unit, which collects key electrical parameters (such as voltage, current, and temperature) within the module in real time and analyzes and processes the collected data using a preset fault diagnosis algorithm. When a fault such as overvoltage, overcurrent, device overheating, or short circuit is detected within the module, the fault detection unit immediately transmits the fault information to adjacent modules and the system management unit via the communication interface.

[0012] Fault isolation: When a fault is detected in a power conversion module, the isolation circuit composed of fast solid-state relays or intelligent power switching devices cuts off the electrical connection between the faulty module and the system within 5ms, and sends a fault signal to the system control center; when a power conversion module fails, the adjacent modules receive the fault information and quickly cut off the electrical connection between themselves and the faulty module by controlling the corresponding electronic switches to achieve fault isolation.

[0013] Redundant switching: After the system control center receives the fault signal, based on the hot standby strategy, the system management unit selects a suitable backup module from the redundant modules according to the fault information and the current system operating status, and immediately activates the corresponding redundant power conversion module. Through pre-charging control and phase synchronization algorithm, the redundant module is seamlessly connected to the system within 10ms to replace the faulty module. During the switching process, a smooth transition control strategy is adopted to adjust the output voltage and current of the backup module and other working modules to ensure a smooth transition of the system output power, avoiding large fluctuations in the system output voltage and current due to module switching.

[0014] Health Management: After isolating the faulty module and switching to the backup module, the system management unit reconfigures the entire PET system based on the status of each active and backup module. By optimizing the control algorithm, it redistributes the load power of each active module, ensuring optimal system operation under the new module combination. Simultaneously, the system management unit monitors the operating status of each module in real time and dynamically adjusts the activation and deactivation of redundant modules based on load changes and module health, achieving a balance between system efficiency and reliability.

[0015] A power electronic transformer implementing the above-mentioned modular redundant design method comprises:

[0016] Multiple power conversion modules with the same structure and function;

[0017] Redundant module group, including redundant input filter module, redundant power conversion module and redundant output rectifier module;

[0018] The fault detection and isolation unit consists of sensors, monitoring circuits, and fast solid-state relays distributed across each module. It is used to monitor module status in real time and isolate faulty modules.

[0019] The system control center uses a controller architecture that combines FPGA and DSP to receive fault signals, control redundant module switching, and adjust power distribution strategies;

[0020] The communication network uses high-precision clock synchronous Ethernet based on IEEE 1588 to realize data interaction and control command transmission between modules;

[0021] The health management module has a built-in state-life prediction model and data analysis algorithm to predict module life and generate maintenance plans.

[0022] Specifically, each of the above-mentioned power conversion modules includes:

[0023] Input filter unit, used for filtering the input AC power;

[0024] Full-bridge or multi-level conversion circuit unit, converting the filtered AC power into high-frequency AC power;

[0025] High-frequency transformer unit to achieve electrical isolation and voltage level conversion;

[0026] The output rectifier unit converts high-frequency AC power into stable DC power or AC power of a specific frequency.

[0027] The method provided by the present invention has the following beneficial effects: 1. High reliability and fault tolerance: Through modular redundant design, when a power conversion module fails, the redundant module can be quickly put into operation to replace the failed module, avoiding the shutdown of the entire PET system due to the failure of a single module, greatly improving the reliability and fault tolerance of the system, and ensuring the stable power supply of the power system.

[0028] 2. Rapid fault handling: Distributed fault detection and diagnosis mechanisms, as well as rapid fault isolation and switching strategies, can complete fault detection, isolation, and backup module access in a short period of time, reducing system fault recovery time and minimizing the impact of faults on the power system.

[0029] 3. Efficient Resource Utilization: Redundant modules are in hot standby mode and monitor their own and system status in real time, enabling rapid commissioning when needed, thus improving redundancy module utilization. Furthermore, system reconfiguration and optimization control strategies dynamically adjust module operating states based on actual system needs, achieving optimal allocation of system resources and reducing system cost and size.

[0030] 4. Flexibility and scalability: The standardized modular structure and interface design make the PET system easy to expand and maintain. The number of power conversion modules and redundant modules can be easily increased or decreased according to actual needs to adapt to different power levels and reliability requirements. In addition, in the event of a module failure, it is easy to replace and repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the processing flow of the method provided by the present invention. DETAILED DESCRIPTION

[0032] The technical solution of the present invention is described in detail below with reference to the accompanying drawings and specific implementation methods.

[0033] Example 1

[0034] like Figure 1 Shown is a flow chart of the modular redundancy design method of the present invention.

[0035] This embodiment takes a three-phase power electronic transformer (PET) applied to a smart microgrid as an example to specifically illustrate the modular redundancy design method of the present invention.

[0036] 1. Modular structure construction

[0037] The PET system is divided into six power conversion modules, each rated at 50 kW. Each power conversion module, designed according to the aforementioned structure, includes an input filter circuit, a high-frequency inverter, a high-frequency transformer, a high-frequency rectifier, and an output filter circuit. The modules are connected via standardized electrical interfaces (such as quick-swap connectors) and communication interfaces (based on the CAN bus protocol) to enable power transmission and information exchange.

[0038] 2. Redundant module configuration

[0039] According to the system reliability requirements, two redundant power conversion modules are configured. The redundant module and the working module have exactly the same hardware and software design. When the system is operating normally, the redundant module is in hot standby state, monitoring its own voltage, current, temperature and other parameters in real time, and maintaining information exchange with the working module and system management unit through the communication interface.

[0040] 3. Fault detection and diagnosis

[0041] An independent fault detection unit is installed within each power conversion module. This unit uses high-precision voltage, current, and temperature sensors to collect key electrical parameters within the module. Pre-set fault diagnosis algorithms include overvoltage detection (identifying an overvoltage fault when the input or output voltage exceeds 120% of the rated value), overcurrent detection (identifying an overcurrent fault when the current exceeds 150% of the rated current), and device overheating detection (identifying an overheat fault when the power semiconductor device temperature exceeds 120°C). When the fault detection unit detects a fault, it immediately transmits information such as the fault type and occurrence time via the CAN bus to adjacent modules and the system management unit.

[0042] 4. Fault isolation and fast switching

[0043] Assume an overcurrent fault occurs in one of the active modules. After its fault detection unit transmits fault information, adjacent modules receive the information and, through the solid-state relays within their control modules, quickly sever the electrical connection to the faulty module, isolating the fault. Simultaneously, the system management unit selects a redundant module in optimal standby status based on the redundant module's status information. It activates its startup circuit and configures the standby module's parameters through the communication interface to match its output voltage and current with those of the other active modules. During the switching process, a smooth transition control strategy based on a dual closed-loop voltage and current system is employed to adjust the outputs of the standby module and the other active modules, ensuring a smooth transition in system output power within 0.1 seconds, thus avoiding any impact on the microgrid.

[0044] 5. System reconstruction and optimization control

[0045] After isolating the faulty module and switching to a backup module, the system management unit redistributes the load power based on the current module status. By optimizing the control algorithm, the load ratios of the active modules are balanced as much as possible, while also considering the efficiency characteristics of each module, ensuring that the entire PET system operates at optimal efficiency under the new module combination. During subsequent operation, the system management unit continuously monitors the status of each module and dynamically adjusts the activation and deactivation of redundant modules as needed when the load changes, ensuring that the system remains efficient and reliable.

[0046] Example 2

[0047] This embodiment provides a power electronic transformer that implements the above-mentioned modular redundancy design method, including:

[0048] Multiple power conversion modules with the same structure and function, each of which includes: an input filter unit for filtering the input alternating current; a full-bridge or multi-level conversion circuit unit for converting the filtered alternating current into high-frequency alternating current; a high-frequency transformer unit for achieving electrical isolation and voltage level conversion; and an output rectifier unit for converting the high-frequency alternating current into stable direct current or alternating current of a specific frequency.

[0049] The redundant module group includes a redundant input filter module, a redundant power conversion module and a redundant output rectifier module.

[0050] The fault detection and isolation unit consists of sensors, monitoring circuits, and fast solid-state relays distributed in each module. It is used to monitor the module status in real time and isolate the faulty module.

[0051] The system control center adopts a controller architecture that combines FPGA and DSP to receive fault signals, control redundant module switching, and adjust power distribution strategies.

[0052] The communication network adopts high-precision clock synchronous Ethernet based on IEEE 1588 to realize data interaction and control instruction transmission between modules.

[0053] The health management module has a built-in state-life prediction model and data analysis algorithm to predict module life and generate maintenance plans.

[0054] Example 3

[0055] This embodiment provides specific operations for analyzing collected data by a data processing module based on a Kalman filter algorithm.

[0056] The following are the key logic and processes for processing collected data:

[0057] 1. The core framework of the data processing module includes:

[0058] Input layer: Receives the raw data collected by the sensor (such as temperature, displacement, speed and other time series signals) and obtains the system state transition model parameters (such as state transition matrix A, observation matrix H).

[0059] Algorithm layer: implements the prediction-update iterative process of Kalman filtering and outputs the optimal state estimate.

[0060] Output layer: provides filtered data sequences or state parameters for subsequent decision making.

[0061] 2. The specific steps of Kalman filtering in data processing are as follows:

[0062] S1: Model definition and initialization

[0063] System modeling: its state equation:

[0064] (w k is the process noise, which obeys the Gaussian distribution N(0,Q)).

[0065] Observation equation: (v k is the observation noise, which obeys the Gaussian distribution N(0,R)).

[0066] Initialization: Set the initial state estimate \hat{x}_0 (such as the initial sensor reading or historical mean).

[0067] Initialize the covariance matrix P0 (reflecting the uncertainty of the initial estimate, usually set to a diagonal matrix).

[0068] S2: Prediction stage (estimating the current status based on historical data)

[0069] State prediction: Calculate the prior state at the current moment based on the optimal estimate at the previous moment:

[0070]

[0071] Covariance prediction: Update the prior covariance to reflect the accumulated uncertainty of the prediction process:

[0072] P k|k-1 =AP k-1 A T +Q.

[0073] S3: Update phase (correcting predictions based on measured data)

[0074] Kalman gain calculation: balance the weights of predicted values ​​and observed values, the formula is:

[0075] K k =P k|k-1 H T (HP k|k-1 H T +R) -1 .

[0076] If the observation noise R is small, the gain K k The larger the value, the more trust is placed in the measured data; otherwise, the more reliance is placed on model predictions.

[0077] Status update: using measured data z k Correct the predicted state to obtain the posterior optimal estimate:

[0078]

[0079] in( ) is the observation residual.

[0080] Covariance update: Reduce the uncertainty of the estimate, the formula is:

[0081] P k =(IK k H)P k|k-1 , the updated P k Will be the P at the next moment k-1 .

[0082] S4: Iterative optimization (loop steps 2-3)

[0083] Each time a new set of data z is obtained k , repeat the prediction-update process to make the state estimate Gradually approaching the true value, while the covariance P k Converges to a stable minimum.

[0084] When using this method, you need to pay attention to Q (process noise covariance): it reflects the accuracy of the system dynamic model. The more accurate the model, the smaller Q. R (observation noise covariance): it can be obtained through sensor calibration or historical data statistics (such as the variance of multiple measurements). Matrix operations can be simplified to recursive formulas, reducing the amount of calculation and adapting to embedded devices (such as FPGA / MCU). Abnormal data processing: add residual threshold judgment, when When the threshold is exceeded, the data is considered abnormal and R can be temporarily increased or reinitialization can be triggered.

[0085] Through the above process, the Kalman filter module can effectively extract the true state of the dynamic system from noisy data, improving data reliability and decision-making accuracy.

[0086] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or some technical features can be replaced by equivalents without departing from the spirit of the technical solution of the present invention. They should all be included in the scope of the technical solution for protection of the present invention.

Claims

1. A modular redundant design method for a power electronic transformer (PET), characterized in that: The steps include: Module division: The overall function of the power electronic transformer is divided into multiple functionally independent power conversion modules. Each power conversion module includes an input filter unit, a full-bridge or multi-level conversion circuit unit, a high-frequency transformer unit, and an output rectifier unit. Each module is electrically connected and communicates with each other through standardized interfaces. Redundancy configuration: Redundant modules are set up at each functional level. Specifically, redundant input filter modules are set up on the input side, a certain ratio (N+M, N is the number of normal working modules, M is the number of redundant modules, 1≤M≤N / 2) of redundant power conversion modules are set up in the power conversion link, and redundant output rectifier modules are set up on the output side; Fault detection: Built-in sensors and monitoring circuits in each power conversion module and key components collect voltage, current, temperature, and switch device status parameters in real time. The collected data is analyzed by a data processing module based on the Kalman filter algorithm, and the module is judged to be faulty based on the preset fault threshold. Fault isolation: When a fault is detected in a power conversion module, the electrical connection between the faulty module and the system is cut off within 5ms through an isolation circuit composed of fast solid-state relays or intelligent power switching devices, and a fault signal is sent to the system control center; Redundancy switching: After receiving a fault signal, the system control center immediately activates the corresponding redundant power conversion module based on the hot standby strategy. Through pre-charge control and phase synchronization algorithms, the redundant module seamlessly connects to the system within 10ms and replaces the faulty module. At the same time, the power allocation strategy of other normally operating modules is adjusted to maintain the stable output of the power electronic transformer. Health management: Establish a module health management system based on the state-life prediction model. By analyzing the module's historical operation data and real-time monitoring data, a particle filter algorithm is used to predict the remaining service life of each module. Module maintenance and replacement plans are planned in advance based on the prediction results.

2. The modular redundancy design method according to claim 1, characterized in that: In the module division step, a combination of function tree decomposition and reliability block diagram analysis is used to determine the division scheme of the power conversion module, so that the coupling degree between modules is less than 0.3 and the impact of a single module failure on the overall system function does not exceed 20%.

3. The modular redundancy design method according to claim 1, characterized in that: In the redundant configuration step, the number of redundant modules M is determined by calculating the number M using a Markov reliability model based on the reliability target of the power electronic transformer (system mean time between failures ≥ 10,000 hours) and the historical failure rate data of each module.

4. The modular redundancy design method according to claim 1, characterized in that: In the fault detection step, a multi-sensor data fusion algorithm is used to fuse and analyze the voltage, current, and temperature sensor data through DS evidence theory, thereby improving the fault diagnosis accuracy to more than 99%.

5. The modular redundancy design method according to claim 1, characterized in that: In the redundant switching step, the pre-charging control adopts a ramp voltage control strategy, which controls the charging current of the capacitor on the input side of the redundant module to control the charging time within 8ms to avoid impact on the system.

6. The modular redundancy design method according to claim 1, characterized in that: The high-frequency transformer unit adopts planar magnetic integration technology and integrates multiple high-frequency transformer cores, thereby reducing the transformer volume by 40% and increasing the power density by 30%.

7. The modular redundancy design method according to claim 1, characterized in that: The system control center also integrates an adaptive power control algorithm to dynamically adjust the duty cycle of each power conversion module according to system load changes and module operating status, so that the overall system efficiency is always maintained above 95%.

8. A power electronic transformer implementing the modular redundant design method, characterized in that: include: Multiple power conversion modules with the same structure and function; Redundant module group, including redundant input filter module, redundant power conversion module and redundant output rectifier module; Fault detection and isolation unit, including sensors, monitoring circuits, and fast solid-state relays distributed in each module, used to monitor module status in real time and isolate faulty modules; The system control center uses a controller architecture that combines FPGA and DSP to receive fault signals, control redundant module switching, and adjust power distribution strategies; The communication network uses high-precision clock synchronous Ethernet based on IEEE 1588 to realize data interaction and control command transmission between modules; The health management module has a built-in state-life prediction model and data analysis algorithm to predict module life and generate maintenance plans.

9. The power electronic transformer according to claim 8, characterized in that: Each of the power conversion modules comprises: Input filter unit, used for filtering the input AC power; Full-bridge or multi-level conversion circuit unit, converting the filtered AC power into high-frequency AC power; High-frequency transformer unit to achieve electrical isolation and voltage level conversion; The output rectifier unit converts high-frequency AC power into stable DC power or AC power of a specific frequency.

Images