A method for adaptive optimization of multi-port energy routing under wide bandgap devices

Abstract

The application provides a wide band gap device lower SST multi-port energy routing adaptive optimization method, comprising: according to the change characteristics of the device junction temperature distribution, the material defect density of the wide band gap device and the multi-port coupling effect of the solid state transformer are jointly analyzed, the electric field abnormal area when the overload protection mechanism is triggered is judged through the real-time comparison of the electric field stress threshold; according to the distribution of the insufficient heat management performance area, combining the electromagnetic interference level of the solid state transformer and the carrier mobility of the wide band gap device, the distributed electric field balancing technology is used to dynamically adjust the port voltage stability, and the corrected power conversion efficiency distribution is obtained; through the corrected power conversion efficiency distribution, the interaction characteristics between the breakdown electric field strength of the wide band gap device under high stress and the dynamic load response of the solid state transformer are analyzed, and the voltage imbalance risk point caused by the multi-port coupling effect is judged by using the real-time data fusion technology.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to an adaptive optimization method for SST multiport energy routing under wide bandgap devices.

[0002] Terminology Definition

[0003] 1. Multi-port coupling effect: refers to the physical phenomenon in which the voltage and current characteristics of different ports of a solid-state transformer (SST) are affected by the interaction of electromagnetic fields, thermal fields or electric fields.

[0004] 2. Coupling Influence Region: This refers to the spatial overlap between the region with high material defect density and the SST multi-port strong coupling region. Electric field stress concentration is prone to occur in this region.

[0005] 3. Voltage imbalance risk point: refers to the specific location or time when the port voltage deviation exceeds the rated value ±5% under multi-port coupling;

[0006] 4. Distributed electric field equalization technology: refers to the control technology that achieves uniform electric field distribution in each region through dynamic voltage compensation based on port load weight and electric field anomaly degree;

[0007] 5. Real-time data fusion technology: refers to the technology of integrating and analyzing multi-source sensor data (temperature, voltage, current, etc.) after noise filtering and weight allocation, using specific algorithms.

[0008] 6. Wide bandgap device parameter database: This refers to a structured database that stores the physical parameters (bandgap width, breakdown electric field strength, carrier mobility, etc.) of wide bandgap devices such as SiC and GaN. The data is collected from device manufacturer datasheets and updated through experimental calibration. Background Technology

[0009] As a core piece of equipment in the next generation of power electronics, solid-state transformers play a crucial role in multi-source and multi-load efficient interconnection and flexible energy distribution in multi-port energy routing systems. Their reliability and energy conversion efficiency directly determine the overall performance of new energy access, microgrid operation, and DC distribution networks.

[0010] In high-voltage, high-power applications, wide-bandgap semiconductor devices have become a key material for achieving high-efficiency, high-power-density conversion.

[0011] While current solutions have made some progress in terms of device performance or system control strategies, they all share a deep-seated problem that is difficult to avoid: when multiple ports are simultaneously subjected to complex power flows, there is a strong interaction between the internal bandgap energy level characteristics of the device and the external high voltage electrical stress. This interaction affects the microscopic efficiency of energy conversion and can trigger avalanche breakdown under extreme conditions, leading to voltage imbalance between ports and uncontrolled propagation of energy runaway.

[0012] Existing design methods often view the material band structure and system-level electrical stress distribution in isolation, making it difficult to capture the coupling nature of the two in real multi-port dynamic operation.

[0013] The bandgap width of a wide bandgap material determines the amount of energy required for an electron to transition from the valence band to the conduction band, and this energy threshold directly determines the device's ability to maintain low conduction loss and low switching loss at high voltage.

[0014] However, when the bandgap is large, while the device can withstand higher electric fields, it is also more prone to electron avalanche multiplication in local areas with excessively high electric fields, which can lead to a sudden surge in current and cause destructive breakdown.

[0015] In a multi-port energy routing system, when a port suddenly becomes overloaded or experiences reverse power flow, it will rapidly change the voltage distribution of adjacent ports, causing the local bandgap electric fields of multiple devices to simultaneously approach or exceed the avalanche critical value. This voltage-electric field-avalanche chain reaction can easily trigger simultaneous failure at multiple points.

[0016] Therefore, how to accurately grasp the influence of the bandgap energy level of wide bandgap devices on conversion efficiency, as well as the triggering mechanism of the bandgap avalanche effect on inter-port voltage imbalance and cascading breakdown during the dynamic operation of multi-port energy routing, has become a key issue determining the long-term stable operation and efficient energy routing capability of the system. Summary of the Invention

[0017] This invention provides an adaptive optimization method for SST multiport energy routing under wide bandgap devices, mainly including:

[0018] Bandgap width and breakdown electric field strength data are obtained from a wide bandgap device parameter database. Simultaneously, port voltage stability and power conversion efficiency information of the solid-state transformer during operation are collected. The local electric field distribution under dynamic load response is calculated using numerical simulation tools to obtain the characteristics of the device junction temperature distribution. Based on these characteristics, a joint analysis is performed on the material defect density of the wide bandgap device and the multi-port coupling effect of the solid-state transformer. Real-time comparison of electric field stress thresholds identifies abnormal electric field regions when the overload protection mechanism is triggered. If the stress value in the abnormal electric field region exceeds the preset electric field stress threshold, the control signal delay data of the solid-state transformer and the dielectric constant data of the wide bandgap device are integrated. The thermal conductivity parameter is corrected through information processing to identify potential areas with insufficient thermal management performance. Based on the distribution of these areas, combined with the electromagnetic interference level of the solid-state transformer and the current carrying capacity of the wide bandgap device, further analysis is conducted. The sub-mobility is used to dynamically adjust the port voltage stability using distributed electric field equalization technology, resulting in a corrected power conversion efficiency distribution. Based on this corrected distribution, the interaction characteristics between the breakdown electric field strength of wide-bandgap devices under high stress and the dynamic load response of the solid-state transformer are analyzed. Real-time data fusion technology is used to identify voltage imbalance risk points caused by multi-port coupling effects. If the voltage imbalance risk points exceed the preset safety range, the local electric field distribution is further optimized using information processing technology by integrating the bandgap width values ​​of the wide-bandgap devices and the overload protection mechanism data of the solid-state transformer, determining the final thermal management performance improvement scheme. Based on this scheme, the electric field stress threshold parameters of the wide-bandgap devices and the control signal delay configuration of the solid-state transformer are updated. Simulation verification technology is used to evaluate the port voltage stability of the overall system, obtaining the voltage distribution optimization results for specific overload scenarios.

[0019] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0020] This invention discloses a comprehensive thermo-electric-field coupling optimization method for high-stress dynamic operation of wide-bandgap devices in solid-state transformers. This method collects real-time data on the bandgap width and breakdown electric field strength parameters of the wide-bandgap devices, as well as the port voltage stability and power conversion efficiency data of the solid-state transformer. It combines this with numerical simulation to calculate the local electric field and junction temperature distribution characteristics under dynamic load, and jointly analyzes the influence of material defect density and multi-port coupling effects on electric field stress. When an abnormal electric field region exceeds the stress threshold during overload, this invention integrates control signal delay and dielectric constant information to correct the thermal conductivity, accurately locating areas with insufficient thermal management. Distributed electric field equalization technology is then used to dynamically adjust voltage stability, significantly improving the corrected power conversion efficiency. Based on this, real-time data fusion is used to assess voltage imbalance risk, and when the risk exceeds the limit, the local electric field distribution is optimized again, ultimately forming a thermal management performance improvement scheme. The electric field stress threshold and control delay configuration are updated to optimize voltage distribution under specific overload scenarios. This invention effectively solves the coupling failure problem of thermal runaway and electric field breakdown of wide-bandgap devices under dynamic high-stress conditions in solid-state transformers, significantly improving system reliability and energy conversion efficiency.

[0021] Through simulation and experimental verification, the specific technical effects of this method are as follows: 1) Under an overload scenario of 1.5 times the rated load, the voltage imbalance rate of the SST port is reduced from 12% in the existing technology to below 3%; 2) The power conversion efficiency is improved by 4.5 to 6 percentage points, and the efficiency is stable at above 97% when running under rated load; 3) The peak junction temperature of the device is reduced by 28 to 35 degrees Celsius, and the incidence of thermal runaway risk is reduced by 80%; 4) The continuous stable operation time of the system is extended by 5000 to 8000 hours, and the probability of multi-point failure caused by multi-port coupling is reduced to below 0.5%. Attached Figure Description

[0022] Figure 1 This is a flowchart of the SST multi-port energy routing adaptive optimization method for wide bandgap devices according to the present invention.

[0023] Figure 2 This is a schematic diagram of the SST multi-port energy routing adaptive optimization method under wide bandgap devices according to the present invention.

[0024] Figure 3 This is a graph showing the junction temperature distribution of the device of the present invention over time.

[0025] Figure 4 This is a comparison chart of electric field stress and power conversion efficiency of the present invention.

[0026] Figure 5 This is a heat map illustrating the risk of multi-port voltage imbalance in this invention.

[0027] Figure 6This is a comparison chart of port voltage stability before and after optimization in this invention.

[0028] Figure 7 This is a schematic diagram of the multi-port topology of the solid-state transformer (SST) of the present invention.

[0029] Figure 8 This is a cross-sectional view of the internal electric field distribution of the wide bandgap device of the present invention.

[0030] Figure 9 This is a schematic diagram of the sensor array layout of the present invention.

[0031] Figure 10 This is a schematic diagram of the thermal management and heat dissipation structure of the present invention. Detailed Implementation

[0032] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0033] like Figure 1As shown, the adaptive optimization method for SST multi-port energy routing under wide bandgap devices provided by this invention includes the following steps: First, step S101 is executed, obtaining the bandgap width value and breakdown electric field strength data from the wide bandgap device parameter database, collecting SST port voltage stability and power conversion efficiency information, and calculating the local electric field distribution under dynamic load through numerical simulation to obtain the device junction temperature distribution change characteristics; then, step S102 is executed, based on the junction temperature distribution change characteristics, jointly analyzing the material defect density and multi-port coupling effect, and judging the abnormal electric field region when overload protection is triggered by real-time comparison of electric field stress threshold; in step S103, if the stress value in the abnormal electric field region exceeds the threshold, the control signal delay data and dielectric constant data are fused to correct the thermal conductivity parameter and determine the region with insufficient thermal management performance; otherwise, the process directly proceeds to step S10. 4. Step S104: Based on the distribution of areas with insufficient thermal management, combined with electromagnetic interference levels and carrier mobility, distributed electric field balancing technology is used to dynamically adjust the port voltage stability, resulting in a corrected power conversion efficiency distribution. Step S105: The interaction characteristics between breakdown electric field strength and dynamic load response are analyzed, and real-time data fusion technology is used to determine the voltage imbalance risk points caused by multi-port coupling effects. In step S106: If the voltage imbalance risk points exceed the safe range, the bandgap width value and overload protection mechanism data are integrated to perform secondary optimization of the local electric field distribution, determining the final thermal management performance improvement scheme; otherwise, proceed directly to step S107. Finally, step S107 is executed to update the electric field stress threshold parameters and control signal delay configuration. The port voltage stability is evaluated through simulation verification, obtaining the voltage distribution optimization results under specific overload scenarios. This process achieves adaptive optimization control through two conditional judgment branches, ensuring that the optimal energy routing configuration can be obtained under different operating conditions.

[0034] like Figure 2As shown, the adaptive optimization method for SST multi-port energy routing under wide bandgap devices of the present invention adopts a four-layer architecture. The data layer includes a wide bandgap device parameter database 10, a numerical simulation module 20, and a data acquisition module 30. The parameter database stores key material parameters such as bandgap width, breakdown electric field strength, carrier mobility, thermal conductivity, and dielectric constant. The numerical simulation module uses the COMSOL finite element method to calculate the local electric field distribution and junction temperature distribution. The data acquisition module acquires port voltage, power conversion efficiency, and temperature sensor data in real time. The analysis layer includes a joint analysis module 40, an electric field stress comparison module 50, and a thermal management correction module 60. The joint analysis module receives material characteristic data and acquired data, performs material defect density analysis and multi-port coupling effect analysis, the electric field stress comparison module performs real-time threshold comparison and abnormal region identification, and the thermal management correction module performs thermal conductivity parameter correction and locates regions with insufficient performance. The optimization layer comprises a distributed electric field equalization module 70, a real-time data fusion module 80, and a secondary optimization module 90. The distributed electric field equalization module dynamically adjusts the port voltage based on abnormal regions and correction parameters. The real-time data fusion module assesses voltage imbalance risks using real-time data, and the secondary optimization module performs fine-grained optimization of the local electric field distribution. The simulation verification module 100 in the verification layer evaluates the port voltage stability of the optimization results and returns the verification results to the distributed electric field equalization module through a feedback adjustment mechanism, forming a closed-loop optimization control. All modules are connected via data streams, enabling a complete adaptive optimization process from data acquisition, analysis and processing, optimization adjustment to verification feedback.

[0035] like Figure 7As shown, the solid-state transformer (SST) of this invention adopts a multi-port topology design. This structure includes an input port 101 for connecting to a 10kV high-voltage DC power supply. The input port 101 is connected to a high-voltage side rectifier module 102, which is responsible for rectifying the input power. The rectified power is transmitted to a high-frequency isolation transformer 103, which is located at the core of the topology, achieving electrical isolation and voltage conversion between the input and output sides. The output side of the transformer 103 is connected to an intermediate DC bus 104, which serves as a hub for energy collection and distribution, supplying power to the three output ports. Output port 1, labeled 105, provides a 400V DC output for connecting a photovoltaic inverter; output port 2, labeled 106, provides a 380V output for connecting an energy storage battery; and output port 3, labeled 107, provides a 220V output for connecting a residential load. Multiple SiC MOSFET switching devices 108 are distributed between the ports and modules to achieve high-frequency switching control and power conversion. The control unit 109 is located at the top of the topology and is connected to each power module via control signal lines to achieve coordinated control of the entire system. The protection circuit 110 is located at the bottom of the topology and monitors the operating status of each module via protection signal lines, providing overvoltage, overcurrent, and short-circuit protection functions in abnormal situations. In the diagram, solid arrows indicate the direction of power flow, dashed lines represent control signals, and dotted lines represent protection signals.

[0036] like Figure 8 As shown, the wide bandgap semiconductor device used in this invention has a vertical cross-sectional structure. Inside the device housing 201, from top to bottom, are arranged a gate 202, a source 203, a drain 204, a SiC epitaxial layer 205, and a SiC substrate 206. The source 203 is located on the upper left side of the device, the drain 204 is located on the upper right side of the device, and the gate 202 is located in the center between the source 203 and the drain 204. A gate oxide layer is disposed below the gate 202.

[0037] A PN junction region 207 is formed in the SiC epitaxial layer 205. This region is a concentrated electric field area, located directly below the source 203 and drain 204. When the device is operating, the electric field is non-uniformly distributed inside the device. A high electric field region 208 is located inside and near the PN junction region 207, with an electric field strength of 2.5-3.0 MV / cm, represented by dense equipotential lines; a medium electric field region 209 is located in the channel region below the gate 202, with an electric field strength of 1.5-2.5 MV / cm, represented by medium-density equipotential lines; and a low electric field region 210 is located at the edge of the device and in the region far from the PN junction, with an electric field strength below 1.5 MV / cm, represented by sparse equipotential lines.

[0038] The arrows in the figure indicate the direction of the electric field. The main electric field direction is vertically downwards from the electrode to the substrate, while in the channel region it is horizontal, pointing from the drain to the source. The electric field strength scale 211 is located in the lower right corner of the figure, using a grayscale gradient to represent the range of electric field strength from 0 to 3.0 MV / cm. This electric field distribution characteristic directly reflects the high critical breakdown electric field strength of wide-bandgap semiconductor materials, enabling the device to operate stably under high voltage conditions.

[0039] like Figure 9 As shown, the sensor array layout of this invention includes an SST main body frame 301 and various types of sensors deployed thereon. The SST main body 301 is functionally divided into four areas: an input stage, a high-frequency transformer area, an isolation stage, and an output stage. Temperature sensors are marked with circular symbols and include: temperature sensor T1 302 located near the input port, temperature sensor T2 303 located in the core of the high-frequency transformer, temperature sensor T3 304 located near output port 1, temperature sensor T4 305 located near output port 2, and temperature sensor T5 306 located near output port 3. Voltage sensors are marked with square symbols and include: voltage sensor V1 307 at the input port, voltage sensor V2 308 at output port 1, voltage sensor V3 309 at output port 2, and voltage sensor V4 310 at output port 3. Current sensors are marked with triangular symbols, and current sensors I1-I4 311 are deployed at each port location to monitor current changes at each port. All sensors are connected to the data acquisition bus 312 via the data connection lines shown by the dashed lines. The data acquisition bus 312 aggregates the multi-source sensor data and transmits it to the data fusion processing unit 313 for unified processing. This layout scheme achieves comprehensive monitoring coverage of key locations in the SST, providing complete data support for subsequent fault diagnosis and condition assessment.

[0040] like Figure 10As shown, the thermal management heat dissipation structure of this invention adopts a multi-level heat conduction design, comprising, from top to bottom: a SiC MOSFET device chip 401, a thermal interface material 406, a thermally conductive insulating layer (DBC substrate) 402, a copper heat sink base 403, and a heat sink fin array 404. The SiC MOSFET device chip 401 serves as the primary heat source, and the heat it generates is conducted through the thermal interface material 406 (approximately 0.2 mm thick) to the DBC substrate 402 (0.5 mm thick). The DBC substrate 402 employs a ceramic-coated copper structure, providing excellent thermal conductivity while ensuring electrical insulation. Heat continues to be transferred through the second thermal interface material to the copper heat sink base 403 (3 mm thick). The copper heat sink base 403 has high thermal conductivity, allowing for rapid lateral heat diffusion. A heat pipe 407 is embedded within the heat sink base, and the heat pipe 407 utilizes the working fluid phase change principle to efficiently transfer heat from the heat source area to the heat sink fin array 404. The heat sink fin array 404 comprises multiple evenly distributed fins (2.5mm spacing, 25mm height) to increase the heat dissipation area. Locally reinforced heat sinks 410 are positioned below areas of concentrated hotspots to enhance heat dissipation in areas with insufficient thermal management. A cooling fan 405 is located below the heat sink fin array 404, generating upward forced convection airflow (direction shown as 409) to remove heat from the fin surface. Temperature monitoring points 408 are distributed at key locations such as the device surface, copper base plate, and heat sink base to monitor the temperature distribution at each level of the thermal management system in real time. Heat sink fixing bolts 411 are used to securely install the entire heat dissipation structure to the system housing.

[0041] Specifically, this embodiment of the SST multi-port energy routing adaptive optimization method under wide bandgap devices may include:

[0042] S101. Obtain the bandgap width value and breakdown electric field strength data from the wide bandgap device parameter database. At the same time, collect the port voltage stability and power conversion efficiency information of the solid-state transformer during operation. Calculate the local electric field distribution under dynamic load response using numerical simulation tools to obtain the variation characteristics of the device junction temperature distribution.

[0043] The bandgap width and breakdown electric field strength data are obtained from the wide bandgap device parameter database. The current operating state of the solid-state transformer is determined based on the collected port voltage and power conversion efficiency data. The local electric field distribution under the current dynamic load is calculated using finite element numerical simulation. The power loss density distribution of each region of the device is obtained through the electric field distribution results. The instantaneous value of the junction temperature distribution is calculated based on the power loss density distribution. If the instantaneous value of the junction temperature distribution exceeds the preset safety threshold, the current operating point is marked as a high-risk state. The instantaneous value of the junction temperature distribution is combined with the junction temperature values ​​of multiple previous time points to obtain the trend of junction temperature change over time.

[0044] like Figure 3The figure shows a comparison of the junction temperature changes over time. The horizontal axis represents time, ranging from 0 to 200 milliseconds; the vertical axis represents junction temperature, ranging from 100 to 200 degrees Celsius. The figure includes three curves: the junction temperature change curve before optimization (solid line), the junction temperature change curve after optimization (dashed line), and the safety threshold reference line (dotted line, 180 degrees Celsius). At the start of the overload (t=50ms), both curves rise from an initial temperature of 100 degrees Celsius. The curve before optimization shows a faster temperature rise rate during the overload, eventually reaching a peak of 185 degrees Celsius, exceeding the safety threshold of 180 degrees Celsius and entering the over-threshold region marked in gray in the figure. The curve after optimization, due to the thermal management scheme of this invention, effectively controls the temperature rise rate, with a peak temperature of only 165 degrees Celsius, remaining below the safety threshold throughout. After the end of the overload (t=150ms), both curves show a downward trend, but the curve before optimization still maintains a higher temperature level. The comparison results show that the optimized solution of the present invention can effectively reduce the peak junction temperature of the device under overload conditions by about 20 degrees Celsius, ensuring that the device operates within a safe temperature range and significantly improving the reliability and service life of the system.

[0045] In one possible implementation, the construction and updating rules for the wide bandgap device parameter database are as follows: First, the stored parameter types include bandgap width (unit: electron volts), breakdown electric field strength (unit: megavolts per centimeter), carrier mobility (unit: square centimeters per volt per second), thermal diffusivity (unit: square meters per second), dielectric constant, thermal conductivity (unit: watts per meter per Kelvin), and carrier lifetime (unit: microseconds); second, the data source mainly consists of raw data from datasheets provided by device manufacturers, calibrated through laboratory testing using a laser flash thermal conductivity meter and a semiconductor parameter analyzer; third, the update mechanism involves updating the database every 6 months, synchronously adding new device model data, and correcting the aging parameters of existing devices; fourth, the access method is through a cloud database interface call, where the system automatically returns the corresponding complete parameter set after inputting the device model (e.g., 4H-SiC MOSFET, GaN HEMT).

[0046] In one possible implementation, obtaining bandgap width and breakdown electric field strength data from a wide bandgap device parameter database can serve as a fundamental step in the optimization design of solid-state transformers.

[0047] For example, suppose we use a database of wide bandgap devices based on SiC materials. This database stores the physical parameters of various devices, where the bandgap width is typically represented by the energy difference required for an electron to transition from the valence band to the conduction band. For SiC devices, the bandgap width is approximately 3.2 eV, which is much higher than the 1.1 eV of traditional silicon devices, allowing the devices to operate at higher temperatures and voltages. The breakdown electric field strength refers to the critical value at which a material breaks down under the influence of an electric field. For SiC devices, this value can reach 3 MV / cm. By querying the database, this data can be extracted for subsequent electric field simulations to ensure the transformer's withstand voltage capability. In practical applications, this step involves connecting to a cloud database interface, inputting the device model such as 4H-SiC MOSFET, and automatically retrieving data to provide parameter support for evaluating the transformer's operating status.

[0048] For example, when determining the current operating status of a solid-state transformer based on the collected port voltage and power conversion efficiency data, a real-time monitoring system can be used to collect the data. Assuming the transformer's input port voltage is 1000 V and the output is 800 V, a power conversion efficiency calculated as output power divided by input power reaching 98% indicates that the transformer is operating at high efficiency. If the efficiency is below 95%, it may indicate overload or a fault.

[0049] Specifically.

[0050] In one possible implementation, voltage data is collected every second by an embedded sensor, and the results are used to determine the system's state (e.g., normal, light load, or heavy load) using the efficiency formula η = P_out / P_in. P_out is calculated by multiplying the current and voltage. This allows for real-time determination of the system's state, providing input conditions for subsequent simulations. This determination process helps identify potential problems early and improves system reliability.

[0051] In one possible implementation, the key numerical analysis method is to use finite element numerical simulation to calculate the local electric field distribution under the current dynamic load.

[0052] For example, using COMSOL Multiphysics software to build a device model, the previously acquired bandgap and breakdown electric field strength are used as material property inputs. Then, a dynamic load is defined, such as a sudden change in load current from 10 A to 50 A. The simulation calculates the electric field distribution by solving the Poisson equation and the continuity equation. The results show that the electric field strength in the edge region of the device may reach 2.5 MV / cm, close to the breakdown threshold. This simulation process includes mesh generation, boundary condition setting, and iterative solution, and typically requires high-performance computing resources to handle complex geometries, thereby obtaining accurate electric field maps and laying the foundation for power loss analysis.

[0053] For example, when obtaining the power loss density distribution of each region of the device through the electric field distribution results, calculations can be performed based on the Joule heating effect.

[0054] Specifically, the power loss density P = σE², where σ is the electrical conductivity and E is the local electric field strength. Substituting the electric field data obtained from simulation into the formula, a distribution map can be mapped. For example, the loss density in the central region of the device is 10 W / cm³, while it is 50 W / cm³ at the edge, reflecting thermal non-uniformity. In practical applications, this distribution helps to identify hot spots and avoid device overheating failure.

[0055] In one possible implementation, the finite element numerical simulation tool uses COMSOL Multiphysics, with the following specific simulation parameters: First, mesh generation: the mesh density in critical areas such as the PN junction of the device is controlled within 2 micrometers, and the mesh density in non-critical areas is controlled within 10 micrometers to ensure the calculation accuracy of critical areas; Second, boundary conditions: the electric field boundary adopts a fixed rated voltage constraint, and the temperature boundary adopts a convection heat transfer constraint, with the heat transfer coefficient set at 200 W / m² / Kelvin; Third, the solution algorithm: the Newton-Raphson iterative algorithm is used, and iteration stops when the electric field strength calculation error does not exceed 1%; Fourth, overload condition settings: the overload multiple ranges from 1.2 times to 2 times the rated load, the overload duration is from 50 milliseconds to 200 milliseconds, and the load change slope is from 5 A / s to 10 A / s.

[0056] In one possible implementation, calculating the instantaneous value of the junction temperature distribution based on the power loss density distribution involves the application of a heat conduction model.

[0057] For example, using the transient thermal equation ∂T / ∂t = α∇²T + Q / ρc, where Q is the loss density and α is the thermal diffusivity, the junction temperature can be obtained through the finite element method, such as an instantaneous junction temperature of 150°C. This calculation takes into account the material's heat capacity and heat dissipation conditions, providing a real-time temperature map.

[0058] For example, if the instantaneous value of the junction temperature distribution exceeds a preset safety threshold such as 200°C, the current operating point is marked as a high-risk state. This step is achieved through threshold comparison and triggers an alarm to protect the equipment.

[0059] In one possible implementation, the instantaneous value of the junction temperature distribution is combined with the junction temperature values ​​at multiple previous time points to obtain the trend of junction temperature change over time. For example, values ​​from the past 10 time points are collected, such as 100°C at t=0 and 120°C at t=1s. The fitted curve shows an upward trend, which can predict potential faults, enable preventive maintenance, and thus improve the overall stability and lifespan of the solid-state transformer.

[0060] In one possible implementation, the preset safety threshold is set according to the following rules: First, the junction temperature safety threshold is based on the highest operating junction temperature indicated in the wide bandgap device manufacturer's datasheet, reduced by 20 degrees Celsius. For example, when the highest operating junction temperature of a SiC device is 200 degrees Celsius, the junction temperature safety threshold is set to 180 degrees Celsius. Second, the electric field stress threshold is set to 80% of the device's breakdown electric field strength. For example, when the breakdown electric field strength of a SiC device is 3 MV / cm, the electric field stress threshold is set to 2.4 MV / cm. Third, the preset range for control signal delay is set to no more than 50 microseconds. This value is determined based on experimental data of the dynamic response time of the SST control loop to ensure that the protection action can be triggered in a timely manner without lag when an overload occurs.

[0061] S102. Based on the variation characteristics of the junction temperature distribution of the device, a joint analysis is performed on the material defect density of the wide bandgap device and the multi-port coupling effect of the solid transformer. By comparing the electric field stress threshold in real time, the abnormal electric field region when the overload protection mechanism is triggered is determined.

[0062] Obtain the junction temperature distribution data sequence, calculate the change trend sequence based on the junction temperature distribution data sequence, and use the change trend sequence to perform segmented correlation analysis on the material defect density to obtain the defect density distribution map. Perform spatial location matching between the defect density distribution map and the multi-port coupling strength to obtain the coupling influence region. Extract the electric field stress value sequence at the corresponding location through the coupling influence region, and compare the electric field stress value sequence with a preset stress threshold in real time. If a point in the electric field stress value sequence exceeds the stress threshold, it is determined as an abnormal region and the occurrence time is recorded. When the overload protection trigger signal arrives, extract the location of the abnormal region closest to the trigger time to obtain the electric field abnormal distribution region corresponding to the instant of overload protection action.

[0063] In one possible implementation, the piecewise correlation analysis of material defect density employs a linear regression model. The core of the model is establishing a linear correspondence between the slope of the junction temperature change trend and the material defect density. For SiC devices, when the slope of the junction temperature change trend does not exceed 0.5 degrees Celsius per second, the first set of regression coefficients (with a slope coefficient of 5 × 10¹) is used. 5 The intercept coefficient is 1 × 10¹ per cubic centimeter per degree Celsius per second. 5 per cubic centimeter); when the slope of the junction temperature change trend exceeds 0.5 degrees Celsius per second, the second set of regression coefficients (slope coefficient of 8 × 10¹) is used. 5 The intercept coefficient is 2 × 10¹ per cubic centimeter per degree Celsius per second. 5 (per cubic centimeter), ensuring the accuracy of defect density analysis under different thermal stress scenarios through segmented calculations.

[0064] For example, in a solid-state transformer system, the first step is to extract key information from an established junction temperature distribution data sequence. This data sequence is typically acquired through real-time monitoring sensors, covering temperature distribution at multiple time points. Acquiring the junction temperature distribution data sequence involves converting sensor signals into a digital sequence. For instance, a thermistor array might be placed at key locations within the device, collecting temperature readings at fixed intervals, such as one second. Noise is then removed using data filtering algorithms to ensure the sequence's accuracy. This sequence records not only the average junction temperature but also the localized temperature values ​​of different regions within the device, such as the gate and drain regions, thus providing a foundation for subsequent analysis.

[0065] In one possible implementation, when calculating the trend sequence based on the junction temperature distribution data sequence, a difference method can be used to quantify the temperature change over time.

[0066] Specifically, the junction temperature values ​​at adjacent time points in the sequence are subtracted to obtain a difference sequence. These differences are then averaged and smoothed using a sliding window to capture trends such as rising, stable, or falling. This calculation helps identify temperature fluctuation patterns; for example, when the load suddenly increases, the trend sequence may show a sharp rise in peak values, reflecting the dynamic response of the device to thermal stress. This trend sequence can then be further correlated with potential issues in material properties.

[0067] For example, the process of performing segmented correlation analysis on the defect density of a material using a trend sequence to obtain a defect density distribution map requires first dividing the trend sequence into several segments, such as based on trend inflection points, and then matching known defect models within each segment. Defect density refers to the distribution concentration of materials such as crystal dislocations or impurities, which affect heat conduction efficiency.

[0068] Specifically, statistical correlation analysis is used to correlate the fluctuations of the trend sequence with the defect density parameter. For example, in the high-temperature trend segment, the defect density may be high, leading to the formation of hot spots. This ultimately generates a two-dimensional distribution map, in which color gradients represent areas of high and low density. This chart visually demonstrates how defects affect the overall thermal distribution.

[0069] In one possible implementation, when the spatial location correspondence between the defect density distribution map and the multi-port coupling strength is matched to obtain the coupling influence area, the multi-port coupling strength refers to the interaction strength of the electromagnetic field or thermal field between different ports in the transformer.

[0070] For example, by overlaying the distribution map onto a 3D model of the transformer, coordinate matching can be used to identify the overlapping areas between high-density defect regions and strongly coupled ports. These areas are often where thermal stress is concentrated, thus defining the boundaries of the affected regions. This matching improves the accuracy of system diagnosis and avoids blind inspections.

[0071] For example, by extracting the electric field stress value sequence at the corresponding location through coupling influence region, the electric field intensity value can be sampled from the region using finite element mesh to form a time series.

[0072] Specifically, an extraction path is defined in the simulation software, and data is collected along the region boundary, such as one sampling point every millisecond. The peak value and average value are recorded in the sequence, which helps to track stress evolution.

[0073] In one possible implementation, the electric field stress value sequence is compared with a preset stress threshold in real time. If a point in the sequence exceeds the threshold, it is identified as an abnormal region and the time of occurrence is recorded. The preset threshold is set based on the tolerance limit of the device material; for example, the threshold for silicon-based devices is 10^6 V / m. The comparison process uses a threshold comparison algorithm to scan the sequence point by point. Once the value exceeds the threshold, the location and timestamp are marked. This real-time monitoring ensures timely warnings.

[0074] For example, when the overload protection trigger signal arrives, the location of the abnormal region closest to the trigger time is extracted to obtain the electric field abnormal distribution region corresponding to the instant the overload protection operates.

[0075] Specifically, by searching for the entry closest to the trigger time from the recorded abnormal moments, tracing back the sequence data, and reconstructing the distribution map, the expansion of the abnormal region is displayed. This extraction helps in post-event analysis of the overload causes and optimization of the protection mechanism design, thereby improving the reliability and efficiency of solid-state transformers.

[0076] S103. If the stress value in the abnormal electric field region exceeds the preset electric field stress threshold, the control signal delay data of the solid-state transformer and the dielectric constant data of the wide bandgap device are integrated, and the thermal conductivity parameter is corrected through the information processing stage to determine the potential areas with insufficient thermal management performance.

[0077] After correcting the thermal conductivity parameters through an information processing workflow, data sets related to electric field stress and thermal management performance are extracted from abnormal areas. A pre-established analysis model is used to initially classify these data sets, revealing the distribution of potential deficiencies. Based on this classification, control signals and delay data related to the solid-state transformer are acquired. Time-series analysis is performed on the delay data to determine if signal delay anomalies exist. If the delay data exceeds a preset range, it is marked as a high-risk area, and its specific location is determined. Using the location of the high-risk area, the dielectric constant data of the wide-bandgap device is acquired. The dielectric constant data is correlated with the thermal conductivity parameters. If the deviation between the dielectric constant data and the thermal conductivity parameters exceeds a preset standard, it is identified as a critical point of insufficient thermal management performance, and its distribution range is determined. Based on the distribution range of the critical points, the correlation information between electric field stress data and thermal management performance is extracted. The thermal conductivity parameters are then corrected a second time through information processing to determine the corrected thermal conductivity parameter values. By using the corrected thermal conductivity parameters, predictive analysis of thermal management performance is performed on areas with insufficient thermal management. If the prediction results show that the thermal management performance is lower than the preset level, a corresponding performance deficiency marker is generated, resulting in a detailed list of marked areas. Based on this detailed list, solid-state transformer control signals and dielectric constant data of wide-bandgap devices are integrated, and targeted thermal management optimization strategies are generated through an information processing flow, determining the execution priority of these strategies. Based on the execution priority of the optimization strategies, data updates are performed on high-risk areas and key points. Thermal management performance is continuously monitored through information processing, resulting in updated performance status data.

[0078] In one possible implementation, the core logic of the thermal conductivity correction is as follows: based on the initial thermal conductivity in the device parameter database, correction is made by combining the measured deviation of the dielectric constant. The larger the deviation of the dielectric constant, the greater the thermal conductivity correction. The dielectric constant influence coefficient is between 0.02 and 0.05, and this coefficient is determined through experimental calibration. The correlation standard between dielectric constant and thermal conductivity is as follows: when the deviation between the measured value and the standard value of the dielectric constant does not exceed 5%, the thermal management status is judged to be normal; when the deviation exceeds 5%, the location is judged to be a critical point of insufficient thermal management performance.

[0079] For example, after correcting the thermal conductivity parameters through information processing, when extracting data sets related to electric field stress and thermal management performance from the abnormal region, a specific solid-state transformer application scenario can be considered, where the abnormal region is the electric field anomaly point obtained through previous junction temperature distribution analysis.

[0080] Specifically, assuming that wide-bandgap devices such as SiC MOSFETs generate hot spots under high-power operation, the information processing flow first uses finite element simulation software to correct the thermal conductivity parameters. For example, the initial thermal conductivity value is adjusted from 150 W / m·K to a more accurate 148 W / m·K. Based on real-time temperature feedback, electric field stress data such as peak intensity of 5 MV / cm and thermal management performance data such as cooling efficiency of 85% are extracted. These data sets form a matrix for subsequent classification.

[0081] In one possible implementation, the process of using a pre-established analytical model to initially classify the data set and obtain the distribution of potential deficient regions can be achieved by a machine learning classifier such as a support vector machine model, which is pre-trained on a historical dataset, including electric field stress threshold and thermal conductivity related samples.

[0082] For example, in a solid-state transformer multi-port system, after the model inputs a set of data, the regions are classified into three categories: normal, slightly deficient, and severely deficient. The output deficient region distribution map shows that 20% of the area near the transformer output port is marked as potentially deficient, which helps to identify thermal management bottlenecks.

[0083] For example, when acquiring control signals and delay data related to solid-state transformers based on the distribution of insufficient regions obtained from classification, the process of performing time-series analysis on the delay data to determine whether there are signal delay anomalies can be executed in actual business by acquiring PWM control signals and response delay values.

[0084] Specifically, if the delay data exceeds the preset range, such as more than 50 μs, it is marked as a high-risk area. For example, in the transformer overload test, the input port signal delay was found to be 60 μs, which led to the area being marked as high-risk and its specific location being determined, such as the coordinates near the device gate (x=10mm, y=15mm). This is connected to the subsequent dielectric constant analysis to ensure system stability.

[0085] In one possible implementation, dielectric constant data of wide bandgap devices is obtained by identifying the specific location of high-risk areas. The process involves correlating and comparing the dielectric constant data with thermal conductivity parameters. This involves measuring the dielectric constant value of a device such as GaNHEMT, which is 9.5, and comparing it with the corrected thermal conductivity of 148 W / m·K. If the deviation exceeds a preset standard, such as 5%, it is determined to be a key point of insufficient thermal management performance.

[0086] For example, in business scenarios, this comparison reveals that the distribution range of key points covers a 5mm radius area at the center of the device, thus providing a basis for secondary calibration.

[0087] For example, based on the distribution range of key points, the correlation information between electric field stress data and thermal management performance can be extracted. The thermal conductivity parameter can be corrected twice through the information processing stage. When determining the corrected thermal conductivity parameter value, an iterative algorithm can be used to adjust the parameter.

[0088] For example, the correlation information shows that electric field stress is negatively correlated with thermal conductivity. After secondary correction, the parameter value stabilizes at 147 W / m·K, which improves thermal dissipation efficiency in high-frequency solid-state transformers.

[0089] In one possible implementation, the process involves predicting and analyzing the thermal management performance of the deficient areas using the corrected thermal conductivity parameter value. If the prediction results show that the thermal management performance is lower than the preset level, a corresponding performance deficiency label is generated, and a detailed list of the labeled areas is obtained.

[0090] For example, predicting that performance will drop below 70% using Monte Carlo simulations generates labels such as "high risk of hot spots" and lists them including location, severity, and recommended actions, which lays the foundation for optimization strategies.

[0091] For example, based on a detailed list of identified areas, solid-state transformer control signals and dielectric constant data of wide-bandgap devices are integrated, and targeted thermal management optimization strategies are generated through information processing. When determining the execution priority of optimization strategies, strategies that increase cooling flow rate in high-risk areas can be prioritized and set to priority level 1.

[0092] In one possible implementation, the process involves optimizing the execution priority of the strategy, updating data for high-risk areas and key points, and continuously monitoring thermal management performance through information processing to obtain updated performance status data. For example, in real-time monitoring, the updated performance status data shows that the efficiency has increased to 90%, which ensures the long-term reliability of the system.

[0093] S104. Based on the distribution of areas with insufficient thermal management performance, and combined with the electromagnetic interference level of solid-state transformers and the carrier mobility of wide-bandgap devices, distributed electric field balancing technology is used to dynamically adjust the port voltage stability, thereby obtaining the corrected power conversion efficiency distribution.

[0094] The spatial location information of the thermally inadequate region is obtained. Based on the spatial location information of the thermally inadequate region, the distribution range of devices with high electromagnetic interference intensity is determined. The carrier mobility data of wide bandgap devices is used to screen the distribution range of devices with high electromagnetic interference intensity to obtain the key influence range. For the key influence range, distributed electric field equalization technology is applied to calculate the required electric field adjustment intensity of each port. The port voltage is dynamically adjusted according to the required electric field adjustment intensity of each port to obtain the adjusted voltage value sequence. The power conversion efficiency at the corresponding time is calculated based on the adjusted voltage value sequence to obtain the efficiency sequence. The efficiency sequence is reorganized according to spatial location to obtain the corrected spatial distribution of power conversion efficiency.

[0095] like Figure 4 The figure shows the comparison between electric field stress and power conversion efficiency at each port of a three-port charging system. The left vertical axis represents electric field stress (unit: MV / cm), and the right vertical axis represents power conversion efficiency (unit: %). The diagonally filled bar chart represents the electric field stress value at each port, where the electric field stress at port 1 is 2.1 MV / cm, port 2 is 2.6 MV / cm, and port 3 is 2.3 MV / cm. The horizontal dashed line represents the electric field stress threshold of 2.4 MV / cm. The electric field stress at port 2 exceeds this threshold, requiring targeted stress management in practical applications. The gray-filled bar chart represents the power conversion efficiency of each port before optimization, which is 94% for port 1, 91% for port 2, and 92% for port 3. The dotted-filled bar chart represents the power conversion efficiency of each port after optimization, which is improved to 97.5% for port 1, 97.2% for port 2, and 96.8% for port 3. The comparison shows that after processing with the optimization method described in this invention, the power conversion efficiency of all three ports is significantly improved, with an average improvement of about 4.4 percentage points, which verifies the effectiveness of the technical solution of this invention in improving the power conversion efficiency of multi-port charging systems.

[0096] For example, in a solid-state transformer system, it is first necessary to obtain spatial location information of areas with insufficient thermal management, which can be achieved by integrating a sensor network.

[0097] Specifically, these sensors are deployed in key parts of the transformer, such as the core winding and heat dissipation module, to locate areas with low thermal conductivity parameters by monitoring temperature distribution data in real time.

[0098] For example, in a typical power conversion device, if the temperature in a certain area is consistently above 85 degrees Celsius while the surrounding areas are normal, that location is marked as having insufficient thermal management. Three-dimensional spatial location information is generated through coordinate mapping technology, thereby providing basic data support for subsequent analysis.

[0099] In one embodiment, the spatial location information of these thermally inadequate regions can be used to determine the distribution range of devices with high current electromagnetic interference intensity. This involves combining location data with electromagnetic field scanning results.

[0100] For example, electromagnetic compatibility (EMC) testing equipment can be used to scan the intensity of interference signals inside a transformer. If the peak interference level in a certain range exceeds a preset threshold, such as 50 dB, it is designated as a high-intensity distribution range. In this way, thermal and electromagnetic issues can be linked together to form a comprehensive view of the device distribution.

[0101] For example, carrier mobility data from wide-bandgap devices can be used to screen for device distribution ranges with high electromagnetic interference (EMI) intensity, thus identifying critical influence ranges. Here, carrier mobility data for wide-bandgap devices such as SiC or GaN materials is typically obtained through laboratory measurements; for instance, a mobility value above 1000 cm² / Vs indicates low EMI sensitivity. During the screening process, this data is compared with the EMI intensity within each range. If a sub-range has device mobility below the average and high EMI intensity, it is identified as a critical influence range, thus focusing on the areas that truly require optimization.

[0102] In one embodiment, distributed electric field equalization technology is applied to calculate the required electric field modulation intensity at each port for the critical influence zone. This technology is based on the principle of uniform electric field distribution and is calculated using simulation software such as finite element analysis tools.

[0103] For example, the interval is divided into multiple ports, and the electric field intensity at each port is adjusted using an iterative algorithm to ensure that the overall field strength deviation is less than 10%. The calculation process includes inputting initial electric field data, applying an equilibrium model for iterative optimization, and finally outputting the specific adjustment intensity value for each port.

[0104] For example, the port voltage can be dynamically adjusted by adjusting the required electric field intensity at each port to obtain a sequence of adjusted voltage values.

[0105] Specifically, a feedback control loop is used to adjust the voltage in real time.

[0106] For example, if the intensity adjustment indicator requires a 5V increase in voltage, the controller applies the change gradually, recording the voltage value at each time point to form a sequence such as a gradient list from 400V to 405V. This helps stabilize the device's operating state and avoids heat buildup caused by overvoltage.

[0107] In one embodiment, the specific control strategy of the distributed electric field equalization technology is as follows: the priority of port voltage adjustment is determined by two factors: first, the port load weight (i.e., the proportion of the real-time load current of this port to the total load current of all ports); and second, the degree of electric field anomaly (i.e., the proportion by which the actual electric field strength in this area exceeds a preset threshold). The weight of the electric field anomaly accounts for 60%, and the port load weight accounts for 40%. Ports with higher priority are given priority for voltage compensation. Voltage compensation uses proportional-integral control logic, with the proportional coefficient ranging from 0.02 to 0.05 and the integral coefficient ranging from 0.001 to 0.003. These two coefficients are used to adjust the compensation speed and stability, ensuring that the electric field quickly becomes uniform.

[0108] In one embodiment, the power conversion efficiency at corresponding times is calculated based on the adjusted voltage value sequence to obtain an efficiency sequence. Here, the efficiency calculation is based on the input-output power ratio.

[0109] For example, for each voltage value, the input power and output power are measured, and the efficiency = output / input * 100%. The sequence may show a variation from 95% to 97%. In this way, the impact of regulation on performance is quantified.

[0110] For example, by recombining the efficiency sequence according to its spatial location, a corrected spatial distribution of power conversion efficiency can be obtained.

[0111] Specifically, the sequence data is mapped back to the original spatial coordinates.

[0112] For example, by mapping time series data to location information and recombining them to form a distribution map, the efficiency values ​​of different regions can be displayed, such as 97% in the central area and 95% in the edge area. This recombination helps to visualize the overall optimization effect and can be used in business to guide further thermal management and electromagnetic compatibility design, achieving higher system reliability.

[0113] S105. By analyzing the interaction characteristics between the breakdown electric field strength of wide bandgap devices under high stress and the dynamic load response of solid-state transformers through the corrected power conversion efficiency distribution, the voltage imbalance risk points caused by multi-port coupling effect are identified using real-time data fusion technology.

[0114] The corrected power conversion efficiency distribution data is obtained, and the breakdown electric field sequence of the wide bandgap device under high stress conditions is obtained through real-time data fusion processing. The breakdown electric field sequence is time-aligned with the dynamic load response data of the solid-state transformer. The interaction intensity sequence between the two is calculated based on the time alignment result. If the interaction intensity sequence exceeds a preset threshold, the current moment is marked as a potential voltage imbalance moment. For the marked potential voltage imbalance moment, the corresponding multi-port coupling state parameters are extracted. The risk location set is determined by the correspondence between the multi-port coupling state parameters and the voltage imbalance moment. The risk location set is sorted to obtain the final high-risk point sequence.

[0115] like Figure 5 As shown, this invention provides a multi-port voltage imbalance risk heatmap to visually display the interaction intensity and voltage imbalance risk of each port connection area at different time points. The X-axis of the heatmap represents time, ranging from 0-200ms with 20ms intervals; the Y-axis represents the three port connection areas: port 1-port 2, port 2-port 3, and port 1-port 3. The heatmap values ​​represent the interaction intensity, ranging from 0-1.0, and a grayscale color gradient is used to represent the risk level: white areas (0-0.3) represent low risk, gray areas (0.3-0.6) represent medium risk, and black areas (0.6-1.0) represent high risk. A risk threshold of 0.8 is set in the figure; exceeding this value indicates a voltage imbalance risk. The heatmap shows that the port 2-port 3 connection area reaches a peak value of 0.85 at t=100ms, exceeding the risk threshold, indicating a high voltage imbalance risk in this connection area at that time, requiring corresponding equalization control measures. The port 1-port 3 connection area is generally in a low-risk state with low interaction intensity. The port 1-port 2 connection area exhibits medium-risk fluctuation characteristics. This heatmap provides a visual basis for voltage imbalance monitoring and early warning in multi-port systems.

[0116] In one embodiment, corrected power conversion efficiency distribution data is first read from the monitoring system of the solid-state transformer. This data is typically based on a spatial mapping adjusted for previous electromagnetic interference.

[0117] For example, in a multi-port solid-state transformer system, this distribution data may show that efficiency drops to 85% in high-load areas while remaining above 95% in low-load areas. This method of acquisition ensures the basis for the accuracy of subsequent analysis.

[0118] For example, in high-frequency switching applications, real-time data fusion processing can integrate temperature, current, and voltage signals collected by sensors to obtain the breakdown electric field sequence of wide-bandgap devices such as silicon carbide MOSFETs under high stress conditions. The specific process involves weighted averaging and fusing multi-source data. First, noise is filtered out, and then the breakdown electric field value sequence is calculated. For example, each value in the sequence represents the critical electric field strength of the device at the peak stress, such as gradually increasing from 2000V / cm to 5000V / cm, thereby reflecting the change in the device's withstand voltage limit. This fusion helps to capture potential fault precursors in a timely manner.

[0119] In one embodiment, the breakdown electric field sequence is time-aligned with the dynamic load response data of the solid-state transformer.

[0120] For example, a timestamp matching method can be used to align the time points of the breakdown sequence with the load response curve. The load response data may include the output voltage fluctuation curve of the transformer during sudden load changes. An interpolation algorithm is used to ensure that the sequence length is consistent, thereby providing a synchronization basis for subsequent calculations.

[0121] For example, the interaction strength sequence between the two can be calculated based on the temporal alignment results.

[0122] Specifically, the degree of coupling between the breakdown electric field and the load response can be quantified by correlation coefficient analysis. For example, Pearson correlation values ​​can be calculated on aligned data points to form an intensity sequence. If the sequence value is close to 1, it indicates a strong interaction. This calculation reveals the correlation between device breakdown risk and system load.

[0123] In one embodiment, the real-time data fusion processing employs an adaptive weighted fusion algorithm, with the following specific steps: First, Gaussian filtering is applied to the temperature, current, and voltage data collected by the sensors to remove noise, using a sliding window of 5 sampling points to smooth the data during filtering; Second, the reliability weights of each data source are calculated: the weight of temperature data is calculated based on the deviation between the actual temperature and the average temperature, with a higher weight for smaller deviations; the weight of current data is calculated based on the deviation between the actual current and the rated current, with a higher weight for smaller deviations; the weight of voltage data is calculated based on the deviation between the actual voltage and the rated voltage, with a higher weight for smaller deviations, and the sum of the weights of the three types of data is 1; Third, the breakdown electric field strengths derived from each of the three types of data are weighted and averaged according to their corresponding weights to obtain the final breakdown electric field sequence.

[0124] In one embodiment, if the interaction intensity sequence exceeds a preset threshold such as 0.8, the current moment is marked as a potential voltage imbalance moment.

[0125] For example, in a real-world distributed energy system, marking the times when the intensity value exceeds the threshold for three consecutive time points helps to identify unstable nodes in the system early, thereby avoiding overall efficiency loss.

[0126] For example, for the marked potential voltage imbalance moments, the corresponding multi-port coupling state parameters are extracted. The specific process includes pulling the mutual inductance coefficient and phase difference value between ports from the system log. For example, in a three-port transformer, the parameters may show that the coupling coefficient between port 1 and port 2 is 0.95, while that with port 3 is 0.7. This extraction ensures that the parameters reflect the real-time state.

[0127] In one embodiment, the set of risk locations is determined by the correspondence between multi-port coupled state parameters and voltage imbalance moments.

[0128] For example, by establishing a mapping model that associates highly coupled parameters with moments of imbalance, the intersection of ports 2 and 3 can be identified as risk locations. This correspondence, trained based on historical data, can improve the accuracy of risk location and enable more reliable voltage stability control in business operations.

[0129] For example, sorting the set of risky locations yields the final sequence of high-risk points.

[0130] Specifically, the maintenance is prioritized according to the severity of the risk, such as the magnitude of voltage deviation. For example, the priority decreases from 5% to 15%, forming a sequence. This not only optimizes the maintenance priority but also improves the overall reliability of solid-state transformers under high-stress conditions.

[0131] S106. If the voltage imbalance risk point exceeds the preset safety range, the local electric field distribution is optimized again by integrating the bandgap width value of the wide bandgap device and the overload protection mechanism data of the solid transformer, and the final thermal management performance improvement scheme is determined.

[0132] By collecting voltage imbalance data, information processing technology is used to perform preliminary analysis of the collected voltage signals to obtain the distribution characteristics of the voltage imbalance. Based on the distribution characteristics of the voltage imbalance, combined with the bandgap width data of wide bandgap devices, information processing technology is used to perform detailed calculations of the local electric field distribution to determine the key regions of the electric field distribution. If the key regions of the electric field distribution exceed the preset safety range, calibration processing is performed using overload protection data from the solid-state transformer to obtain the response parameters of the overload protection. Based on the response parameters of the overload protection, the parameters of the thermal management module are adjusted to determine the preliminary optimization results of the thermal management performance. If the preliminary optimization results of the thermal management performance do not meet the preset standard, information processing technology is used to perform secondary analysis of the electric field distribution and overload protection data to obtain the direction for improving the thermal management performance. Based on the improvement direction, a pre-established thermal management model is used to simulate and verify the adjusted parameters to determine the final thermal management performance improvement scheme. Based on the final thermal management performance improvement scheme, the system operating parameters are updated in real time to obtain the stable operating state of the system.

[0133] For example, when collecting data related to voltage imbalance, it is first necessary to use a sensor network to monitor the multi-phase voltage signals in the power system in real time. For instance, voltage transformers can be deployed at the input and output ports of solid-state transformers to capture instantaneous voltage fluctuation data. This data includes the amplitude and phase difference of the three-phase voltages. Information processing techniques such as Fourier transform are used to perform spectral analysis on the signals, thereby extracting harmonic components and imbalance factors, and obtaining the distribution characteristics of voltage imbalance, such as the specific distribution pattern of phase A voltage being higher and phase B voltage being lower when the load changes. This preliminary analysis helps to identify potential sources of instability in the system.

[0134] In one possible implementation, based on the obtained voltage imbalance distribution characteristics and combined with the bandgap data of wide-bandgap devices (e.g., the bandgap of SiC devices is approximately 3.2 eV), finite element analysis software can be used to perform detailed calculations of the local electric field distribution. Specifically, the bandgap width is input as a material parameter into the electric field simulation model to calculate the electric field intensity gradient inside the device, identifying critical regions such as high-electric-field regions near the PN junction. If the electric field in this region exceeds the material breakdown threshold, it is marked as a risk point. This step ensures accurate mapping of the electric field distribution, providing a foundation for subsequent safety assessments.

[0135] For example, if the critical area of ​​the electric field distribution exceeds a preset safety range, such as the electric field strength exceeding the safety threshold of 10kV / mm, overload protection data from the solid-state transformer is used for calibration. This data includes the response time and current threshold of the protection circuit. The electric field calculation parameters are adjusted through a comparison algorithm to obtain response parameters such as a protection activation delay of 0.1ms. Thus, the calibrated parameters are more closely aligned with actual operating conditions, avoiding misjudgments.

[0136] In one possible implementation, parameters of the thermal management module are adjusted based on the overload protection response parameters. For example, the cooling fan speed or the thermal conductivity of the heat sink material is modified. An initial assessment is then made to determine if the optimization results, such as a 5-degree Celsius temperature drop, meet the requirements. This adjustment process is implemented through a feedback loop to ensure an initial improvement in thermal management performance and reduce the risk of device overheating.

[0137] For example, if the initial optimization results do not meet the standards, a secondary analysis is performed on the electric field distribution and overload protection data. Machine learning algorithms, such as clustering methods, are used to identify patterns and determine improvement directions, such as enhancing local cooling. This analysis reveals the causes and consequences of heat accumulation, such as voltage imbalance leading to uneven current and resulting in hot spots, allowing for optimization of thermal management in this direction.

[0138] In one possible implementation, by improving the direction, a pre-established thermal management model, such as a finite difference heat conduction model, is used to simulate and verify the adjustment parameters. For example, the temperature distribution under different wind speeds is simulated to determine the final solution, such as adding an auxiliary heat sink, thereby improving the overall thermal stability of the system.

[0139] For example, according to the final solution, the system operating parameters are updated in real time, such as adjusting the transformer switching frequency and monitoring the stable state, such as voltage fluctuations of less than 1%. This not only improves reliability but also achieves more efficient energy conversion in business operations and reduces downtime losses.

[0140] In one possible implementation, the secondary optimization of the local electric field distribution employs an improved particle swarm optimization algorithm. The optimization objective is to minimize the uniformity error of the electric field intensity in each region (uniformity error refers to the sum of squares of the deviations between the electric field intensity in each region and the average electric field intensity, divided by the total number of region divisions). The algorithm parameters are set as follows: the number of particles participating in the optimization is 50, the maximum number of iterations is 100, the inertia weight is linearly decreased from 0.9 to 0.4 to balance global and local search capabilities, and the learning factor is always set to 2 to adjust the speed at which particles approach their own optimal solution and the global optimal solution. When the uniformity error change does not exceed 1% over five consecutive iterations, the optimization is considered converged, and the final electric field distribution parameters are output.

[0141] S107. Based on the final thermal management performance improvement scheme, update the electric field stress threshold parameters of the wide bandgap device and the control signal delay configuration of the solid-state transformer. Use simulation verification technology to evaluate the port voltage stability of the overall system and obtain the voltage distribution optimization results for specific overload scenarios.

[0142] Based on the correlation between thermal management performance and electric field stress threshold, the latest value of improved thermal management performance under the current operating condition is obtained. Based on the latest thermal management performance value, the adjustment amount of the electric field stress threshold parameter for the wide bandgap device is determined. Using the adjusted electric field stress threshold parameter, the threshold setting in the wide bandgap device protection model is updated. Based on the updated wide bandgap device protection model, the matching configuration value of the solid-state transformer control signal delay is obtained. Using the solid-state transformer control signal delay configuration value, the delay parameter of the system's real-time control loop is modified. For the modified delay parameter and protection model, system-level port voltage simulation calculations including overload conditions are performed. Based on the system-level port voltage simulation calculation results, the optimized results of port voltage distribution under specific overload conditions are obtained.

[0143] like Figure 6 As shown, this invention systematically optimizes multi-port voltage stability, and the optimization effect is verified through comparative experiments. The horizontal axis in the figure represents five key performance indicators, including voltage imbalance rate, port 1 voltage deviation, port 2 voltage deviation, port 3 voltage deviation, and continuous operating time improvement factor; the vertical axis represents the corresponding values, in percentage or multiple units. Hollow bars represent performance data before optimization, while filled bars represent performance data after optimization.

[0144] In terms of voltage imbalance rate, the rate was 12% before optimization and decreased to 2.1% after optimization, a reduction of 82%. Regarding voltage deviation at each port, the voltage deviation at port 1 decreased from 5% before optimization to 1.2% after optimization, a reduction of 76%; the voltage deviation at port 2 decreased from 8% before optimization to 1.5% after optimization, a reduction of 81%; and the voltage deviation at port 3 decreased from 6% before optimization to 1.3% after optimization, a reduction of 78%. Furthermore, the continuous operating time of the system increased from 1.0 times the baseline to 1.8 times, an improvement of 80%. The above experimental data demonstrate that the optimization scheme proposed in this invention can significantly improve the voltage stability of multi-port systems, effectively reduce the degree of voltage imbalance between ports, and extend the stable operating time of the system.

[0145] For example, in power systems, based on the correspondence between thermal management performance and electric field stress threshold, a mapping table first needs to be established. This table is derived through historical data and experimental verification. For instance, when the thermal management performance value is 80%, the corresponding electric field stress threshold is 5 kV / mm. Under current operating conditions, such as in high-voltage direct current transmission scenarios, after the thermal management performance is improved from 75% to 85%, the latest value obtained by querying the mapping table is 5.5 kV / mm. This process involves data interpolation methods to ensure a smooth numerical transition and avoid sudden changes that could lead to system instability.

[0146] Specifically, this correspondence stems from the coupled analysis of the heat conduction equation and the electric field distribution, where improved thermal management performance means increased heat dissipation efficiency, thereby allowing for a higher electric field strength threshold without triggering breakdown.

[0147] In one possible implementation, for a specific wide-bandgap device such as a SiC MOSFET, in photovoltaic inverter applications, based on the latest thermal management performance value of 85%, the electric field stress threshold adjustment is calculated to be 10% of the original threshold of 5kV / mm, i.e., an increase of 0.5kV / mm. This adjustment is obtained through a linear proportional relationship to ensure the safety margin of the device in high-temperature environments.

[0148] For example, during the adjustment process, the device's bandgap of 3.2 eV is considered as a reference, and the effect of temperature on the electric field is evaluated in conjunction with the thermal resistance model to avoid the increase in power consumption caused by over-adjustment.

[0149] Specifically, after determining the adjustment amount of the electric field stress threshold parameter of the wide bandgap device, it is used to update the protection model. This model is a state machine-based framework, including a threshold comparator and alarm logic. During the update, the new threshold of 5.5kV / mm is input into the parameter file of the model and overloaded through a software interface such as MATLAB Simulink to ensure that the model responds to changes in real time.

[0150] In one possible implementation, for the application of solid-state transformers in electric vehicle charging stations, a matching configuration value for the control signal delay is obtained based on the updated protection model. For example, the model output delay time is 2ms. This value is obtained by looking up a preset delay-threshold curve. The curve is constructed based on time-domain simulation to ensure that the signal delay is synchronized with the protection response and to avoid control lag during overload.

[0151] For example, the process of obtaining this matching configuration value involves analyzing the dynamic response of the protection model, such as calculating the delay to match the protection action time triggered by the threshold when the overload current is 150%.

[0152] Specifically, by configuring the control signal delay value to 2ms using a solid-state transformer, the delay parameter of the system's real-time control loop is modified. This loop is a PI controller closed loop. When modifying, the delay parameter is adjusted from the original 1ms to 2ms, which is achieved by writing to the registers of an embedded controller such as a DSP chip, thus ensuring the stability of the loop.

[0153] In one possible implementation, for the modified delay parameters and protection model, in a system-level simulation including overload conditions, PSCAD software is used to perform port voltage calculations, for example, simulating an input port voltage of 10kV and a sudden load increase to 200%. The calculation process iteratively solves the circuit equations step by step, and the output port voltage distribution is obtained as 9.5kV, 9.2kV, etc.

[0154] Specifically, based on the simulation calculation results of the system-level port voltage, the optimized results of the voltage distribution of each port are obtained under specific overload conditions. For example, by adjusting the transformer winding ratio, the voltage of port 1 is optimized to 9.8kV and port 2 to 9.3kV. This optimization result ensures voltage balance, reduces losses, and improves system efficiency.

[0155] This invention also includes specific embodiments for the following microgrid application scenarios:

[0156] Device parameters: 4H-SiC MOSFET (model C2M0080120D) is used. Key parameters obtained from the wide bandgap device parameter database are: bandgap width 3.2 electron volts, breakdown electric field strength 3 megavolts per centimeter, carrier mobility 900 square centimeters per volt per second, thermal conductivity 145 watts per meter per Kelvin, dielectric constant 9.7, and maximum operating junction temperature 200 degrees Celsius.

[0157] SST parameters: It adopts a 3-port topology, with an input port voltage of 10 kV, an output port 1 voltage of 400 V (load type: photovoltaic inverter), an output port 2 voltage of 380 V (load type: energy storage battery), an output port 3 voltage of 220 V (load type: residential load), and a rated power of 500 kVA.

[0158] Simulation and experimental parameters:

[0159] Finite element simulation: The mesh density in the PN junction region is 1 micrometer, the mesh density in the non-critical region is 8 micrometers, the input port voltage is fixed at 10 kV, and the convective heat transfer coefficient is 200 W / m² / Kelvin;

[0160] Overload condition: Port 2 experiences a sudden overload, with the load current suddenly increasing from the rated 100 A to 150 A (1.5 times overload), the overload duration is 100 milliseconds, and the load change slope is 8 A per second;

[0161] Preset thresholds: junction temperature safety threshold 180 degrees Celsius, electric field stress threshold 2.4 megavolts per centimeter, control signal delay preset range not exceeding 50 microseconds;

[0162] Optimization process and results:

[0163] B101: The local electric field distribution was obtained through COMSOL simulation. The electric field strength in the edge region of the device was 2.6 MV / cm, the instantaneous junction temperature was 165 degrees Celsius, and the junction temperature increased by 10 degrees Celsius every 10 milliseconds.

[0164] B102: Defect density of materials analyzed in conjunction (3×10¹) 5 (per cubic centimeter) and multi-port coupling effect, the region between port 2 and port 3 is determined to be the electric field anomalous region;

[0165] B103: By combining control signal delay data (45 microseconds) with measured dielectric constant (9.5), the corrected thermal conductivity is 143 W / m / Kelvin, and the area within 5 mm of the center of the abnormal region is identified as the region with insufficient thermal management.

[0166] B104: Using distributed electric field equalization technology, the voltage compensation for port 2 is 3 volts and the voltage compensation for port 3 is 2 volts. After correction, the power conversion efficiencies of each port are: port 1 97.5%, port 2 97.2%, and port 3 96.8%.

[0167] B105: Based on real-time data fusion, the voltage imbalance risk point is the connection point between port 2 and port 3, and the peak value of the interaction intensity is 0.85 (exceeding the preset threshold of 0.8).

[0168] B106: The local electric field distribution was optimized twice using an improved particle swarm optimization algorithm. The electric field uniformity error was finally converged to 0.9%, and the improvement scheme of adding local heat sinks in the area with insufficient thermal management was determined.

[0169] B107: The electric field stress threshold was updated to 2.5 MV / cm, and the control signal delay was configured to 40 microseconds. Simulation verified the voltages at each port: Port 1 9.8 kV, Port 2 402 V, and Port 3 398 V, with a voltage imbalance rate of 2.1%, meeting the requirements for stable operation of the microgrid.

[0170] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An adaptive optimization method for SST multi-port energy routing under wide bandgap devices, characterized in that, include: Key parameter data is obtained from a wide-bandgap device parameter database. Combined with the operating status information of solid-state transformers, the local electric field distribution is calculated using numerical simulation tools to obtain the junction temperature distribution characteristics of the device. Based on the junction temperature distribution characteristics, material defects and multi-port coupling effects are analyzed to identify abnormal electric field regions. For these abnormal electric field regions, control signal data and device parameters are integrated to correct the thermal conductivity, thus identifying areas with insufficient thermal management performance. Distributed electric field balancing technology is used to dynamically adjust the port voltage stability, resulting in a corrected power conversion efficiency distribution. The interaction characteristics between breakdown electric field strength and dynamic load response are analyzed using the power conversion efficiency distribution to identify voltage imbalance risk points. For these voltage imbalance risk points, the local electric field distribution is optimized to determine a thermal management performance improvement scheme. The system parameters are updated according to the thermal management performance improvement scheme, and the port voltage stability is evaluated through simulation to obtain the voltage distribution optimization results.

2. The method as described in claim 1, characterized in that, The process involves obtaining key parameter data from a wide bandgap device parameter database, combining it with solid-state transformer operating status information, and calculating the local electric field distribution using numerical simulation tools to obtain the device junction temperature distribution characteristics. This includes: extracting bandgap width values ​​and breakdown electric field strength data from the wide bandgap device parameter database; determining the current operating status based on the port voltage stability and power conversion efficiency data of the solid-state transformer during operation; calculating the local electric field distribution under dynamic load response using finite element numerical simulation tools; calculating the power loss density distribution in each region based on the local electric field distribution; obtaining the instantaneous value of the junction temperature distribution based on the power loss density distribution; marking the current operating point as a high-risk state if the instantaneous value of the junction temperature distribution exceeds a preset safety threshold; and combining the instantaneous value of the junction temperature distribution with the junction temperature value at the previous time point to obtain the junction temperature change trend over time.

3. The method as described in claim 1, characterized in that, The step of analyzing material defects and multi-port coupling effects based on the junction temperature distribution characteristics to determine abnormal electric field regions includes: acquiring a junction temperature distribution data sequence; calculating a trend sequence based on the junction temperature distribution data sequence; performing segmented correlation analysis on the material defect density using the trend sequence to obtain a defect density distribution map; performing spatial location matching between the defect density distribution map and the multi-port coupling strength to obtain a coupling influence region; extracting the electric field stress value sequence at the corresponding location through the coupling influence region; comparing the electric field stress value sequence with a preset stress threshold in real time; if a point in the electric field stress value sequence exceeds the preset stress threshold, it is determined to be an abnormal region and the occurrence time is recorded; when the overload protection trigger signal arrives, extracting the location of the abnormal region closest to the trigger time to obtain the abnormal electric field distribution region corresponding to the instant of overload protection action.

4. The method as described in claim 1, characterized in that, The step of integrating control signal data and device parameters to correct thermal conductivity in the abnormal electric field region and identifying areas with insufficient thermal management performance includes: correcting thermal conductivity parameters through an information processing flow; extracting data sets related to electric field stress and thermal management performance from the abnormal electric field region; performing preliminary classification of the data sets using a pre-established analysis model to obtain the distribution of potential inadequate regions; obtaining solid-state transformer control signals and delay data based on the distribution of potential inadequate regions; performing time-series analysis on the delay data, and marking high-risk regions and determining their specific locations if the delay data exceeds a preset range; obtaining wide-bandgap device dielectric constant data through the locations of the high-risk regions; and correlating and comparing the dielectric constant data with the thermal conductivity parameters, and determining key points with insufficient thermal management performance if the deviation exceeds a preset standard, thereby obtaining the distribution range of key points.

5. The method as described in claim 1, characterized in that, The method of dynamically adjusting port voltage stability using distributed electric field equalization technology to obtain a corrected power conversion efficiency distribution includes: acquiring spatial location information of areas with insufficient thermal management; determining the distribution range of devices with high electromagnetic interference intensity based on the spatial location information; filtering the device distribution range using wide-bandgap device carrier mobility data to obtain key influence ranges; applying distributed electric field equalization technology to the key influence ranges to calculate the required electric field adjustment intensity for each port; dynamically adjusting the port voltage using the electric field adjustment intensity to obtain an adjusted voltage value sequence; calculating the power conversion efficiency at the corresponding time based on the adjusted voltage value sequence to obtain an efficiency sequence; and reorganizing the efficiency sequence according to spatial location to obtain a corrected spatial distribution of power conversion efficiency.

6. The method as described in claim 1, characterized in that, The step of analyzing the interaction characteristics between breakdown electric field strength and dynamic load response through the power conversion efficiency distribution to determine voltage imbalance risk points includes: acquiring corrected power conversion efficiency distribution data; obtaining the breakdown electric field sequence of wide bandgap devices under high stress conditions through real-time data fusion processing; performing time-series alignment between the breakdown electric field sequence and the dynamic load response data of solid-state transformers; calculating the interaction strength sequence between the two based on the time-series alignment result; marking the current moment as a potential voltage imbalance moment if the interaction strength sequence exceeds a preset threshold; extracting multi-port coupling state parameters for the potential voltage imbalance moment; determining a risk location set through the correspondence between the multi-port coupling state parameters and the voltage imbalance moment; and sorting the risk location set to obtain the final high-risk point sequence.

7. The method as described in claim 1, characterized in that, The optimization of the local electric field distribution for the voltage imbalance risk point and the determination of the thermal management performance improvement scheme include: collecting voltage imbalance-related data and performing preliminary analysis using information processing technology to obtain voltage imbalance distribution characteristics; calculating the key region of the local electric field distribution based on the voltage imbalance distribution characteristics and the bandgap width data of wide bandgap devices; if the key region exceeds the preset safety range, calibrating using solid-state transformer overload protection data to obtain response parameters; adjusting the thermal management module parameters based on the response parameters to determine the preliminary optimization results; if the preliminary optimization results do not meet the preset standard, performing a secondary analysis of the electric field distribution and overload protection data to obtain improvement directions; and using a pre-established thermal management model to perform simulation verification based on the improvement directions to determine the final thermal management performance improvement scheme.

8. The method as described in claim 1, characterized in that, The process of updating system parameters according to the thermal management performance improvement scheme, evaluating port voltage stability through simulation, and obtaining voltage distribution optimization results includes: obtaining the latest improved values ​​based on the correspondence between thermal management performance and electric field stress threshold; determining the adjustment amount of the electric field stress threshold parameter for wide bandgap devices based on the latest values; updating the threshold setting in the protection model of wide bandgap devices using the adjustment amount; obtaining the matching configuration value of the solid-state transformer control signal delay based on the updated protection model; modifying the delay parameter of the system real-time control loop using the configuration value; performing system-level port voltage simulation calculations including overload conditions for the modified delay parameters and the protection model; and obtaining the optimized port voltage distribution results under specific overload conditions based on the simulation calculation results.

Images