A multi-mode redundant power supply system based on bidirectional inverter power boards

Abstract

The application relates to the technical field of power electronics, and discloses a multi-mode redundant power supply system based on a bidirectional inverter power board, which comprises a topological structure acquisition module, which is used for acquiring a reconfigurable circuit topological structure; an optimal structure generation module, which is used for generating an optimal topological structure according to the reconfigurable circuit topological structure and a multi-parameter real-time monitoring system; a parameter acquisition module, which is used for dynamically adjusting control parameters under different temperature conditions according to a temperature characteristic model library; a predicted state acquisition module, which is used for obtaining a predicted state of the multi-mode redundant power supply system in a future time period according to a preset digital twin model; and a predicted control module, which is used for controlling the multi-mode redundant power supply system according to the predicted state. Through the fusion of multi-stage topological structures, wide-bandgap device characteristic optimization and digital twin technology, the application realizes efficient operation in a wide load range, strong adaptability under extreme environmental conditions and high reliability in long-term operation.

Description

Technical Field

[0001] This application relates to the field of power electronics technology, and more specifically, to a multi-mode redundant power supply system based on a bidirectional inverter power supply board. Background Technology

[0002] In applications such as aerospace, polar research stations, deep-sea equipment, and industrial control systems, power systems face multiple technical challenges. These power systems require long-term stable operation under extreme environmental conditions, including high temperatures (-55°C to 250°C), high radiation (up to 100 klad total dose), and strong magnetic fields (up to 10 Tesla). Simultaneously, load power demands typically fluctuate frequently between light and heavy loads, placing stringent requirements on the efficiency and dynamic response of the power system. Furthermore, in these environments far from maintenance facilities, power system failures can lead to catastrophic consequences, thus demanding extremely high reliability and adaptability.

[0003] Existing technologies suffer from the following shortcomings: traditional bidirectional inverters with a single topology struggle to maintain high efficiency over a wide load range, especially experiencing significant efficiency losses during frequent switching between light and heavy loads; conventional control methods are typically optimized for single extreme factors (such as high temperature), making it difficult to cope with the combined effects of multiple extreme conditions; although wide-bandgap devices such as silicon carbide and gallium nitride possess excellent high-temperature and high-frequency characteristics, their switching behavior, parasitic parameters, and temperature characteristics differ drastically from traditional silicon-based devices, making it difficult for existing control strategies to fully realize their performance potential; device characteristics degrade and age when operating in extreme environments for extended periods, making it difficult to maintain optimal control parameters; the lack of system-level real-time health assessment and predictive fault management mechanisms limits system reliability; and there is a lack of effective solutions to resonance and EMI problems caused by the high di / dt and dv / dt characteristics of wide-bandgap devices. Summary of the Invention

[0004] This application provides a multi-mode redundant power supply system based on a bidirectional inverter power supply board, which solves the technical problems of low efficiency, poor reliability and difficult EMI control of power supply systems in extreme environments.

[0005] This application provides a multi-mode redundant power supply system based on a bidirectional inverter power supply board, including:

[0006] A topology acquisition module is used to acquire a reconfigurable circuit topology; wherein, the reconfigurable circuit topology includes power semiconductor switches, magnetic components, filter capacitors, and driving circuits; the reconfigurable circuit topology can be switched to a full-bridge topology, a half-bridge topology, a three-level topology, or a hybrid topology through different combinations of devices;

[0007] The optimal structure generation module is used to generate the optimal topology based on the reconfigurable circuit topology and the multi-parameter real-time monitoring system.

[0008] The parameter acquisition module is used to dynamically adjust control parameters under different temperature conditions based on the temperature characteristic model library.

[0009] The predicted state acquisition module is used to obtain the predicted state of the multi-mode redundant power system in the future time period based on the preset digital twin model; wherein, the preset digital twin model is trained based on ground test data and updated with on-orbit data.

[0010] The predictive control module is used to control the multi-mode redundant power supply system based on the predicted state.

[0011] The beneficial effects of this application are as follows:

[0012] High-efficiency operation and peak efficiency are achieved through multi-level topology fusion and adaptive switching; the system can operate stably in extreme temperature environments and has enhanced radiation resistance and electromagnetic interference resistance through wide-bandgap device temperature adaptive control and multi-physics coupling; the system's mean time between failures is increased and the lifespan of key components is extended through predictive fault-tolerant control and health management, which greatly improves the system's reliability. Attached Figure Description

[0013] Figure 1 This is a block diagram of a multi-mode redundant power supply system based on a bidirectional inverter power supply board according to this application. Detailed Implementation

[0014] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, some features described in the examples may be combined in other examples.

[0015] At least one embodiment of this application discloses a multi-mode redundant power system 100 based on a bidirectional inverter power board, such as... Figure 1 As shown, it includes:

[0016] The topology acquisition module 110 is used to acquire a reconfigurable circuit topology; wherein, the reconfigurable circuit topology includes power semiconductor switches, magnetic components, filter capacitors and driving circuits; the reconfigurable circuit topology can be switched to a full-bridge topology, a half-bridge topology, a three-level topology or a hybrid topology by combining different devices.

[0017] Specifically, the topology acquisition module includes:

[0018] A reconfigurable circuit topology including full-bridge, half-bridge, and three-level structures is constructed. This architecture consists of power semiconductor switches, magnetic components, filter capacitors, and drive circuits. By combining different components, it can be switched to full-bridge, half-bridge, three-level, or hybrid topologies, realizing flexible switching of multiple conversion modes.

[0019] The system uses wide-bandgap semiconductor devices such as silicon carbide (SiC) MOSFETs or gallium nitride (GaN) HEMTs as switching devices, and integrates current, voltage and temperature sensors in the drive circuit to monitor the operating status.

[0020] The output filter network adopts an LC structure, but the resonant frequency and damping coefficient can be adaptively adjusted according to different topology modes.

[0021] The optimal structure generation module 120 is used to generate the optimal topology based on the reconfigurable circuit topology and the multi-parameter real-time monitoring system.

[0022] Specifically, the optimal structure generation module includes:

[0023] The mapping table generation unit is used to generate a four-dimensional topology efficiency mapping table based on the efficiency characteristics of different topologies under different input voltage, output voltage, load current and temperature conditions.

[0024] By analyzing the efficiency characteristics of each topology under different input voltage, output voltage, load current and temperature conditions, a four-dimensional topology efficiency mapping table is generated.

[0025] For topology Its efficiency function is expressed as:

[0026] ;

[0027] in, Output power represents the effective power output by the system to the load. Input power represents the total power that the system draws from the power source; Power loss represents the power loss of the system during the energy conversion process; System efficiency is expressed as the ratio of output power to input power. It indicates a specific topology type, such as a full-bridge, half-bridge, or three-level structure; Indicates output power; Indicates the input voltage; This indicates the output voltage.

[0028] The power loss can be further decomposed into:

[0029] ;

[0030] in, Conduction loss refers to the energy loss of a power semiconductor device due to its internal resistance when it is in the conduction state. Switching losses refer to the energy loss generated by power semiconductor devices during the turn-on and turn-off processes. For magnetic component losses, including iron losses (hysteresis losses and eddy current losses) and copper losses (winding resistance losses) in transformers and inductors. Auxiliary circuit losses refer to the energy consumed by auxiliary systems such as drive circuits, control circuits, and protection circuits. It indicates a specific topology type, such as a full-bridge, half-bridge, or three-level structure; Indicates output power; Indicates the input voltage; This indicates the output voltage.

[0031] A complete efficiency mapping table is generated by interpolating measured data under various operating conditions and stored in the controller's non-volatile memory. This mapping table provides the decision-making basis for the topology selection algorithm.

[0032] The demand acquisition unit is used to collect load parameters from the multi-parameter real-time monitoring system to obtain the current load status, and to obtain the predicted power demand within a preset time period based on historical load data.

[0033] A multi-parameter real-time monitoring system is adopted to collect data including input voltage. Output voltage Load current Parameters such as power flow direction and ambient temperature are included. The sampling frequency is set according to the system's dynamic response requirements, typically ranging from 10kHz to 50kHz.

[0034] Meanwhile, short-term power demand is predicted based on historical load data, and the prediction algorithm uses an autoregressive integral moving average model:

[0035] ;

[0036] in, for Predicted power after time, , , Representing time respectively , and The actual power value, The number of autoregressive terms. , , respectively represent time. , and The weighting of the actual power value on the predicted value. , , Representing time respectively , and The prediction error (white noise term). , Representing time respectively and The model uses the prediction error as a correction weight for the current prediction. By comprehensively considering historical power data and prediction errors, it achieves accurate predictions of future power demand.

[0037] The prediction results are used for topology switching decisions to avoid system instability caused by frequent switching.

[0038] The optimal structure generation unit is used to generate the optimal topology based on the current load state, predicted power demand, and a four-dimensional topology efficiency mapping table.

[0039] Here, based on the current load status and prediction results, and combined with the efficiency mapping table, the efficiency of each topology under the current conditions is calculated, and the optimal topology is determined by comparison. :

[0040] ;

[0041] in, It is a set of all available topologies, including switchable topologies such as full-bridge, half-bridge, and three-level; Indicates a specific topology type; Current power demand represents the output power required by the system under current load. Input voltage represents the system's input power supply voltage; Output voltage, representing the voltage value that the system needs to output; System efficiency is expressed as the ratio of output power to input power. This represents the optimal topology that is most efficient under the current operating conditions. This represents the independent variable that makes the function reach its maximum value, i.e., finding the efficiency. Largest topology .

[0042] To prevent frequent switching under similar efficiency conditions, a lag interval and minimum duration constraint are introduced: switching only occurs when the efficiency improvement exceeds a threshold. (Typical value 1.5%) and the current topology has been in continuous use for more than [time period missing]. The switch is executed only after a typical 10-second interval.

[0043] At the power threshold point Nearby, a smooth transition strategy is adopted: before the source topology is shut down, the target topology is pre-started and gradually increases its power sharing ratio to achieve a disturbance-free handover. Representing topology and topology The power point at which the efficiencies are equal is the point at which the efficiencies of the two topologies are equal. Equal power values A set; , They represent the power Lower topology , Efficiency; source topology refers to the currently running topology; target topology refers to the new topology that the system will switch to.

[0044] Through the above steps, the system can automatically select the optimal topology based on load changes, achieving high-efficiency operation over a wide load range.

[0045] The parameter acquisition module 130 is used to dynamically adjust the control parameters under different temperature conditions based on the temperature characteristic model library.

[0046] Specifically, the parameter acquisition module includes:

[0047] The model library acquisition unit is used to acquire a temperature characteristic model library, which includes switching time temperature model, on-resistance temperature model, breakdown voltage temperature model and threshold voltage temperature model.

[0048] A mathematical model of the temperature-dependent characteristics of wide-bandgap devices is established to accurately describe the variation of key parameters with temperature. This mainly includes the following parameter models:

[0049] Switching time-temperature model:

[0050] ;

[0051] in, For temperature The following switch-on and switch-off times (including activation time) and shutdown time ), indicating that wide bandgap devices are available at temperatures The time required to complete the activation or deactivation process under certain conditions; Reference temperature The switching time at standard room temperature represents the reference switching time value of the device under standard room temperature conditions. The switching time temperature coefficient represents the relative rate of change of switching time caused by a unit change in temperature. This refers to the actual operating temperature of the current device; The reference temperature serves as the baseline for calculating temperature characteristics. This formula describes a linear relationship between the switching time of a wide-bandgap semiconductor device and temperature.

[0052] On-resistance temperature model:

[0053] ;

[0054] in, For temperature The on-resistance represents the equivalent resistance of a power device in the on-state. Reference temperature The on-resistance is expressed as the reference value of the on-resistance measured under standard reference temperature conditions. This refers to the actual temperature of the current device. For reference temperature; The temperature coefficient represents the sensitivity of the on-resistance to temperature changes; the larger the value, the more significant the effect of temperature on the on-resistance. The nonlinear exponent describes the degree of nonlinearity in the change of on-resistance with temperature. A value greater than 1 indicates that the on-resistance increases at a faster rate as the temperature rises.

[0055] Breakdown voltage-temperature model:

[0056] ;

[0057] in, For temperature The breakdown voltage at a given temperature indicates the voltage at which the voltage drops below the set temperature. The maximum voltage that a wide-bandgap device can withstand under certain conditions; Reference temperature The breakdown voltage at the standard reference temperature represents the reference breakdown voltage value of the device under standard reference temperature conditions. The breakdown voltage temperature coefficient represents the relative rate of change of breakdown voltage caused by a unit change in temperature. This refers to the actual operating temperature of the current device; The reference temperature serves as the baseline for calculating temperature characteristics. This formula describes a linear relationship between the breakdown voltage of a wide-bandgap semiconductor device and temperature.

[0058] Threshold voltage-temperature model:

[0059] ;

[0060] in, For temperature The threshold voltage at a given temperature indicates the voltage at which the temperature is below the threshold voltage. The minimum gate voltage required to turn on the device under certain conditions; Reference temperature The threshold voltage is the reference threshold voltage value measured under standard room temperature conditions. This refers to the actual operating temperature of the current device; The reference temperature serves as the benchmark for calculating temperature characteristics. The threshold voltage temperature coefficient represents the change in threshold voltage caused by a unit change in temperature. A negative sign indicates that the threshold voltage typically decreases as temperature increases. This formula describes a linear relationship between the threshold voltage of a wide-bandgap semiconductor device and temperature.

[0061] Through experimental measurements, the specific parameter values ​​of each device in the above model were determined, and an accurate temperature characteristic model library was established.

[0062] Junction temperature unit, used to estimate the device junction temperature based on the thermal resistance network model.

[0063] A thermal resistance network model is used to estimate the device junction temperature in real time, avoiding the complexity and cost of direct measurement.

[0064] ;

[0065] in, Junction temperature is the actual temperature of the PN junction inside a semiconductor chip and is a key parameter for device reliability and performance. Ambient temperature (measured by a thermistor or thermocouple) indicates the temperature of the environment or cooling medium surrounding the device; For the first Power loss can be categorized into different types, including conduction loss, switching loss, and drive loss. The corresponding thermal resistance represents the thermal impedance during the process of heat transfer from the heat source to the environment. The larger the value, the worse the heat dissipation performance. This indicates that the summation is performed on all loss types, taking into account the combined effect of all heat sources in the device.

[0066] Power loss Calculated based on the real-time operating status of the device:

[0067] Conduction loss:

[0068] ;

[0069] in, Conduction loss represents the power loss generated by a power device when it is in the conduction state. The effective value current represents the root mean square value of the current flowing through the device. For junction temperature The on-resistance is the equivalent resistance between the drain and source when the power device is turned on under the current junction temperature conditions. The junction temperature is the actual temperature of the PN junction inside the semiconductor chip. This formula describes the relationship between the conduction loss of a power device and the product of the square of the current and the on-resistance.

[0070] Switching losses:

[0071] ;

[0072] in, Switching losses represent the power losses generated by power devices during the turn-on and turn-off processes. The switching frequency represents the number of times the power device switches per second. "Power-on energy" refers to the energy consumed by a power device during the transition from an off state to an on state. Turn-off energy represents the energy consumed by a power device during the transition from an on-state to an off-state. and These are all functions of current, voltage, and temperature, meaning their values ​​change with variations in the device's operating current, voltage, and temperature, typically increasing with increasing current, voltage, and temperature. This formula describes the relationship between the switching losses of a power device and the product of the switching frequency and the energy of a single switching operation.

[0073] The system places temperature sensors at multiple key points and uses a Kalman filter algorithm to fuse direct measurements and model estimates, thereby improving the accuracy of junction temperature estimation.

[0074] ;

[0075] in, For the first The estimated temperature of the first step is represented by the temperature at the first step. The estimated junction temperature of the device calculated by the Kalman filter algorithm at each time step; For the first The estimated junction temperature of the step represents the estimated junction temperature of the previous time step; To measure temperature, it indicates the time in the first... The temperature value directly measured by the temperature sensor at each time step; Kalman gain is a weighting coefficient used to balance the confidence of model predictions and actual measurements. Its value is dynamically calculated based on the characteristics of system noise and measurement noise. The observation matrix represents the mapping relationship between the state variable (estimated junction temperature) and the observed variable (measured temperature). In simple cases, it can be an identity matrix. This formula describes the state update process in the Kalman filter algorithm, which obtains a more accurate junction temperature estimate by fusing model predictions and actual measurements.

[0076] The parameter adjustment unit is used to dynamically adjust the device control parameters based on the estimated device junction temperature.

[0077] Based on the estimated junction temperature, the following control parameters are dynamically adjusted to optimize device performance:

[0078] Switching frequency adjustment:

[0079] ;

[0080] in, For temperature The actual switching frequency under the current junction temperature indicates the switching frequency value used by the system under the current junction temperature conditions. The rated switching frequency represents the reference switching frequency value designed for the system under standard operating conditions. This represents the junction temperature of the current device, indicating the actual temperature of the PN junction inside the semiconductor chip; The maximum allowable junction temperature represents the upper limit of the temperature at which the device can operate safely. Exceeding this temperature may lead to performance degradation or permanent damage to the device. Rated junction temperature indicates the standard junction temperature of the device under normal operating conditions; This is an adjustment factor that represents the sensitivity of temperature to switching frequency adjustment (typical value 0.7). The larger the value, the more significant the effect of temperature change on the switching frequency. This means taking the smaller value between the expression within the parentheses and 1 to ensure that the switching frequency does not exceed the rated value. When the junction temperature rises, this formula will appropriately reduce the switching frequency to reduce switching losses, thereby controlling the rise in device temperature.

[0081] Dead time adjustment:

[0082] ;

[0083] in, For temperature The dead time in the figure represents the delay time set by the power device during complementary switching to prevent shoot-through short circuit; Minimum dead time represents the shortest safe dead time value allowed by the system, used under low temperature conditions; Maximum dead time, representing the longest dead time value set by the system, used under high temperature conditions; This represents the junction temperature of the current device, indicating the actual temperature of the PN junction inside the semiconductor chip; This is the lowest operating junction temperature, representing the lower limit of normal device operation; The maximum allowable junction temperature represents the upper limit of the temperature at which the device can operate safely. As the temperature increases, the dead time should be appropriately increased to avoid shoot-through, as high temperatures affect the switching characteristics and turn-off speed of the device.

[0084] Gate drive voltage adjustment:

[0085] ;

[0086] in, For temperature The gate drive voltage represents the actual voltage applied to the gate of the power device under the current junction temperature conditions. The rated gate drive voltage represents the reference gate drive voltage value designed under standard operating conditions. "Adjustment Range" indicates the maximum adjustable range of the gate voltage. This represents the junction temperature of the current device, indicating the actual temperature of the PN junction inside the semiconductor chip; Rated junction temperature indicates the standard junction temperature of the device under normal operating conditions; The maximum permissible junction temperature represents the upper temperature range at which the device can operate safely. Temperature drift of the threshold voltage is compensated for by adjusting the drive voltage, ensuring that the device maintains stable switching characteristics at different temperatures.

[0087] Overcurrent protection threshold adjustment:

[0088] ;

[0089] in, For temperature The overcurrent protection threshold value indicates the overcurrent protection trigger current value set by the system under the current junction temperature conditions. The rated overcurrent protection threshold represents the reference overcurrent protection current value designed at standard operating temperature; The adjustment coefficient represents the sensitivity of temperature changes to the overcurrent protection threshold adjustment; the larger the value, the more the overcurrent protection threshold decreases when the temperature rises. This represents the junction temperature of the current device, indicating the actual temperature of the PN junction inside the semiconductor chip; Rated junction temperature indicates the standard junction temperature of the device under normal operating conditions; The maximum permissible junction temperature represents the upper limit of the temperature at which the device can operate safely. As the temperature increases, this formula appropriately lowers the overcurrent protection threshold to enhance device protection and prevent damage from overcurrent under high-temperature conditions.

[0090] The management unit is used to perform heat dissipation management based on multi-point temperature monitoring for modules with power exceeding the preset power.

[0091] To address the issue of uneven temperature distribution in high-power modules, a heat dissipation management system based on multi-point temperature monitoring is implemented.

[0092] Constructing a device temperature distribution map: A two-dimensional map of the surface temperature distribution of the device is generated through multi-point temperature sensing and thermal model interpolation.

[0093] Calculate the temperature gradient:

[0094] ;

[0095] in, The temperature gradient vector represents the rate of change and direction of the temperature field in space; Let be the partial derivative of temperature in the x-direction, representing the rate of change of temperature along the x-axis; The partial derivative of temperature in the y-direction represents the rate of temperature change along the y-axis. The magnitude of the temperature gradient vector reflects the drasticness of temperature change, and its direction points in the direction of the fastest temperature increase; it is an important parameter for heat flow analysis and thermal stress assessment.

[0096] Dynamic load distribution: For parallel devices, the current sharing ratio of each device is adjusted according to the temperature distribution.

[0097] ;

[0098] in, For the first The current allocated to the device represents the current allocated to the first device. The actual operating current value of each parallel device; The total current represents the total current flowing through the entire parallel device group; The temperature of the device represents the first... The current operating temperature of each device; The highest temperature among all devices indicates the temperature of the device with the highest temperature in the parallel device group; The distribution index represents the sensitivity coefficient of temperature difference to current distribution; the larger the value, the more significant the effect of temperature difference on current distribution. This refers to the total number of devices connected in parallel. Indicates all The formula sums the current from the parallel devices. This achieves dynamic load distribution based on temperature, where devices with lower temperatures will carry more current, thus achieving temperature balance and preventing hot spots from forming.

[0099] Through the above steps, the system can adaptively adjust the control parameters according to the device temperature state, fully utilize the performance potential of wide bandgap devices under extreme temperature conditions, and ensure the safe and reliable operation of the system.

[0100] The prediction state acquisition module 140 is used to obtain the predicted state of the multi-mode redundant power system in the future time period based on the preset digital twin model; wherein, the preset digital twin model is trained based on ground test data and the parameters of the digital twin model are updated based on on-orbit data.

[0101] Specifically, the predicted state acquisition module includes:

[0102] The model determination unit is used to obtain a multi-physics coupled model that includes the electrical domain, thermal domain, radiation domain, and mechanical stress domain, and describes the state changes and interaction relationships of each physical quantity through a set of differential equations.

[0103] A multiphysics coupled model encompassing the electrical, thermal, radiation, and mechanical stress domains is constructed, and the state changes and interactions of each physical quantity are described through a system of differential equations:

[0104] ;

[0105] in, The system state vector includes electrical state variables (current, voltage, etc.), thermal state variables (temperature distribution), cumulative radiation dose, and mechanical stress distribution, etc. The control vector contains switching signals, modulation parameters, etc. This is a vector of environmental parameters, including ambient temperature, radiation intensity, and magnetic field strength, etc. It represents the rate of change of the system state vector over time, describing the dynamic evolution of the system state; It is a nonlinear state equation, representing a functional relationship that maps the current state of the system, control input, and environmental conditions to the rate of change of state; The time variable represents the time of system evolution. This equation is the core mathematical expression of the multiphysics coupled digital twin model, describing the dynamic changes of the system's state over time in multiple physical domains, including electrical, thermal, radiation, and mechanical domains.

[0106] The physical domain models are as follows:

[0107] Electrical domain model: The electrical characteristics of power electronic systems are described using either an average value model or a switching model.

[0108] ;

[0109] in, The inductance value in the circuit represents the inductance of the energy storage element. The rate of change of current represents the amount of change in current per unit time. Input voltage, representing the voltage value supplied by the power source; This represents the output voltage, indicating the voltage value at the load terminal. The equivalent resistance of the circuit represents the loss resistance in the circuit. The current in the circuit; The capacitance value in the circuit represents the capacitance of the energy storage element. Voltage rate of change represents the amount of voltage change per unit time; This is the voltage across the capacitor; The load resistance represents the equivalent resistance of the electrical equipment connected to the circuit output. Inductance is one of these parameters. ,capacitance and resistance May vary with temperature The fact that they change indicates that they are functions of temperature, and their values ​​change as the temperature rises or falls.

[0110] Thermal domain model: The temperature distribution of the system is described using a finite element method or an equivalent thermal resistance network model.

[0111] ;

[0112] in, Density is the mass of a material per unit volume. Specific heat capacity represents the amount of heat required to raise the temperature by one unit per unit mass of material. Let be the partial derivative of temperature with respect to time, representing the rate of change of temperature over time; It is a divergence operator used to calculate the divergence of a vector field; Thermal conductivity represents a material's ability to conduct heat. The temperature gradient represents the rate and direction of temperature change in space. Let be the volumetric heat source density, representing the heat generated per unit volume. In this system, this heat is mainly generated by electrical losses (such as switching losses and conduction losses). This equation is the fundamental equation for heat conduction, describing the distribution of the temperature field in the system over time and space.

[0113] Radiation effect model: describes the impact of cumulative radiation dose on device parameters.

[0114] ;

[0115] in, These are parameters affected by radiation (such as threshold voltage shift, leakage current increase, etc.), representing the electrical parameter values ​​of the device that change over time under radiation conditions; The rate of change of the parameter over time represents the rate of degradation of the device parameter under radiation. Radiation dose rate, representing the radiation dose received by the device per unit time; The radiation sensitivity coefficient represents the degree to which device parameters are sensitive to radiation; it is a proportionality coefficient relating device characteristics to the type of radiation. This equation describes the degradation process of semiconductor device parameters with cumulative radiation dose under space radiation conditions and is the core expression of the radiation effect model.

[0116] Mechanical stress model: describes the distribution of mechanical stress caused by thermal expansion and external vibration.

[0117] ;

[0118] in, Here, is the divergence operator, representing the divergence of the tensor field; For stress tensor, it represents the stress state at various points inside the material; Volume force refers to the external force acting per unit volume, such as gravity. Density is the mass of a material per unit volume. Let be the second partial derivative of the displacement field with respect to time, and let represent the acceleration field; Let be the displacement field, describing the displacement vector of each point in the material relative to its initial position. This equation is the fundamental equation describing the relationship between material deformation and stress in solid mechanics.

[0119] Thermal stress through the coefficient of thermal expansion Coupled with temperature field:

[0120] ;

[0121] in, Thermal strain represents the relative dimensional change of a material due to temperature variations. is the coefficient of thermal expansion, which represents the relative change in size caused by a unit change in temperature; The current temperature indicates the actual temperature at which the material is located. The reference temperature is denoted as the base temperature under no thermal strain. This equation describes the thermal expansion effect of a material under temperature changes and is a key expression for thermomechanical coupling.

[0122] The coupling relationships between different physical domains are achieved through the following mechanisms:

[0123] Electrical → Heat: Electrical losses are treated as a heat source.

[0124] Thermal → Electrical: Temperature affects the electrical characteristics of devices;

[0125] Radiation → Electrical: Radiation damage affects electrical parameters;

[0126] Thermal → Mechanical: Temperature gradient causes thermal stress;

[0127] Mechanical → Electrical: Stress affects the contact resistance of devices and the reliability of interconnections.

[0128] The calibration unit is used to perform online calibration of the multiphysics coupling model according to the recursive parameter identification algorithm to obtain the target digital twin model.

[0129] To ensure the accuracy of the digital twin model, a recursive parameter identification algorithm is used to achieve online calibration of the model.

[0130] State observation equations:

[0131] ;

[0132] in, The system measurement output represents the measured values ​​of various physical quantities collected from the actual system, such as voltage, current, and temperature. The observation function represents the mathematical relationship that maps the system state and parameters to the observable output; it is a nonlinear function. This is the system state vector, which contains various state variables describing the dynamic behavior of the system, such as electrical state and thermal state. The parameters to be identified represent the set of parameters in the digital twin model that need to be updated through online learning methods, such as physical parameters like resistance, inductance, and thermal resistance. For measurement noise, this represents the unavoidable random error during the measurement process. This equation describes the relationship between the actual measured values ​​and the model predictions of the system, and is the basis for parameter identification and model calibration.

[0133] Parameter identification algorithm: The model parameters are updated in real time using the recursive least squares method.

[0134] ;

[0135] in, This is the model parameter vector for the next time step, representing the updated parameter estimates; This is the model parameter vector at the current moment, representing the current parameter estimates; The gain matrix controls the step size and direction of parameter updates, which are determined based on the estimation error covariance and the input signal. This is the output vector obtained from the actual system measurements, which includes the measured data from each sensor; The output vector predicted by the digital twin model, based on the current parameters. Calculated; Prediction error represents the deviation between the model output and the actual measured value, and is used to drive the parameter update process.

[0136] ;

[0137] in, The Kalman gain matrix represents the weighting coefficients for parameter updates, which determine the degree of influence of the new measurements on the parameter estimates. The parameter estimation error covariance matrix represents the magnitude of the uncertainty in the current parameter estimation. The observation matrix represents the sensitivity of the observation output to the parameters, and is the result of calculating the partial derivative of the observation function with respect to the parameters at the current parameter estimate. The noise covariance matrix is ​​used to represent the statistical characteristics of noise during the measurement process. It is the identity matrix, used in matrix calculations to maintain the consistency of the matrix dimensions; Let be the updated parameter estimation error covariance matrix, representing the uncertainty after the parameter update. (Superscript) Indicates matrix transpose, superscript This represents finding the inverse of a matrix.

[0138] Model self-calibration trigger condition: When the error between the measured output and the model prediction exceeds a threshold, parameter updates are triggered.

[0139] ;

[0140] in, Let L be the Euclidean norm (L2 norm), representing the measured output vector. With model predicted output vector The distance between them is calculated by taking the square root of the sum of the squares of the differences between the components of the two vectors, with the same units as the measured quantity. The preset error threshold represents the critical error value that triggers model self-calibration. When the error between the measured value and the model's predicted value exceeds this threshold, the system will initiate parameter identification and model update processes, with the unit being the same as the measured quantity. This inequality is the trigger condition for the model self-calibration mechanism, ensuring the synchronization accuracy between the digital twin model and the actual system.

[0141] At the same time, the system periodically injects small signal disturbances to ensure parameter identifiability.

[0142] The prediction unit is used to obtain the predicted state of the multi-mode redundant power system in the future time period based on the target digital twin model.

[0143] The future state of the system is predicted using a calibrated digital twin model, and the prediction error is reduced through feedback compensation.

[0144] State prediction: The numerical integration method is used to solve the system state equations and predict future state evolution.

[0145] ;

[0146] in, For predicted future moments The system state vector contains a set of variables describing the state of each physical domain of the system; For the current moment The system state vector represents the current physical state of the system. The system state equations describe the mathematical relationships that govern the evolution of the system state over time. For state variables, representing the instantaneous state of the system during the integration process; The control input vector represents the control signal applied to the system; This is an environmental parameter vector, representing the external environmental conditions that affect the system's behavior; Let be the integral variable, representing the time from the current moment. To the future Any point in time between; The prediction time step represents the time span of the state prediction. This equation describes the computational process of predicting future states based on the current state and the system dynamics model using numerical integration methods.

[0147] In actual calculations, a suitable numerical integration method is selected based on the system characteristics, such as the fourth-order Runge-Kutta method or the adaptive step size method.

[0148] Error compensation unit: Corrects the predicted state obtained from the target digital twin model according to the error model; wherein, the error model is used to describe the error between the actual system response and the predicted state obtained from the target digital twin model; and describes the error evolution characteristics according to the autoregressive moving average model.

[0149] Error compensation: An error model is established by comparing the actual system response with the digital twin prediction.

[0150] ;

[0151] in, For error vector, it represents the difference between the actual system output and the predicted output of the digital twin model, with the same unit as the corresponding physical quantity; For the actual system at time The output vector contains various state variables measured from the physical system, such as measured values ​​of voltage, current, and temperature. For digital twin models at time The predicted output vector contains the predicted values ​​of the corresponding state variables calculated by the model. This equation defines the basis of the error compensation mechanism, providing a basis for subsequent error modeling and prediction correction by calculating the deviation between the actual system and the digital twin model output.

[0152] The error evolution characteristics are described using an Autoregressive Moving Average (ARMA) model:

[0153] ;

[0154] in, For the current moment The error vector represents the deviation between the actual system and the output of the digital twin model; For the past moment The error vector represents historical error data; The order of the autoregressive term represents the number of time steps of the historical error data considered. This is the autoregressive coefficient matrix, representing the weights of the influence of historical errors on current errors; The order of the moving average term represents the number of time steps of the historical noise data considered. This is the moving average coefficient matrix, representing the weights of historical noise on the current error; For a moment The white noise vector represents unpredictable random disturbances in the system.

[0155] The prediction results of digital twins are corrected using an error model:

[0156] ;

[0157] in, The corrected prediction output vector after error compensation represents the future time step. The final result of the system state prediction after error correction; For digital twin models to directly predict future moments The output vector contains the predicted values ​​of each state variable obtained from the model calculation; For future moments predicted based on the ARMA error model The error vector represents the estimated value of the error in future predictions; The current moment; The prediction time step represents the time interval from the current time to the target prediction time. This equation describes the calculation process of obtaining a corrected prediction result after error compensation by adding the original prediction result of the digital twin model to the prediction error estimate, thereby improving the accuracy of system state prediction.

[0158] A building unit is used to construct a visualization interface for the target digital twin model; the visualization interface is used to display the state distribution of each physical domain.

[0159] Build a visual interface for the digital twin to display the state distribution of each physical domain in real time:

[0160] Electrical domain: Displays real-time changes in topology, current paths, voltage distribution, and power flow.

[0161] Thermal domain: Displays temperature distribution in the form of a heat map, marking hotspot locations and temperature gradients.

[0162] Radiation domain: Displays the cumulative radiation dose distribution and the radiation damage status of sensitive devices.

[0163] Mechanical domain: Displays stress distribution and potential mechanical failure risk points.

[0164] It also provides time series plots and spectral analysis results for each physical quantity, and enhances data understanding through visual elements such as color scales, contour lines, and vector arrows.

[0165] Through the above steps, the system establishes a high-fidelity digital twin model, realizing virtual mapping and state prediction of the actual physical system, and providing a theoretical basis for adaptive control and health management.

[0166] The predictive control module 150 is used to control the multi-mode redundant power supply system based on the predicted state.

[0167] Specifically, the following steps are included:

[0168] The predictive control unit is used to control the multi-mode redundant power supply system according to the predictive control algorithm. The predictive control algorithm uses rolling time-domain prediction for control and is based on the control objective function, which includes efficiency control, thermal management control, and reliability control.

[0169] Utilizing digital twin models to predict future system states enables prediction-based optimal control.

[0170] Control objective function construction: Define a multi-objective function that includes efficiency, thermal management, and reliability:

[0171] ;

[0172] in, Let be the overall control objective function, and let represent the overall performance evaluation index of the system. This is the system state vector, which includes physical quantities such as voltage, current, and temperature. The control input vector includes control parameters such as switching frequency and duty cycle; This is an environmental parameter vector, representing external conditions such as temperature and radiation. As an efficiency sub-objective, evaluate the system's energy conversion efficiency; As a sub-objective of thermal management, evaluate temperature distribution uniformity and maximum temperature; As a reliability sub-objective, assess the system lifetime attrition rate; , , Let these represent the weighting coefficients for efficiency, thermal management, and reliability, respectively, and satisfy the following conditions: , .

[0173] The efficiency sub-objective is defined as the negative value of the electrical energy conversion efficiency:

[0174] ;

[0175] in, The efficiency sub-objective function represents the negative value of the system efficiency and is used to transform the efficiency maximization problem into the objective function minimization problem during the optimization process. The system's energy conversion efficiency; The system output power represents the effective power received by the load. Let be the system input power, and represent the total power drawn from the power source. This formula maximizes efficiency by taking the negative of the efficiency value during the optimization process of minimizing the objective function.

[0176] The thermal management sub-objective is defined as a weighted sum of temperature uniformity and maximum temperature:

[0177] ;

[0178] in, Let be the objective function for thermal management, representing a comprehensive evaluation index of the system's temperature distribution and maximum temperature; This is the temperature uniformity weighting coefficient, with a value range of [value range missing]. This is used to balance the uniformity of temperature distribution and the relative importance of the highest temperature in the objective function; This represents the highest temperature in the system, indicating the temperature value at the point where the temperature is highest. The system average temperature represents the average temperature value at all monitoring points in the system. The term indicates the degree of unevenness in temperature distribution; The rated maximum operating temperature of the system indicates the maximum allowable operating temperature of the system as designed. The value represents the ratio of the actual maximum temperature to the rated temperature, reflecting the system's temperature margin.

[0179] The reliability sub-target is defined as lifetime attrition rate:

[0180] ;

[0181] in, Let be the reliability sub-objective function, representing the system lifetime attrition rate; Damage accumulation rate represents the increase in system damage per unit time. Cumulative damage represents the total degree of damage the system has already suffered; The current moment; This is an integral variable, representing the time range from system startup to the current moment; This is the damage accumulation function, which describes the rate at which system damage accumulates under given operating conditions; Electrical stress represents the stress on the device at time [time]. The electrical stress it bears; Temperature represents the time at which the device is in motion. Operating temperature; The radiation dose rate represents the device's radiation dose rate at time [time]. The intensity of radiation absorbed. This formula quantifies the rate of change of system reliability over time by calculating the time derivative of damage accumulation, and is used to minimize system lifetime loss in control optimization.

[0182] Control parameter optimization: Find the optimal sequence of control parameters in the prediction time domain that minimizes the objective function.

[0183] ;

[0184] in, The optimal control parameter vector represents the combination of control parameters that minimizes the objective function. This represents finding the parameters that minimize the objective function. Operators; Let be the overall control objective function, and let represent the overall performance evaluation index of the system. This is the system state vector, which includes physical quantities such as voltage, current, and temperature. The control input vector includes control parameters such as switching frequency and duty cycle; This is an environmental parameter vector, representing external conditions such as temperature and radiation. This formula describes the conditions under a given system state. and environmental conditions Next, find the objective function that makes Control parameters for achieving minimum value The optimization process.

[0185] For uncertain environmental parameters The minimum-maximum optimization strategy is adopted:

[0186] ;

[0187] in, The optimal control parameter vector represents the control strategy that minimizes the objective function under worst-case conditions. This represents the search for the control parameter vector that minimizes the objective function. ; Represents the set of all possible environmental parameters Find the maximum (worst) objective function value; The overall control objective function; This is the system state vector; To control the input vector; An environmental parameter vector; This represents the set of possible values ​​for environmental parameters, indicating the range of all environmental conditions the system may face.

[0188] Predictive control algorithm implementation: A model-based rolling temporal predictive control (MPC) strategy is adopted.

[0189] At the present moment Predicting the future using digital twin models System state within a time step;

[0190] Solve the optimization problem to obtain the control sequence. ;

[0191] Only the first control action in the sequence is executed. ;

[0192] In the next moment Obtain the new system state and repeat the above process;

[0193] To reduce computational complexity when solving optimization problems, continuous control parameters are discretized into a finite set, and an approximate optimal solution is quickly found through intelligent search algorithms (such as improved particle swarm optimization).

[0194] The predictive control module is also used for fault prediction and early intervention;

[0195] Based on digital twin models and historical operational data, early prediction and proactive intervention of system failures can be achieved.

[0196] Device cumulative stress and degradation model:

[0197] Establish a cumulative stress and degradation model for power devices to describe the changes in device health status over time and under operating conditions:

[0198] ;

[0199] in, The cumulative damage represents the device's damage from its initial state to time [time value missing]. The total extent of accumulated damage; This is the damage accumulation function, which describes the rate at which system damage accumulates under given operating conditions; Electrical stress represents the stress on the device at time [time]. The electrical stress it bears; Temperature represents the time at which the device is in motion. Operating temperature; The radiation dose rate represents the device's radiation dose rate at time [time]. The intensity of radiation received; Let be the integral variable, representing the time from system startup to the current moment. Time range; The current time indicates the point in time at which the cumulative damage is assessed.

[0200] For different failure mechanisms, corresponding physical damage models are established:

[0201] Solder joint fatigue caused by thermal cycling:

[0202] ;

[0203] in, The number of thermal cycles that the solder joint can withstand before failure; These are constants related to materials and structures; This represents the temperature variation range during the thermal cycle. This is a temperature amplitude index, characterizing the degree to which the amplitude of temperature change affects lifespan; Average temperature; The average temperature index characterizes the degree to which average temperature affects lifespan; This refers to the thermal cycling frequency; It is a frequency index that characterizes the degree to which cycle frequency affects lifespan.

[0204] Chip cracking caused by thermal stress:

[0205] ;

[0206] in, This refers to the chip's failure time due to cracking. These are constants related to the chip materials and structure; Represents an exponential function; Thermal activation energy characterizes the activation effect of temperature on the failure process; Boltzmann's constant has a value of eV / K; For chip temperature; The mechanical stress that the chip withstands; The stress index characterizes the degree of influence of mechanical stress on failure time.

[0207] Gate oxide breakdown:

[0208] ;

[0209] in, This refers to the gate oxide breakdown time. These are constants related to oxide materials and processes; The activation energy is that of oxides. Boltzmann's constant has a value of eV / K; This refers to the device's operating temperature. The electric field strength that the gate oxide withstands; is the electric field index, which characterizes the degree of influence of electric field strength on breakdown time; This represents an exponential function.

[0210] Cumulative radiation damage:

[0211] ;

[0212] in, This represents the offset of the device threshold voltage, characterizing the degree of device parameter drift caused by radiation. The saturation offset coefficient represents the maximum possible threshold voltage offset. The radiation susceptibility coefficient; Cumulative radiation dose represents the total radiation energy absorbed by the device. This represents an exponential function.

[0213] Remaining Useful Life (RUL) Prediction:

[0214] Based on the cumulative damage model and the current state, predict the remaining service life of key system components:

[0215] ;

[0216] in, The remaining useful life is indicated at time [time]. Predicted remaining system availability time; The failure threshold represents the critical value of cumulative damage when the system reaches a failure state; The current cumulative damage is dimensionless, representing the system's current value at time t. The total extent of damage already suffered; The current damage rate is represented at time [time value missing]. The rate of increase in system damage; This is the damage accumulation function, which describes the rate at which system damage accumulates under given operating conditions; Electrical stress represents the stress on the device at time [time]. The electrical stress it bears; Temperature represents the time at which the device is in motion. Operating temperature; The radiation dose rate represents the device's radiation dose rate at time [time]. The intensity of radiation it withstands.

[0217] To improve prediction accuracy, the RUL distribution is updated using a particle filter algorithm:

[0218] ;

[0219] in, Given an observation history Remaining service life under certain conditions The posterior probability distribution represents a probability estimate of the remaining lifespan given historical observation data. Given the current damage state Remaining service life under certain conditions The conditional probability distribution represents the prediction of remaining life under a specific damage state; Given an observation history Current damage state under the given conditions The posterior probability distribution represents the estimate of the current damage state based on historical observation data, and this distribution is updated in real time using a particle filtering algorithm. From the start of system operation to the current moment The historical collection of all observation data, including sensor measurements, operating parameters, and other information; For the current moment The system damage state indicates the degree of system degradation; Represents all possible damage states Perform integration to calculate the marginal probability distribution.

[0220] Early intervention strategy for faults:

[0221] Based on the RUL prediction results, implement tiered intervention measures:

[0222] when Time: Maintain normal operation and update forecasts regularly;

[0223] when At this time: Enter early warning mode and adjust control parameters to reduce stress;

[0224] when Time: The maintenance plan is triggered, and preparations are made to replace the component;

[0225] when Time: Immediately enter safe mode to prevent catastrophic failure.

[0226] Intervention measures include reducing power, decreasing temperature cycling amplitude, and adjusting switching frequency. The specific strategy depends on the predicted failure mechanism and system operating requirements.

[0227] The predictive control module also features a multi-level fault-tolerant control structure;

[0228] Construct a fault-tolerant control structure with multi-level protection to maintain system functionality under abnormal conditions:

[0229] Level 1 fault tolerance: Component-level redundancy;

[0230] Key components employ hardware redundancy design, such as:

[0231] N+1 redundant power supply module configuration;

[0232] Parallel power devices (with margin design);

[0233] Redundant sensors and signal processing channels.

[0234] When a component malfunction is detected, the system automatically switches to a backup resource and records the malfunction event.

[0235] Level 2 fault tolerance: control-level adaptation;

[0236] When faced with non-fatal anomalies, maintain system functionality by adjusting control strategies:

[0237] Functional degradation: Reduce output power or adjust performance parameters;

[0238] Parameter reconfiguration: Recalibrate control parameters or switch control modes;

[0239] Algorithm adjustment: Select a control algorithm with lower computational complexity but higher robustness.

[0240] Control-level adaptation is triggered via the Fault Detection and Isolation (FDI) system:

[0241] ;

[0242] in, This is the residual vector, representing the difference between the actual system output and the model's predicted output; This is the actual output vector, containing the system's output at time [time]. The actual measured values, such as physical quantities like voltage, current, and temperature; The model predicts the output vector, representing the output vector based on the system model at time [time value missing]. Predicted output value; The norm of the residual vector represents the magnitude of the residual and is used to quantify the degree of deviation between the actual output and the predicted output. This is the fault detection threshold. When the residual norm exceeds this value, anomaly detection is triggered. Its setting is based on the residual statistical characteristics during normal system operation. The detection window length represents the minimum time during which the residual exceeds the threshold, used to avoid false alarms caused by transient interference.

[0243] Level 3 fault tolerance: System-level refactoring;

[0244] In the face of severe anomalies, the system performs topology reconstruction and switching of operating modes:

[0245] Isolate the faulty area: Disconnect the faulty part by switching devices;

[0246] Adjust power flow path: Reconfigure energy transfer path;

[0247] Switching operating modes: such as switching from grid-connected mode to emergency off-grid mode.

[0248] System-level refactoring is triggered by a security monitor, based on state assessment results and a predefined decision table.

[0249] The predictive control module is also used for health status monitoring and assessment;

[0250] Real-time monitoring and assessment of the health status of key system components:

[0251] Definition of health indicators:

[0252] Define multidimensional health indicators for different components:

[0253] Power semiconductor devices: rate of change of on-resistance, rate of change of switching time, rate of increase of leakage current, rate of change of thermal resistance;

[0254] Capacitors: Equivalent series resistance (ESR) change rate, capacitance change rate, leakage current;

[0255] Magnetic components: rate of increase in core loss, rate of change in winding resistance;

[0256] Interconnection structure: contact resistance change rate, temperature distribution non-uniformity;

[0257] Health indicators are normalized and mapped to a range of 0-1, where 1 represents a completely new state and 0 represents complete failure.

[0258] Comprehensive health status assessment:

[0259] By weighted and fused health indicators of various components, the overall system health status index (HSI) is calculated:

[0260] ;

[0261] in, The overall health status index represents a comprehensive score of the overall health level of the system, with a value range of 0-1. For the first The health metric of a component represents the health status of a single component, with a value range of 0-1, where 1 represents complete health and 0 represents complete failure. Let be the weighting factor, representing the th The degree of influence of each component on the overall health of the system, typically ranging from 0 to 1, and satisfying the following conditions: ; This represents the total number of components monitored in the system. The component's index number, from 1 to... Weighting factors The determination is based on the importance, reliability, and impact on system functionality of the components.

[0262] Based on HSI, the system health status is divided into five levels: excellent (0.9-1.0), good (0.75-0.9), normal (0.6-0.75), caution (0.4-0.6) and warning (0-0.4).

[0263] Visualization and Reporting of Health Status:

[0264] The system provides an intuitive health status visualization interface, including:

[0265] Dashboard view: Displays key health indicators and HSI;

[0266] Trend chart: Shows the trend of health indicators over time;

[0267] Alert List: Displays triggered health alerts and recommended actions;

[0268] Lifetime prediction: Displays the estimated remaining lifespan of critical components.

[0269] It also generates health status reports regularly, recording key events and maintenance recommendations.

[0270] Through the above steps, the system achieves predictive control and health management based on digital twins, significantly improving reliability and service life in extreme environments.

[0271] The resonance suppression and EMI control module is used to suppress resonance and control electromagnetic interference for the high-speed switching characteristics of wide-bandgap devices in multi-mode redundant power supply systems.

[0272] Specifically, the resonance suppression and EMI control module includes:

[0273] The parameter acquisition unit is used to acquire key switch parameters.

[0274] The characteristic determination unit is used to obtain the switching characteristics of wide-bandgap devices in a multi-mode redundant power supply system based on key switching parameters.

[0275] Establish a measurement system for the switching characteristics of SiC and GaN devices to obtain key switching parameters:

[0276] Ascent Time and descent time ;

[0277] Rate of change of current and voltage change rate ;

[0278] Ringing frequency during switching process and damping characteristics ;

[0279] Analysis of the influence of parasitic parameters:

[0280] Establishing parasitic inductance in the power loop and parasitic capacitance Relationship with ringing characteristics:

[0281] ;

[0282] in, The ringing frequency during the switching process; This refers to the parasitic inductance in the power circuit. Let be the parasitic capacitance in the power circuit. This formula represents the natural oscillation frequency of the LC resonant circuit formed by the parasitic parameters.

[0283] ;

[0284] in, This refers to the voltage overshoot amplitude during the switching process; This refers to the parasitic inductance in the power circuit. Let be the rate of change of current during the switching process. This formula describes the voltage overshoot phenomenon caused by the interaction between parasitic inductance and rapid current change.

[0285] ;

[0286] in, This refers to the current overshoot amplitude during the switching process; This refers to the parasitic capacitance in the power circuit. Let be the rate of voltage change during the switching process. This formula describes the current overshoot phenomenon caused by the interaction between parasitic capacitance and rapid voltage changes.

[0287] Establish an EMI spectrum characteristic model:

[0288] ;

[0289] in, This refers to the voltage overshoot amplitude during the switching process; This refers to the parasitic inductance in the power circuit. Let be the rate of change of current. This formula represents the voltage spike caused by the interaction between parasitic inductance and rapid current change.

[0290] ;

[0291] in, This refers to the current overshoot amplitude during the switching process; This refers to the parasitic capacitance in the power circuit. Let be the rate of change of voltage. This formula represents the current spikes caused by the interaction between parasitic capacitance and rapid voltage changes.

[0292] Ringing suppression factor analysis:

[0293] Damping ratio Relationship with ringing attenuation:

[0294] ;

[0295] in, This represents the change in ringing voltage over time. This is the initial ringing peak voltage; The damping ratio; It is the undamped natural angular frequency; The frequency is damped. For time; This represents the cosine function. This formula describes the decay process of the ringing voltage.

[0296] The correction unit is used to obtain a drive signal correction strategy based on the switching characteristics of the wide bandgap device.

[0297] To achieve high switching speeds for wide-bandgap devices, a leading edge trimming technique for the drive signal was developed. and Precise control:

[0298] Active gate drive technology:

[0299] Design a gate drive circuit with adjustable impedance to achieve precise control over the switching process:

[0300] The turn-on process is divided into three stages: a pre-charge stage, a Miller plateau stage, and a full-conduction stage. This is achieved by adjusting the gate resistance at each stage. control;

[0301] The turn-off process is controlled in stages: it is divided into an initial descent stage, a Miller plateau stage, and a complete turn-off stage, which is achieved by adjusting the gate resistance in each stage. control.

[0302] Driven waveform trimming algorithm:

[0303] Develop a drive waveform trimming algorithm to dynamically adjust the drive signal according to operating conditions:

[0304] ;

[0305] in, The gate-source voltage is a function of time. The nominal gate voltage is the set target gate voltage value. It is a time variable; is the base of the natural logarithm; To compensate for the time constant, the rate of rise of the gate voltage is determined and dynamically adjusted according to operating conditions:

[0306] ;

[0307] in, To compensate for the time constant; The reference time constant represents the reference value of the time constant under standard conditions; This is the current compensation coefficient, used to adjust the degree of influence of current changes on the time constant; This is the temperature compensation coefficient, used to adjust the degree of influence of temperature changes on the time constant; This is the actual load current; The rated current of the device indicates the maximum current that the device can continuously withstand. This is the actual junction temperature of the device, representing the actual operating temperature inside the semiconductor chip; This is a reference temperature, typically room temperature or standard operating temperature. The maximum permissible junction temperature represents the highest temperature limit at which the device can operate safely. This formula dynamically adjusts the compensation time constant based on real-time changes in load current and junction temperature, achieving precise control of the gate drive signal.

[0308] Adaptive drive strength control:

[0309] Achieve adaptive drive intensity control based on switching state and operating conditions:

[0310] Continuity strength control:

[0311] ;

[0312] in, This represents the gate resistance value during the conduction process; The reference gate resistor is used to turn on the circuit, representing the basic gate resistance value set under standard conditions. This is the load current; This refers to the junction temperature of the device. An adaptive adjustment function is used for the conduction process, which dynamically adjusts the gate resistance value based on current and temperature.

[0313] Shutdown intensity control:

[0314] ;

[0315] in, This represents the gate resistance value during the turn-off process. The reference gate resistor is used for shutdown, representing the basic gate resistance value set under standard conditions. This is the load current; This refers to the junction temperature of the device. This is an adaptive adjustment function for the turn-off process, which dynamically adjusts the gate resistance value based on current and temperature.

[0316] Adaptive function and This is typically achieved through table lookup or polynomial fitting. These functions calculate the optimal gate resistance adjustment factor based on the real-time load current and device junction temperature to balance the relationship between switching speed, losses, and EMI.

[0317] By employing a leading-edge trimming technique for the drive signal, the system can effectively control the signal while maintaining efficient switching of wide-bandgap devices. and This reduces resonance and EMI.

[0318] The timing control unit is used to control the switching timing according to the soft-switching timing optimization algorithm.

[0319] By precisely controlling the switching timing, switching losses are reduced while resonance is suppressed.

[0320] Zero-voltage switching (ZVS) condition optimization:

[0321] Establish a ZVS conditional analysis model to ensure that the following conditions are met:

[0322] ;

[0323] in, The energy stored in the parasitic inductance; The parasitic inductance in the circuit; The current required to achieve zero-voltage switching; The minimum energy required to achieve zero-voltage switching; The equivalent output capacitance includes the sum of the device's output capacitance and the circuit's parasitic capacitance. The bus voltage is given. This formula indicates that, to achieve zero-voltage switching, the energy stored in the parasitic inductance must be greater than or equal to the energy required to charge and discharge the equivalent capacitance.

[0324] Adaptive adjustment of dead time based on different operating points :

[0325] ;

[0326] in, Dead time refers to the delay between the turn-off and turn-on of a power device, used to prevent shoot-through short circuits. Optimal dead time represents the best dead time value calculated under specific working conditions; The reference dead time represents the basic dead time value set under standard conditions. This is the load adjustment factor, used to adjust the dead time according to the load conditions; This refers to the bus voltage. This is the load current; Equivalent output capacitance represents the combined equivalent value of the output capacitance of the power device and the parasitic capacitance in the circuit.

[0327] Calculate the optimal switching timing:

[0328] Calculate the valley bottom detection time With switching time Optimal delay between :

[0329] ;

[0330] in, The optimal delay time represents the best delay time between the valley detection time and the switching time. The maximum rate of change of current represents the maximum allowable rate of change of current in the circuit. Maximum voltage change rate represents the maximum allowable rate of voltage change in the circuit; Parasitic inductance in a circuit represents the unavoidable parasitic inductance component in a circuit. Parasitic capacitance in a circuit represents the unavoidable parasitic capacitance component in a circuit. The resonant frequency represents the frequency at which ringing occurs in the circuit. The damping coefficient represents the system's ability to suppress oscillations; a larger value indicates a stronger ability to suppress oscillations. This formula indicates that the optimal delay time is a function of these parameters, and the optimal switching timing is determined by comprehensively considering these factors.

[0331] Multivariate collaborative optimization algorithm:

[0332] Construct a virtual switch trajectory model based on the operating point:

[0333] Jointly optimize switching timing and drive parameters to minimize:

[0334] ;

[0335] in, The objective function for the switch optimization represents the overall performance index that needs to be minimized; This is the switching loss weighting coefficient, used to adjust the importance of switching loss in the objective function; Switching energy loss represents the energy loss generated by power devices during the switching process; This is the resonant energy weighting coefficient, used to adjust the importance of resonant energy in the objective function; Resonant energy represents the energy contained in the ringing phenomenon generated during the switching process of a circuit; This is the EMI peak weighting coefficient, used to adjust the importance of the EMI peak in the objective function; The EMI peak value represents the maximum amplitude in the electromagnetic interference spectrum. This objective function comprehensively considers three key indicators—switching losses, resonant energy, and EMI peak value—through a weighted summation method, achieving multi-objective synergistic optimization.

[0336] The suppression unit is used to suppress EMI in various preset frequency bands based on a pre-defined multi-stage EMI filter network and electromagnetic shielding structure. The multi-stage EMI filter network and electromagnetic shielding structure are designed to effectively suppress EMI in various frequency bands.

[0337] Multi-stage filter network design:

[0338] Low frequency band (150kHz-1MHz): Uses large inductors ( , ) and capacitor ( , An LC filter composed of )

[0339] Mid-frequency band (1MHz-30MHz): Utilizing a damped RC network and ferrite beads ( Provides frequency-dependent impedance;

[0340] High frequency band (>30MHz): High frequency impedance matching and distributed parameter filtering structure are adopted.

[0341] Electromagnetic shielding structure optimization:

[0342] Targeted shielding of critical components, shielding effectiveness satisfy:

[0343] ;

[0344] in, Absorption loss represents the degree to which electromagnetic waves are absorbed and converted into heat energy in the shielding material, thus attenuating them. Reflection loss represents the degree of attenuation of electromagnetic waves reflected off the surface of a shielding material and preventing them from entering the material's interior. The multiple reflection loss represents the additional attenuation effect caused by multiple reflections of electromagnetic waves inside the shielding material. Total shielding effectiveness refers to the overall attenuation capability of the shielding structure against electromagnetic interference signals, and it is required to be no less than 40dB to ensure effective suppression of EMI.

[0345] Grounding layout optimization:

[0346] Multi-point low-impedance grounding design to ensure grounding impedance satisfy:

[0347] ;

[0348] in, The absolute value of the grounding impedance is expressed at a frequency of [frequency value missing]. The impedance of the grounding network; The maximum permissible noise voltage represents the maximum acceptable noise voltage level in the system. frequency The maximum common-mode current represents the maximum common-mode interference current flowing through the grounding system at a specific frequency. This formula indicates that the grounding impedance must be less than the ratio of noise voltage to common-mode current to ensure that the voltage drop generated when the common-mode current flows through the grounding system does not exceed the maximum allowable noise voltage level of the system, thereby effectively suppressing electromagnetic interference.

[0349] The adjustment unit is used to dynamically adjust the suppression strategy based on the preset adaptive resonance suppression algorithm and the real-time state of the system.

[0350] Develop an adaptive resonance suppression algorithm to dynamically adjust the suppression strategy based on the real-time system state:

[0351] Real-time monitoring of resonance characteristics:

[0352] Extracting ringing signals in characteristic frequency bands using a bandpass filter bank:

[0353] ;

[0354] in, The extracted ringing signal represents the resonant ringing waveform obtained after bandpass filtering; Let i be the i-th bandpass filter, used to extract signal components within a specific frequency range; The original voltage waveform of the switching node represents the voltage signal measured during the switching process of the power device; The time variable is used. This formula represents the process of processing the switching node voltage using a bandpass filter bank to extract the resonant ringing signal in a specific frequency band, which is then used for subsequent resonance characteristic analysis and suppression algorithms.

[0355] Calculate the resonant energy index:

[0356] ;

[0357] in, Resonant energy represents the energy contained in the ringing phenomenon generated during the switching process of a circuit; The starting time point indicates the moment when the measurement of the resonant energy begins; The resonant period represents the length of the time window for measuring the resonant energy. The resonant voltage represents the ringing signal in the characteristic frequency band extracted by the bandpass filter; Let be a time infinitesimal element, representing the time increment during the integration process. This formula calculates the energy during the resonance process by integrating the square of the ringing voltage over one resonant period, and is used to evaluate the resonant state of the system.

[0358] Suppression parameters are adaptively adjusted:

[0359] Based on the resonant energy index, the driving parameter vector is dynamically adjusted. :

[0360] ;

[0361] in, This is the driving parameter vector for the next time step, representing the set of driving parameters after adaptive adjustment. This is the current driving parameter vector, representing the set of driving parameters currently in use; The adaptive step size is used to control the magnitude of parameter updates. A larger step size makes the adjustment faster but may lead to instability, while a smaller step size makes the adjustment smoother but converges more slowly. Let be the gradient of the resonant energy with respect to the parameters, representing the resonant energy. For the driving parameter vector The partial derivatives indicate the direction of parameter adjustment to minimize the resonant energy.

[0362] Switching between multi-mode resonance suppression strategies:

[0363] Switching between different suppression strategies based on the operating point and resonance characteristics:

[0364] Mode 1: Primarily for leading edge trimming, suitable for high-power operating points;

[0365] Mode 2: Primarily soft switching, suitable for light-load operating points;

[0366] Mode 3: Hybrid suppression, suitable for medium load operating points;

[0367] Closed-loop feedback control:

[0368] Construct a closed-loop system for resonance suppression, with feedback control law:

[0369] ;

[0370] in, Let be the resonance suppression control output vector, representing the resonance suppression control signal output in the k-th control cycle; This is the feedback control gain matrix, used to adjust the strength of the feedback control. Let be the resonance error vector, representing the deviation between the actual resonance state and the desired resonance state; This is the feedforward control gain matrix, used to adjust the strength of the feedforward control; Let be the reference signal vector, and let represent the desired target value of the resonant state. This control law combines feedback control and feedforward control, achieving precise suppression of the resonance phenomenon by adjusting the gain matrix.

[0371] Real-world application examples of this implementation method:

[0372] To verify the feasibility and effectiveness of this implementation method, a practical application example in the power system of a deep space probe is given below.

[0373] This implementation method is applied to the redundant power supply system of a deep space probe. The system needs to operate stably in extreme temperature environments ranging from -150°C to 200°C, radiation doses up to 70 krad, and frequently changing load requirements. It also requires extremely high reliability and adaptability, while strictly controlling electromagnetic interference.

[0374] The power system configuration is as follows:

[0375] Input power: Solar array (200 to 400V), lithium-ion battery (100V rated);

[0376] Output requirements: +28V / 20A (main output), +5V / 10A (logic power supply), + / -15V / 5A (analog power supply);

[0377] Power rating: 1kW (peak power up to 1.5kW);

[0378] Efficiency requirement: >92% (full load range);

[0379] Service life: >10 years (without manual maintenance);

[0380] EMI requirement: Compliant with MIL-STD-461G standard.

[0381] Implementation process example:

[0382] Multi-level topology fusion and adaptive switching implementation:

[0383] This application employs three basic topologies: full-bridge (suitable for heavy loads), half-bridge (suitable for medium loads), and quasi-resonant (suitable for light loads). Based on measured data, a complete efficiency mapping table has been established. The following data demonstrates the efficiency performance of different topologies under specific conditions:

[0384] Efficiency of each topology when the load power is 200W (20% of rated power):

[0385] Overall bridge coverage: 83.5%;

[0386] Half-bridge: 91.2%;

[0387] Quasi-resonance: 96.8%.

[0388] Efficiency of each topology when the load power is 500W (50% of rated power):

[0389] Overall bridge coverage: 95.2%;

[0390] Half-bridge: 94.8%;

[0391] Quasi-resonance: 91.5%.

[0392] Efficiency of each topology when the load power is 1000W (100% rated power):

[0393] Overall bridge: 97.9%;

[0394] Half-bridge: 93.1%;

[0395] Quasi-resonance: 85.2%.

[0396] In actual operation, the system uses the ARIMA model to predict short-term load changes with a prediction accuracy of 94.7%. Based on the prediction results, the system initiates topology switching preparation 1.5 seconds in advance to ensure a seamless transition to the optimal topology during load changes. In a typical probe mission cycle (24 hours), the system achieved 21 disturbance-free topology switches, with an average efficiency of 95.8% over the entire cycle.

[0397] Temperature adaptive control of wide bandgap devices is achieved:

[0398] This application uses a SiC MOSFET (1200V / 50A) as the main power device. Through temperature characteristic testing, a complete characteristic model of the device was established within the range of -150℃ to 200℃. For example, the measured data of the device's on-resistance Rds,on as a function of temperature are as follows:

[0399] -150℃: 12mΩ;

[0400] -50℃: 18mΩ;

[0401] 25℃: 25mΩ;

[0402] 100℃: 45mΩ;

[0403] 200℃: 82mΩ;

[0404] Based on this data, the system established a temperature characteristic model of the device and dynamically adjusted the control parameters under different temperature conditions. For example, at a low temperature of -100℃, the system adjusted the gate drive voltage from the standard +20V / -5V to +18V / -3V and reduced the dead time from 200ns to 120ns; while at a high temperature of 175℃, the system reduced the switching frequency from 100kHz to 65kHz and increased the gate resistance from 2Ω to 5Ω.

[0405] Through multi-point temperature monitoring and Kalman filtering algorithm, the system achieved a junction temperature estimation accuracy of ±3℃. In a simulated temperature cycling test (-120℃ to 180℃, 5℃ / min), the system successfully maintained all key parameters within safe ranges, and the conversion efficiency fluctuation did not exceed 2.5%.

[0406] Digital twin model and predictive control implementation:

[0407] The system constructs a digital twin model encompassing four physical domains: electrical, thermal, radiation, and mechanical. Through training with ground-based test data and continuous updates with on-orbit data, the model's prediction accuracy has reached: voltage prediction error <1%, temperature prediction error <5%, and lifetime prediction error <12%.

[0408] In a simulation of a typical solar flare event, the system detected an impending increase in radiation levels. The digital twin model predicted potential changes in device parameters over the next 72 hours and generated an optimal control strategy: reducing the switching frequency by 20%, increasing the dead time by 15%, and increasing the participation of redundant units. After implementing this strategy, the system maintained 93.2% of its rated performance when the accumulated radiation dose reached 50 krad, while the performance of the comparison system using traditional fixed-parameter control decreased by 23.5%.

[0409] Predictive health management is achieved by:

[0410] The system conducted comprehensive health monitoring of key components and established a comprehensive assessment model including 45 health indicators. Through several weeks of data collection, the system learned the characteristic distribution under normal conditions and can detect minor abnormal deviations.

[0411] In a simulated power MOSFET aging test, the system detected an abnormal increase in the device's on-resistance 48 hours before the conventional failure threshold was triggered, predicting that the device might reach the failure threshold within 5 days.

[0412] The system automatically activated fault tolerance measures:

[0413] Transfer part of the load to redundant units;

[0414] Reduce the power level of the affected units;

[0415] Adjusting drive parameters slows down the aging process.

[0416] These measures extended the anticipated 5-day failure window to 15 days, providing the ground team with ample response time while maintaining full system functionality. The Health Status Index (HSI) only decreased from 0.92 to 0.85, well above the warning threshold of 0.6.

[0417] EMI control implementation:

[0418] To address the stringent EMI requirements of deep space environments, the system developed adaptive leading-edge control and multi-stage filtering techniques. In resonant testing, the switching time of the SiC device was extended from the typical 12ns to 35ns, while leading-edge trimming technology was applied to reduce dv / dt from the original 35V / ns to 12V / ns, resulting in a 73% reduction in EMI emissions.

[0419] By optimizing the PCB layout and implementing multi-layer shielding design, the system's radiated EMI is 15 dB lower than the MIL-STD-461G standard requirement in the 150 kHz-30 MHz band and 10 dB lower in the 30 MHz-1 GHz band. Simultaneously, the voltage ringing amplitude during switching is reduced from the original 180 V to 45 V, effectively protecting the devices from voltage stress damage.

[0420] Technical effectiveness verification:

[0421] Efficiency and temperature adaptability verification:

[0422] The efficiency test results of the system under different temperature and load conditions are shown in Table 1:

[0423] Table 1: Efficiency test results of the system under different temperature and load conditions;

[0424]

[0425] It is evident that the system maintains excellent efficiency performance across an extreme temperature range (-120℃ to 175℃), with a maximum efficiency of 98.2%. Even under conditions of -120℃ and 175℃, the full-load efficiency remains at 97.8% and 96.2%, respectively. More significantly, under light load conditions (10% load), the system still maintains an efficiency of over 91%, far exceeding the typical efficiency (70%-80%) of traditional single-topology systems under light load.

[0426] Reliability and health management effectiveness verification:

[0427] The system underwent accelerated aging testing, simulating 10 years of on-orbit operation. The test results are shown in Table 2.

[0428] Table 2: Comparison of 10-year accelerated aging test results between the conventional system and the system of this embodiment;

[0429]

[0430] During 10 years of simulated operation, this implementation predicted and prevented 23 potential failure events, while the conventional system experienced 8 actual failures and 17 performance degradations under the same conditions. Particularly noteworthy is that in a simulated high-radiation environment (total dose 70 krad), the performance degradation rate of this system was only 8.5%, while that of the conventional system was 32%, demonstrating the superior adaptability and reliability of this solution in extreme environments.

[0431] The above technical performance verification results fully demonstrate the high efficiency, high reliability and strong adaptability of this implementation method in extreme environments, meeting the stringent requirements of high-end applications such as deep space exploration.

[0432] The embodiments of this application have been described above, but these embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments based on the guidance of these embodiments, and all of them are within the protection scope of these embodiments.

Claims

1. A multi-mode redundant power supply system based on a bidirectional inverter power supply board, characterized in that, include: A topology acquisition module is used to acquire a reconfigurable circuit topology; wherein, the reconfigurable circuit topology includes power semiconductor switches, magnetic components, filter capacitors, and driving circuits; the reconfigurable circuit topology can be switched to a full-bridge topology, a half-bridge topology, a three-level topology, or a hybrid topology through different combinations of devices; The optimal structure generation module is used to generate the optimal topology based on the reconfigurable circuit topology and the multi-parameter real-time monitoring system. The parameter acquisition module is used to dynamically adjust control parameters under different temperature conditions based on the temperature characteristic model library. The predicted state acquisition module is used to obtain the predicted state of the multi-mode redundant power system in the future time period based on the preset digital twin model; wherein, the preset digital twin model is trained based on ground test data and updated with on-orbit data. The predictive control module is used to control the multi-mode redundant power supply system based on the predicted state.

2. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 1, characterized in that, The optimal structure generation module includes: The mapping table generation unit is used to generate a four-dimensional topology efficiency mapping table based on the efficiency characteristics of different topologies under different input voltage, output voltage, load current and temperature conditions. The demand acquisition unit is used to collect load parameters from the multi-parameter real-time monitoring system to obtain the current load status, and to obtain the predicted power demand within a preset time period based on historical load data. The optimal structure generation unit is used to generate the optimal topology based on the current load state, predicted power demand, and a four-dimensional topology efficiency mapping table.

3. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 2, characterized in that, The parameter acquisition module includes: The model library acquisition unit is used to acquire a temperature characteristic model library; wherein, the temperature characteristic model library includes switching time temperature model, on-resistance temperature model, breakdown voltage temperature model and threshold voltage temperature model; Junction temperature unit, used to estimate the device junction temperature based on the thermal resistance network model; The parameter adjustment unit is used to dynamically adjust the device control parameters based on the estimated device junction temperature. The management unit is used to perform heat dissipation management based on multi-point temperature monitoring for modules with power exceeding the preset power.

4. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 3, characterized in that, The predicted state acquisition module includes: The model determination unit is used to obtain a multi-physics coupled model that includes the electrical domain, thermal domain, radiation domain, and mechanical stress domain, and describes the state changes and interaction relationships of each physical quantity through a set of differential equations; The calibration unit is used to perform online calibration of the multiphysics coupling model according to the recursive parameter identification algorithm to obtain the target digital twin model; The prediction unit is used to obtain the predicted state of the multi-mode redundant power system in the future time period based on the target digital twin model.

5. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 4, characterized in that, Following the prediction unit, the prediction state acquisition module further includes: A building unit is used to construct a visualization interface for the target digital twin model; the visualization interface is used to display the state distribution of each physical domain.

6. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 4, characterized in that, The coupling relationships between the physical domains in the multiphysics coupling model include: between the electrical domain and the thermal domain: electrical losses act as a heat source; between the thermal domain and the electrical domain: temperature affects the electrical characteristics of the device; between the radiation domain and the electrical domain: gray damage affects electrical parameters; between the thermal domain and the mechanical domain: temperature gradients cause thermal stress; between the mechanical domain and the electrical domain: stress affects the contact resistance of the device and the reliability of the interconnection.

7. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 4, characterized in that, Following the prediction unit, the prediction state acquisition module further includes: Error compensation unit: Corrects the predicted state obtained from the target digital twin model according to the error model; wherein, the error model is used to describe the error between the actual system response and the predicted state obtained from the target digital twin model; and describes the error evolution characteristics according to the autoregressive moving average model.

8. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 4, characterized in that, The predictive control module includes: The predictive control unit is used to control the multi-mode redundant power supply system according to the predictive control algorithm. The predictive control algorithm uses rolling time-domain prediction for control and is based on the control objective function, which includes efficiency control, thermal management control, and reliability control.

9. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 4, characterized in that, The system also includes: The resonance suppression and EMI control module is used to achieve resonance suppression and electromagnetic interference control for the high-speed switching characteristics of wide-bandgap devices in multi-mode redundant power supply systems.

10. The multi-mode redundant power supply system based on a bidirectional inverter power board according to claim 9, characterized in that, The resonance suppression and EMI control module includes: The parameter acquisition unit is used to acquire key switch parameters; The characteristic determination unit is used to obtain the switching characteristics of wide-bandgap devices in a multi-mode redundant power supply system based on key switching parameters. The correction unit is used to obtain a drive signal correction strategy based on the switching characteristics of the wide bandgap device; The timing control unit is used to control the switching timing according to the soft-switching timing optimization algorithm; The suppression unit is used to suppress EMI in each preset frequency band according to the preset multi-level EMI filter network and electromagnetic shielding structure; The adjustment unit is used to dynamically adjust the suppression strategy based on the preset adaptive resonance suppression algorithm and the real-time state of the system.

Images