Time-triggered an-thermal-magnetic coordinated management of multi-source power system

Abstract

This invention discloses a multi-source power system based on time-triggered safety-thermal-magnetic coordinated management. The system includes: a cloud-edge collaborative management layer, an FPGA collaborative management hub, a time-triggered bus network, a heterogeneous power conversion unit cluster, a global sensing node cluster, and a multi-level DC power bus. The cloud-edge collaborative management layer is connected to the FPGA collaborative management hub for setting operating modes and issuing scheduling targets. The FPGA collaborative management hub connects to the heterogeneous power conversion unit cluster and the global sensing node cluster via the time-triggered bus network, enabling deterministic communication. The FPGA collaborative management hub integrates a safety-thermal-magnetic collaborative decision-making unit, which generates control commands such as power derating, path reconstruction, frequency offset, phase rearrangement, active heat dissipation, and fault isolation based on the rule of "safety first, thermal management second, and electromagnetic optimization last." This system possesses safety, thermal, and magnetic coordinated management capabilities, improving the reliability, adaptability, and compatibility of the multi-source power system.

Description

Technical Field

[0001] This invention relates to the field of power electronics systems technology, specifically to a multi-source power system based on time-triggered safety-thermal-magnetic coordinated management. This system can be applied to modular power routers, distributed energy storage systems, electric vehicle charging and swapping stations, data center power supply systems, ship integrated power systems, aviation ground power supply systems, and other multi-port power conversion applications requiring high reliability, high power density, and high electromagnetic compatibility. Background Technology

[0002] With the rapid development of new energy power generation, energy storage, electric transportation, and data center power supply systems, power electronic power conversion systems are characterized by multi-port, multi-voltage level, multi-power flow direction, and high-density integration. A typical system may simultaneously contain various heterogeneous power conversion units, such as three-phase Vienna rectifier units, single-phase power factor correction units, LLC resonant converter units, phase-shifted full-bridge converter units, bidirectional DC / DC converter units, single-phase inverter units, three-phase inverter units, and energy storage interface units. These units are typically interconnected via high-voltage, medium-voltage, or low-voltage DC buses and work collaboratively within the same cabinet, compartment, or compact space.

[0003] However, high power density and multi-unit integration present significant thermal, safety, and electromagnetic coupling challenges to the system. First, in terms of thermal management, when multiple power conversion units operate simultaneously, localized hot spots can easily form at power devices, magnetic components, and busbar connections. Traditional independent fan control or single-point over-temperature protection can only handle localized anomalies and cannot achieve system-level thermal equilibrium, often resulting in a situation where "localized overheating occurs while the overall heat dissipation capacity remains redundant."

[0004] Secondly, in terms of safety management, multi-level DC busbars mean that insulation degradation, grounding faults, overcurrent, overvoltage, or arcing faults in a single unit can propagate to other units via the busbars. Existing protection methods mostly rely on local hardware protection or software polling, which can achieve rapid action locally, but lack coordination mechanisms across units and across busbars, easily leading to selective mismatch, cascading tripping, or expansion of the fault range.

[0005] Furthermore, regarding electromagnetic radiation management, when multiple high-frequency switching units operate at the same or similar switching frequencies, switching harmonics and near-field radiation may superimpose in specific frequency bands, thereby interfering with communication buses, sensor nodes, and nearby sensitive electronic equipment. Traditional shielding, filtering, and grounding measures are passive suppression methods, which suffer from problems such as large size, high cost, and difficulty in adapting to dynamic operating conditions.

[0006] The core idea of ​​time-triggered protocols is to divide communication time into defined periods and time slots, with each node sending data only within its allocated time slot, thereby achieving deterministic latency, predictable bandwidth, and good fault tolerance. Existing research has attempted to apply the time-triggered approach to communication scheduling in power conversion systems, but most schemes remain at the level of addressing communication latency uncertainties or simple state acquisition, failing to integrate thermal management, safety protection, and electromagnetic radiation control into the same deterministic decision-making framework. When conflicts arise between different objectives, such as changing the switching frequency to reduce radiation but increasing switching losses and exacerbating thermal load, existing systems often lack clear priorities and collaborative arbitration mechanisms.

[0007] Therefore, it is necessary to propose a time-triggered, safety-thermal-magnetic coordinated management multi-source power system that completes multi-source sensing, state fusion, priority arbitration, and instruction execution within a unified bus cycle, thereby improving the reliability, predictability, and electromagnetic compatibility of complex power conversion systems. Summary of the Invention

[0008] The purpose of this invention is to provide a multi-source power system based on time-triggered safety-thermal-magnetic coordinated management, which solves the problems of disconnected safety management, thermal management and electromagnetic radiation management, conflicting control objectives and insufficient system-level coordination capabilities in existing multi-unit power conversion systems.

[0009] To achieve the above objectives, the specific technical solution adopted by the present invention is as follows:

[0010] A multi-source power system based on time-triggered safety-thermal-magnetic collaborative management includes: a cloud-edge collaborative management layer, an FPGA collaborative management hub, a time-triggered bus network, a heterogeneous power conversion unit cluster, a global sensing node cluster, a high-voltage DC power bus, a medium-voltage DC power bus, and a low-voltage DC power bus.

[0011] The cloud-edge collaborative management layer is connected to the FPGA collaborative management hub, used to set the system operation mode, issue energy scheduling targets, and receive system status feedback. The FPGA collaborative management hub is connected to the heterogeneous power conversion unit cluster and the global sensing node cluster through a time-triggered bus network, and is configured to generate deterministic time slot scheduling, uniformly receive, fuse, and prioritize safety parameters, thermal parameters, and electromagnetic parameters, and issue fault isolation, power derating, load migration, active heat dissipation, electromagnetic suppression, or time slot phase rearrangement commands to the heterogeneous power conversion unit cluster. The time-triggered bus network adopts a time-division multiple access mechanism based on a static scheduling table, allocating an independent transmission time slot to each connected node. The heterogeneous power conversion unit cluster consists of multiple power conversion units, each connected to a high-voltage DC power bus, a medium-voltage DC power bus, or a low-voltage DC power bus, to realize AC-DC, DC-DC, or DC-AC power conversion, and to perform power regulation, path switching, or fault isolation according to instructions issued by the FPGA collaborative management center. The global sensing node cluster consists of multiple independent sensor nodes, each of which uploads thermal parameters, electrical safety parameters, electromagnetic radiation parameters, and operating status parameters through the time-triggered bus network. The high-voltage DC power bus, the medium-voltage DC power bus, and the low-voltage DC power bus achieve energy transfer and distribution through the heterogeneous power conversion unit cluster.

[0012] Furthermore, the FPGA collaborative management hub includes: an energy routing strategy parsing unit, a time-triggered scheduling unit, a global clock synchronization unit, a static scheduling table storage unit, a bus monitoring and fault isolation unit, a status summary and uplink feedback unit, and an ampere-thermal-magnetic collaborative decision-making unit. The ampere-thermal-magnetic collaborative decision-making unit interacts with the energy routing strategy parsing unit, the time-triggered scheduling unit, the bus monitoring and fault isolation unit, and the status summary and uplink feedback unit to construct a state diagram based on the received multi-dimensional sensing parameters and generate collaborative control instructions according to preset nested priority rules.

[0013] Furthermore, the FPGA collaborative management hub is configured to divide the bus cycle into synchronization time slots, sensing data fusion time slots, collaborative arbitration time slots, instruction execution time slots, and protection interval time slots according to a static scheduling table. The synchronization time slot corresponds to the global clock synchronization function; the sensing data fusion time slot corresponds to the data uploading and normalization processing function of the global sensing node cluster; the collaborative arbitration time slot corresponds to the state diagram construction, threshold comparison, and priority arbitration function of the safety-thermal-magnetic collaborative decision-making unit; the instruction execution time slot corresponds to the collaborative control instruction issuance function; and the protection interval time slot corresponds to the bus commutation blanking, scheduling table double-buffer switching, and hard-wired bypass signal establishment functions.

[0014] Furthermore, the safety-thermal-magnetic collaborative decision-making unit is configured to perform arbitration according to nested priority rules, which include: a first priority of absolute safety priority, where an unconditional unit isolation command is triggered when the insulation impedance of any conversion unit is lower than a danger threshold, or when continuous overcurrent, overvoltage, grounding fault, or arcing fault characteristics are detected; a second priority of thermal management, where, in the absence of a first priority event, if the junction temperature of the power device of any conversion unit exceeds a first thermal threshold but is lower than a second thermal threshold, a power reduction, load migration, or active heat dissipation command is triggered; if it exceeds the second thermal threshold, a shutdown command is triggered; and a third priority of electromagnetic radiation management, where, in the absence of the first and second priority events, if the radiation intensity of any conversion unit in a specific frequency band exceeds the electromagnetic compatibility limit, a switching frequency offset, frequency jitter, or time slot phase rearrangement command is triggered.

[0015] Furthermore, the FPGA collaborative management hub also includes a dedicated hard-wired bypass signal independent of the time-triggered bus network (3), which is used to bypass the bus scheduling cycle and directly force the disconnection of the drive enable or power circuit of the fault conversion unit when a safety absolute priority event is triggered, while broadcasting a fault isolation frame through the time-triggered bus network (3).

[0016] Furthermore, the safety-thermal-magnetic collaborative decision-making unit also maintains the health status table of each conversion unit in the heterogeneous power conversion unit cluster, and updates the health status value according to the number of thermal cycles, the highest junction temperature record, the number of switching cycles, the radiation exposure time, and the number of fault derating cycles; when allocating power tasks, the FPGA collaborative management center prioritizes assigning high-power, high-frequency tasks to units without abnormal working conditions.

[0017] Furthermore, the time-triggered bus network adopts a redundant dual-channel RS-485 bus, CAN-FD bus, or Ethernet physical layer, and time-division multiple access is implemented by FPGA hardware; the global sensing node cluster and the heterogeneous power conversion unit cluster share the same physical bus and are allocated different time slot sets, or adopt a dual-bus structure with separate control bus and sensing bus.

[0018] Furthermore, the global sensing node cluster includes infrared thermopile sensor nodes, power device junction temperature estimation sensor nodes, high-frequency current injection insulation monitoring sensor nodes, DC bus voltage and current sampling sensor nodes, near-field electromagnetic probe sensor nodes, and vibration sensor nodes; each sensor node has an independent local clock, node identifier, and time slot transmission enable signal, and transmits data frames only within the sensing data time slot allocated to it by the static scheduling table.

[0019] Furthermore, the heterogeneous power conversion unit cluster includes at least two of the following: a three-phase Vienna rectifier unit, a single-phase power factor correction unit, an LLC resonant converter unit, a phase-shifted full-bridge converter unit, a bidirectional DC / DC unit, a single-phase inverter unit, a three-phase inverter unit, and an energy storage interface unit; each unit is equipped with a local controller for performing local closed-loop control, while system-level start-up, mode switching, reference value update, derating, and isolation actions are uniformly scheduled by the FPGA collaborative management center through a time-triggered bus network.

[0020] Furthermore, the rated voltage of the high-voltage DC power bus is 600V, the rated voltage of the medium-voltage DC power bus is 48V, and the rated voltage of the low-voltage DC power bus is 24V. The rectifier unit in the heterogeneous power conversion unit cluster injects power into the high-voltage DC power bus, the LLC resonant converter unit realizes bidirectional conversion between the high-voltage DC power bus and the medium-voltage DC power bus, the phase-shifting full-bridge unit realizes bidirectional conversion between the medium-voltage DC power bus and the low-voltage DC power bus, and the inverter unit draws power from the low-voltage DC power bus and supplies power to the AC load.

[0021] The FPGA collaborative management hub is configured to load the initial scheduling table in the static scheduling table and establish a global synchronization clock through the time-triggered bus network.

[0022] Active heat dissipation commands are used to send speed, flow rate, or valve opening adjustment commands to intelligent air-cooled, liquid-cooled, or phase-change heat dissipation nodes before the local heat accumulation reaches the power degradation threshold, and combine them with the power timing adjustment of adjacent conversion units to reduce thermal coupling interference.

[0023] The switching frequency offset command shifts the switching frequency of the unit with excessive radiation upwards or downwards by a preset step size, centered on the original frequency, and simultaneously adjusts the output filter or control loop parameters. If the radiation still exceeds the limit after one offset, periodic frequency dithering modulation is used to disperse the spectral energy. The time slot phase rearrangement command is used to suppress the superposition of radiation peaks caused by synchronous switching of multiple units: the FPGA is configured to calculate the initial phase of the power switching action of each conversion unit. When the radiation peaks of two or more units are detected to overlap in time, the PWM synchronization phase or time slot trigger phase of the corresponding unit is adjusted within the protection interval time slot, so that the switching actions of each unit are staggered by a preset time interval.

[0024] Beneficial effects

[0025] 1) Integrate safety management, thermal management, and electromagnetic compatibility management into the same system architecture, and avoid multi-objective conflicts through deterministic time-triggered communication and FPGA hardware decisions.

[0026] 2) Static time slot allocation and hardware pipeline processing ensure that the uploading of sensing data, state fusion, arbitration judgment and instruction issuance have microsecond-level deterministic latency.

[0027] 3) The absolute safety priority strategy, combined with an independent hard-wired bypass signal, can quickly isolate events such as insulation faults, overcurrent, and arcing, preventing the spread of faults.

[0028] 4) Supports active heat dissipation and load migration for localized heat accumulation, improving the efficiency of heat dissipation resource utilization.

[0029] 5) Actively suppress electromagnetic radiation by switching frequency offset, frequency jitter and time slot phase rearrangement to adapt to dynamic operating conditions.

[0030] 6) Introduce health tables and lifespan-aware scheduling to extend the overall service life of the system. Attached Figure Description

[0031] Figure 1 This is a diagram of the overall system architecture of the present invention;

[0032] Figure 2 This is a structural diagram of the FPGA collaborative management hub of the present invention;

[0033] Figure 3 This is a diagram showing the time-triggered bus cycle and time slot structure of the present invention;

[0034] Figure 4 This is a flowchart of the safety-thermal-magnetic collaborative decision-making process of this invention;

[0035] Figure 5 This is a schematic diagram illustrating the effect of time slot phase rearrangement in a multi-unit radiation superposition scenario according to the present invention. Detailed Implementation

[0036] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0037] See Figure 1 This is a diagram illustrating the overall system architecture of the present invention, showing the connections between the cloud-edge collaborative management layer, the FPGA collaborative management hub, the time-triggered bus network, the heterogeneous power conversion unit cluster, the global sensing node cluster, and the multi-level DC energy bus. The system of the present invention consists of the cloud-edge collaborative management layer, the FPGA collaborative management hub, the time-triggered bus network, the heterogeneous power conversion unit cluster, the global sensing node cluster, and the multi-level DC energy bus.

[0038] The cloud-edge collaborative management layer runs on industrial control computers, edge computing devices, or host servers. It is used to set up modes such as grid connection, off-grid operation, energy storage charging and discharging, emergency power supply, economic operation, and maintenance. It also generates medium- and long-term energy routing strategies based on load forecasting, energy prices, or equipment health status. This layer connects to the FPGA collaborative management hub via Ethernet, industrial Ethernet, or other communication interfaces, but does not directly participate in microsecond-level real-time protection and arbitration.

[0039] See Figure 2 This diagram illustrates the FPGA collaborative management hub structure of the present invention, showing the data interaction relationships between the energy routing strategy parsing unit, time-triggered scheduling unit, global clock synchronization unit, static scheduling table storage unit, bus monitoring and fault isolation unit, status summary and uplink feedback unit, and safety-thermal-magnetic collaborative decision-making unit. The FPGA collaborative management hub is the real-time control core of the present invention, implemented using a mid-to-high-end FPGA or a system-on-a-chip containing FPGA programmable logic. Internally, the time-triggered scheduling unit generates bus cycles, time slot numbers, and transmission enable based on the static scheduling table; the global clock synchronization unit broadcasts synchronization frames to each node and receives node synchronization status; the bus monitoring and fault isolation unit monitors bus occupancy, frame verification, node timeouts, and abnormal retransmissions; and the safety-thermal-magnetic collaborative decision-making unit performs hardware pipeline calculations on multi-source sensing data and outputs collaborative control commands.

[0040] The time-triggered bus network can be implemented using a redundant dual-channel RS-485 bus as a low-cost solution, or CAN-FD, industrial Ethernet, or fiber optic links can be used depending on the system bandwidth requirements. Regardless of the physical layer used, the media access control layer is implemented by the FPGA using deterministic time-division multiple access according to a static scheduling table, ensuring that sensing data frames, control command frames, and status feedback frames do not preempt each other.

[0041] The heterogeneous power conversion unit cluster includes, but is not limited to, a three-phase Vienna rectifier unit, a single-phase power factor correction unit, an LLC resonant converter unit, a phase-shifted full-bridge converter unit, a bidirectional DC / DC unit, a single-phase inverter unit, a three-phase inverter unit, and an energy storage interface unit. Each converter unit is equipped with a local DSP, MCU, or digital controller to perform closed-loop control of voltage, current, power, or frequency; while system-level actions such as start-up, shutdown, mode switching, reference value updates, derating, and isolation are uniformly scheduled by the FPGA collaborative management center through time-triggered instructions.

[0042] The global sensing node cluster is set up independently of the power conversion unit and can be deployed near power device heat sinks, magnetic components, DC buses, weak insulation points, near switching transistors, near communication cables, and in critical airflow locations within the cabinet. Independently configured sensing nodes do not occupy local control resources of the conversion unit, and unified time slot management ensures that various parameters have comparable time bases.

[0043] The multi-stage DC power bus includes a high-voltage DC power bus (600V), a medium-voltage DC power bus (48V), and a low-voltage DC power bus (24V). Energy conversion and isolation are performed between different voltage levels through corresponding conversion units, and the system reconfigures the energy path according to the operating mode and fault state.

[0044] See Figure 3This diagram illustrates the time-triggered bus cycle and time slot structure of the present invention. Taking a 10.0 ms bus cycle as an example, it shows the timing relationship between the synchronization time slot, the sensing data fusion time slot, the cooperative arbitration time slot, the instruction execution time slot, and the protection interval time slot. The system of the present invention divides one bus cycle into synchronization time slots, sensing data fusion time slots, cooperative arbitration time slots, instruction execution time slots, and protection interval time slots. The bus cycle length is configured according to the number of nodes, communication bandwidth, and real-time requirements, preferably between 1 ms and 10 ms.

[0045] Synchronization slots are used by the FPGA collaborative management center to broadcast periodic synchronization frames. Each conversion unit and sensing node calibrates its local counters based on the synchronization frames. The synchronization frame contains a period number, a global timestamp, a scheduling table version number, and a checksum field. Through this mechanism, each node establishes a local slot boundary consistent with that of the FPGA collaborative management center.

[0046] The sensing data fusion time slot is used by the global sensing node cluster to upload multi-source data in a preset order. Each node's data frame adopts a fixed-length or variable-length format and includes at least a node identifier, unit identifier, sensor type, sampling timestamp, sampled value, status word, and CRC checksum. The FPGA collaborative management center performs frame verification, unit conversion, normalization, and outlier handling while receiving data.

[0047] The collaborative arbitration time slot is used by the safety-thermal-magnetic collaborative decision-making unit to perform hardware comparison and decision generation. Within this time slot, the safety-thermal-magnetic collaborative decision-making unit compares temperature, insulation impedance, bus voltage and current, overcurrent and overvoltage indicators, radiation field strength, and vibration amplitude with preset thresholds and trend criteria, and generates a decision vector according to nested priority rules.

[0048] The instruction execution slot is used to issue instructions to the heterogeneous power conversion unit cluster and heat dissipation nodes. Instruction types include power reference values, limiting coefficients, switching frequency offset values, frequency jitter parameters, PWM phase offset values, heat dissipation speed or flow commands, shutdown flags, bypass flags, fault isolation flags, and status query commands.

[0049] The guard interval time slot is used for bus commutation blanking, node status confirmation, hard-wired bypass signal establishment, schedule table double-buffer switching, and mode switching preparation. By setting the guard interval, bus contention, command overlap, or power loop surges are avoided during critical switching moments.

[0050] The system comprises M power conversion units, N sensing nodes, and P types of status parameters. During the k-th bus cycle, the FPGA collaborative management center maps the data uploaded by each sensing node into a state diagram. The time dimension of the state diagram is the bus cycle number or sampling timestamp; the spatial dimension is the conversion unit identifier and sensor placement location; and the parameter dimensions include temperature, insulation resistance, overcurrent flag, overvoltage flag, radiated field strength, vibration amplitude, heat dissipation status, and health status.

[0051] When constructing the state diagram, the FPGA collaborative management center first completes the data attribution based on node identifiers and unit identifiers; secondly, it performs spatiotemporal alignment based on timestamps and time slot numbers; thirdly, it removes or marks missing data, CRC error data, and data that exceeds the physical range; and finally, it normalizes parameters of different dimensions to a unified risk scale so that the safety-thermal-magnetic collaborative decision-making unit can perform parallel comparisons.

[0052] For temperature parameters, the system directly acquires the surface temperature of the heat sink, or estimates the junction temperature of the power device by combining current, switching frequency, heat dissipation conditions, and thermal resistance models. For safety parameters, the system acquires insulation impedance, leakage current, bus voltage, current change rate, and arc characteristics. For electromagnetic parameters, the system acquires near-field radiation intensity at different frequency bands and compares it with the corresponding electromagnetic compatibility limits.

[0053] See Figure 4 This is a flowchart illustrating the safety-thermal-magnetic collaborative decision-making process of the present invention, showing the logical flow of state diagram construction, priority arbitration, instruction generation, and instruction execution. In the system of this invention, the safety-thermal-magnetic collaborative decision-making unit adopts a three-layer nested priority arbitration rule of safety, thermal, and electromagnetic layers. The safety layer has the highest priority; any event that meets the safety hazard conditions will trigger isolation or shutdown actions first, and may interrupt the thermal management and electromagnetic optimization processes. Safety hazard conditions include insulation impedance below a danger threshold, bus voltage exceeding a protection threshold, overcurrent duration exceeding a set value, and detection of grounding fault or arc fault characteristics, etc.

[0054] In the absence of any safety hazards, the safety-thermal-magnetic collaborative decision-making unit enters the thermal management layer for judgment. If the junction temperature or estimated junction temperature of a certain converter unit exceeds the warning threshold but does not reach the protection threshold, thermal stress is reduced primarily through active cooling, power derating, or load migration; if the protection threshold is exceeded, the unit is triggered to shut down or exit the current energy path. The thermal management layer also determines whether to take measures in advance based on the temperature rise rate to prevent thermal inertia from causing subsequent overheating.

[0055] When both safety and thermal conditions are met, the safety-thermal-magnetic collaborative decision-making unit enters the electromagnetic radiation optimization layer. If a unit exceeds the radiation limit in a specific frequency band, the system first determines whether the frequency band is related to the switching frequency and its harmonics; if related, it issues a switching frequency offset or frequency jitter command; if the radiation peak is mainly caused by the superposition of synchronous switching actions of multiple units, it issues a time slot phase rearrangement command to stagger the switching actions.

[0056] When different control objectives conflict, the system prioritizes the objective with higher priority. For example, if a unit simultaneously experiences radiation exceeding limits and junction temperature approaching the protection threshold, the system will not prioritize frequency offset strategies that may increase losses, but will instead execute power reduction or load migration first; if an insulation hazard occurs, isolation will be executed directly without waiting for the completion of a normal bus command cycle.

[0057] Unit isolation commands are implemented using both bus commands and hard-wired bypass signals. Hard-wired bypass signals directly affect the drive enable, relays, contactors, or electronic switches of the faulty unit; bus isolation frames are used to notify other nodes to update their status tables and scheduling tables, preventing further task assignment to the faulty unit.

[0058] The power derating command multiplies the target unit's power reference value by a derating factor. The derating factor is dynamically calculated based on the temperature overrun, the rate of temperature rise, and the current load demand. If the temperature continues to rise after derating, the power reference value is further reduced or a shutdown is triggered.

[0059] When redundant conversion units or alternative energy paths exist, the load migration command migrates part or all of the load from high-risk units to low-risk units. The migration process adopts a gradual strategy of first increasing the power of the target unit and then decreasing the power of the original unit, and completes the double-buffered switching of the scheduling table within the protection interval time slot.

[0060] Active cooling commands are used to control air cooling, liquid cooling, or other heat dissipation nodes. The system selects the corresponding air duct or cooling branch based on the hotspot location and adjusts the power timing of adjacent units in conjunction to reduce thermal coupling.

[0061] Frequency shift and frequency jitter commands are used for electromagnetic radiation suppression. Frequency shift causes major harmonics to avoid sensitive frequency bands, while frequency jitter disperses concentrated spectral energy. To avoid degrading output power quality, the system synchronously updates filter compensation parameters or control loop parameters when issuing frequency change commands.

[0062] The time-slot phase reordering command is used to adjust the PWM carrier phase or trigger phase of multiple converter units, causing their switching actions to be staggered in time. (See also...) Figure 5 The diagram shows the effect of the time slot phase rearrangement command, illustrating the phase adjustment of multiple converter units after EMC exceedance. This method does not change the average output power and is suitable for handling the problem of superimposed radiation peaks from multiple units.

[0063] The FPGA collaborative management center maintains the health status table for each conversion unit. The health status is updated incrementally based on parameters such as thermal cycle count, highest junction temperature record, over-temperature duration, high-current switching count, cumulative radiation exceeding limits, fault derating count, and total operating time. The initial health status value is set to 100. When it falls below the first health threshold, the system reduces the priority of that unit for high-power tasks; when it falls below the second health threshold, the system prompts for maintenance or replacement.

[0064] During the energy routing task allocation process, the An-thermal-magnetic collaborative decision-making unit considers not only the current load and topology reachability, but also differences in health status. For units with higher health status, high-power or frequent dynamic tasks are prioritized; for units with lower health status, light-load, standby, or low-frequency tasks are prioritized. This method avoids a unit from being subjected to high thermal stress or high electromagnetic exposure for extended periods, thereby extending the overall system lifespan.

[0065] Example 1: Load migration in a heat accumulation scenario

[0066] In a 48V power supply system for a data center, two LLC resonant converter units are connected in parallel to supply power to the low-voltage DC bus, each with a rated power of 1kW. During operation, the global sensing node detected that the heatsink temperature of unit A rose to 92℃, estimating the junction temperature to be close to the warning threshold, while the junction temperature of unit B was 78℃, and the system did not detect any safety faults or radiation exceeding limits. The FPGA collaborative management center determined this to be a thermal management event within the collaborative arbitration time slot, issued a power derating command to unit A, reducing its output power from 1kW to 0.7kW, and simultaneously issued a power boost command to unit B, migrating the 300W load to unit B. After two bus cycles, the temperature of unit A decreased, the temperature of unit B increased slightly, and the bus voltage fluctuation remained within the allowable range.

[0067] Example 2: Rapid Isolation of Insulation Faults

[0068] During the operation of the three-phase Vienna rectifier unit, the insulation monitoring node detected a rapid drop in DC-side insulation impedance from its normal value to below the danger threshold. The sensing node uploaded this data within its dedicated time slot, and simultaneously set its local fault flag. The FPGA collaborative management center determined this to be a safety-priority event, immediately blocking the faulty unit's drive pulse via a hard-wired bypass signal and broadcasting a fault isolation frame via the time-triggered bus network. Other conversion units received the isolation frame and updated their energy path status to prevent the fault from spreading further to the medium-voltage DC bus or adjacent equipment.

[0069] Example 3: Active suppression of excessive radiation

[0070] During laboratory electromagnetic compatibility testing, when a single-phase inverter unit operated at a fixed switching frequency, a high-order harmonic fell into the sensitive frequency band, causing near-field radiation to exceed limits. Since the system did not detect thermal over-limits or safety faults, the collaborative decision-making unit entered the electromagnetic radiation optimization layer and issued a switching frequency offset command to the inverter unit, shifting the main harmonic out of the sensitive frequency band. Subsequently, based on the radiation feedback results, a small-range frequency jitter was further enabled to disperse spectral energy. After adjustment, the radiation peak value decreased, and the output power quality remained within the preset range.

[0071] Example 4: Time-slot phase rearrangement of multi-element radiation superposition

[0072] When three three-phase inverter units operate in parallel, their PWM carriers have the same initial phase, resulting in highly synchronized switching actions. A near-field electromagnetic probe detects that the radiation peak value exceeds the limit in a specific frequency band. The FPGA collaborative management center generates a time-slot phase rearrangement command based on the radiation peak time and the switching phase of each unit, delaying the PWM synchronization phase of the second and third units by a preset angle, thus staggering the switching actions of the three units. From the next bus cycle onwards, the combined radiation peak value decreases, while the average output power of each unit remains unchanged.

[0073] Example 5: Lifetime-Aware Scheduling

[0074] The system contains two identical single-phase power factor correction (PFC) units. Unit A has a longer cumulative operating time and a health level below 60%, while Unit B is a newly replaced unit with a health level close to 100%. Under normal load, the system allocates most of the power load to Unit B, while Unit A operates under light load. When the load increases, the system prioritizes increasing the power of Unit B and limits the maximum output of Unit A. This lifespan-aware scheduling slows down the thermal aging of the weakened units and reduces the probability of unplanned downtime.

[0075] Example 6: Coordinated Control of Heat Dissipation Nodes

[0076] The system is configured with an intelligent air-cooled node, which acts as a controlled node on the time-triggered bus network to receive speed commands. When localized heat accumulation occurs near a certain LLC converter unit but has not yet reached the power degradation threshold, the safety-thermal-magnetic collaborative decision unit sends a speed increase command to the corresponding air duct and temporarily adjusts the power timing of adjacent units to reduce thermal coupling. If the temperature difference returns to normal after several bus cycles, the cooling node reduces its speed in a stepped manner to avoid frequent start-stop operations.

[0077] This invention can be widely applied to modular power routers, distributed energy storage systems, electric vehicle charging and swapping stations, data center power supply systems, ship integrated power systems, aviation ground power supplies, and other multi-port high-power-density power conversion scenarios. Its FPGA-based hardware pipeline decision structure enables deterministic response, its time-triggered bus network ensures predictable communication in multi-node environments, and its global sensing node cluster provides thermal, safety, and electromagnetic status data under a unified time reference. Therefore, this invention has significant engineering implementation value and promising prospects for industrial application.

[0078] In summary, the time-triggered multi-source power system for safety, heat accumulation, and electromagnetic radiation management provided by this invention comprehensively addresses safety faults, heat accumulation, and electromagnetic radiation issues within the same system architecture through its configured deterministic time slot partitioning mechanism, independent global sensing node cluster, FPGA hardware arbitration unit, nested priority strategy module, and collaborative control command generation unit. This system not only improves system reliability and fault isolation capabilities but also enhances electromagnetic compatibility and lifetime balancing capabilities under dynamic operating conditions. All equivalent substitutions, improvements, or combinations made within the spirit and principles of this invention should be included within the scope of protection of this invention.

[0079] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-source power system based on time-triggered safety-thermal-magnetic coordinated management, characterized in that, include: The cloud-edge collaborative management layer, FPGA collaborative management hub, time-triggered bus network, heterogeneous power conversion unit cluster, global sensing node cluster, high-voltage DC energy bus, medium-voltage DC energy bus, and low-voltage DC energy bus. The cloud-edge collaborative management layer is connected to the FPGA collaborative management center and is used to set the system operation mode, issue energy scheduling targets, and receive system status feedback. The FPGA collaborative management hub is connected to the heterogeneous power conversion unit cluster and the global perception node cluster through the time-triggered bus network. It is configured to generate deterministic time slot scheduling, uniformly receive, fuse, and prioritize safety parameters, thermal parameters, and electromagnetic parameters, and issue fault isolation, power derating, load migration, active heat dissipation, electromagnetic suppression, or time slot phase rearrangement instructions to the heterogeneous power conversion unit cluster. The time-triggered bus network adopts a time-division multiple access mechanism based on a static scheduling table, allocating an independent transmission time slot to each connected node; The heterogeneous power conversion unit cluster consists of multiple power conversion units, each of which is connected to a high-voltage DC power bus, a medium-voltage DC power bus, or a low-voltage DC power bus. It is used to realize AC-DC, DC-DC, or DC-AC power conversion and to perform power regulation, path switching, or fault isolation according to the instructions issued by the FPGA collaborative management center. The global sensing node cluster consists of multiple independent sensor nodes. Each node uploads thermal parameters, electrical safety parameters, electromagnetic radiation parameters, and operating status parameters through a time-triggered bus network. The high-voltage DC power bus, medium-voltage DC power bus, and low-voltage DC power bus achieve energy transfer and distribution through a heterogeneous power conversion unit cluster.

2. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 1, characterized in that, The FPGA collaborative management hub includes: an energy routing strategy parsing unit, a time-triggered scheduling unit, a global clock synchronization unit, a static scheduling table storage unit, a bus monitoring and fault isolation unit, a status summary and uplink feedback unit, and an ampere-thermal-magnetic collaborative decision-making unit. The safety-thermal-magnetic collaborative decision-making unit interacts with the energy routing strategy parsing unit, the time-triggered scheduling unit, the bus monitoring and fault isolation unit, and the status summary and uplink feedback unit to construct a status diagram based on the received multi-dimensional sensing parameters and generate collaborative control commands according to preset nested priority rules.

3. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 2, characterized in that, The FPGA collaborative management hub is configured to divide the bus cycle into synchronization time slots, sensing data fusion time slots, collaborative arbitration time slots, instruction execution time slots, and protection interval time slots according to a static scheduling table. The synchronization time slot corresponds to the global clock synchronization function; the sensing data fusion time slot corresponds to the data uploading and normalization processing function of the global sensing node cluster; the collaborative arbitration time slot corresponds to the state diagram construction, threshold comparison, and priority arbitration function of the safety-thermal-magnetic collaborative decision-making unit; the instruction execution time slot corresponds to the collaborative control instruction issuance function; and the protection interval time slot corresponds to the bus commutation blanking, scheduling table double-buffer switching, and hard-wired bypass signal establishment functions.

4. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 2, characterized in that, The An-thermal-magnetic collaborative decision-making unit is configured to perform arbitration according to nested priority rules, which include: The first priority is absolute safety. When the insulation impedance of any conversion unit is lower than the danger threshold, or when continuous overcurrent, overvoltage, grounding fault, or arcing fault characteristics are detected, an unconditional unit isolation command is triggered. The second priority is thermal management. In the absence of a first priority event, if the junction temperature of the power device of any conversion unit exceeds the first thermal threshold but is lower than the second thermal threshold, a power reduction, load migration, or active heat dissipation command is triggered. If it exceeds the second thermal threshold, a shutdown command is triggered. The third priority is electromagnetic radiation management. In the absence of first and second priority events, if the radiation intensity of any conversion unit in a specific frequency band exceeds the electromagnetic compatibility limit, a switching frequency offset, frequency jitter, or time slot phase rearrangement command will be triggered.

5. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 4, characterized in that, The FPGA collaborative management hub also includes a dedicated hard-wired bypass signal independent of the time-triggered bus network (3), which is used to bypass the bus scheduling cycle and directly force the disconnection of the drive enable or power circuit of the fault conversion unit when a safety absolute priority event is triggered, while broadcasting a fault isolation frame through the time-triggered bus network (3).

6. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 2, characterized in that, The An-Temperature-Magnetic Cooperative Decision Unit also maintains the health table of each conversion unit in the heterogeneous power conversion unit cluster, and updates the health value according to the number of thermal cycles, the highest junction temperature record, the number of switching cycles, the radiation exposure time, and the number of fault derating cycles; when allocating power tasks, the FPGA Cooperative Management Center prioritizes allocating high-power, high-frequency tasks to units without abnormal working conditions.

7. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 1, characterized in that, The time-triggered bus network adopts a redundant dual-channel RS-485 bus, CAN-FD bus, or Ethernet physical layer, and time division multiple access is implemented by FPGA hardware; the global sensing node cluster and the heterogeneous power conversion unit cluster share the same physical bus and are allocated different time slot sets, or adopt a dual-bus structure with separate control bus and sensing bus.

8. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 1, characterized in that, The global sensing node cluster includes infrared thermopile sensor nodes, power device junction temperature estimation sensor nodes, high-frequency current injection insulation monitoring sensor nodes, DC bus voltage and current sampling sensor nodes, near-field electromagnetic probe sensor nodes, and vibration sensor nodes. Each sensor node has an independent local clock, node identifier, and time slot transmission enable signal, and transmits data frames only within the sensing data time slots allocated to it by the static scheduling table.

9. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 1, characterized in that, The heterogeneous power conversion unit cluster includes at least two of the following: a three-phase Vienna rectifier unit, a single-phase power factor correction unit, an LLC resonant converter unit, a phase-shifted full-bridge converter unit, a bidirectional DC / DC unit, a single-phase inverter unit, a three-phase inverter unit, and an energy storage interface unit. Each unit is equipped with a local controller to perform local closed-loop control, while system-level start-up, mode switching, reference value update, derating, and isolation actions are uniformly scheduled by the FPGA collaborative management center through a time-triggered bus network.

10. The multi-source power system based on time-triggered safety-thermal-magnetic coordinated management according to claim 1, characterized in that, The rated voltage of the high-voltage DC power bus is 600V, the rated voltage of the medium-voltage DC power bus is 48V, and the rated voltage of the low-voltage DC power bus is 24V. The rectifier unit in the heterogeneous power conversion unit cluster injects power into the high-voltage DC power bus, the LLC resonant converter unit realizes bidirectional conversion between the high-voltage DC power bus and the medium-voltage DC power bus, the phase-shifting full-bridge unit realizes bidirectional conversion between the medium-voltage DC power bus and the low-voltage DC power bus, and the inverter unit draws power from the low-voltage DC power bus and supplies power to the AC load.

Images