An intelligent fuse for a data center and a control method thereof
Patent Information
- Application Number
- CN202611208617.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-11
- Publication Date
- 2026-11-03
AI Technical Summary
Existing fuses cannot reliably interrupt small-scale overloads in high-voltage DC power supply systems for data centers. Their fusing time is uncontrollable, they cannot be linked with the system controller, their cooling function is limited, and they cannot detect disturbance currents, leading to protection blind spots and device damage.
Design an intelligent fuse that integrates a control module, a heat dissipation module, and a fuse module. It generates a breaking signal through current detection and signal control, controls heat dissipation using liquid cooling channels and solenoid valves, adjusts the fusing time with an auxiliary heating module, and combines quartz sand to extinguish the electric arc, thereby achieving active fusing and system collaborative protection.
It achieves reliable protection across the entire current range, eliminates blind spots in small-multiple overload protection, possesses early-stage active fault isolation capability, improves the current carrying capacity and safety of the equipment, and meets the rapid disconnection requirements of high-voltage DC systems in data centers.
Smart Images

Figure CN123033928A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of fuse technology, and particularly relates to a smart fuse for data centers and its control method. Background Technology
[0002] Currently, large data centers generally adopt a 240V–400V high-voltage DC power supply architecture. In such systems, load current fluctuations are large, and small overloads of 2–5 times the rated current occur frequently. Furthermore, the IT of power semiconductor devices (IGBTs, SiCMOSFETs) is limited. 2 The tolerance is limited, requiring the protection device to complete the disconnection within milliseconds.
[0003] Existing fuses mainly have the following shortcomings: Small-scale overloads cannot reliably interrupt circuits. The minimum breaking current of traditional fuses is typically 3 to 10 times the rated current. When an overload of 2 to 3 times the rated current occurs, the arc pre-arc time can be as long as tens of seconds, or even fail to interrupt the circuit, resulting in a protection blind zone. Even though some fuses employ liquid cooling or other forced heat dissipation measures to improve rated current carrying capacity, under small-scale overload conditions, forced heat dissipation actually further prolongs the time required for the fusible element to reach its melting temperature, making it even more difficult for the fuse to complete the interruption within the withstand time of the power semiconductor device.
[0004] The fusing time cannot be precisely controlled. The fusing time of traditional fuses depends entirely on the fault current amplitude and the heat-dissipation balance of the fusible element itself, and cannot actively adjust the fusing time window according to different overload conditions. In data center HVDC systems, the Ifusing of power semiconductor devices under different overload multiples... 2 Due to varying withstand times, existing fuses cannot dynamically adjust their fusing strategy based on real-time thermal conditions, easily leading to the dilemma of premature fusing causing unplanned power outages or premature fusing causing device damage.
[0005] Third, it cannot be linked with the system controller. Traditional fuses rely entirely on their own heat accumulation to break the circuit and cannot receive external commands from the EMS / BMS. When the system detects signs of an impending fault, the fuse cannot break the circuit actively, leading to the escalation of the fault.
[0006] Fourth, the cooling function is limited. Existing fuses with cooling structures only use the cooling system for normal heat dissipation and do not play an active role in the breaking process, resulting in low hardware utilization.
[0007] Fifth, there is a lack of awareness of disturbance currents. Frequent transient disturbance currents in data center power distribution networks can cause abnormal fuse blowing due to heat accumulation, and existing fuses cannot proactively intervene.
[0008] In summary, existing fuses cannot meet the requirements of full current range protection and system collaborative protection for data center systems. Summary of the Invention
[0009] To address the aforementioned issues, this invention proposes an intelligent fuse and its control method for data centers. Data centers have extremely high power density requirements; by reducing fuse size through efficient heat dissipation, this invention directly meets the needs of data centers to increase rack power density and save space. The power density of IT loads within data centers is constantly increasing, current fluctuations are large, and small-scale overloads occur frequently. This invention solves the pain points of traditional fuses, such as unreliable interruption in small-scale overload ranges and susceptibility to semiconductor device damage. Typical power monitoring and battery management systems in data centers demonstrate the integration of fuses into the overall data center protection strategy, achieving a system-level effect of early proactive fault isolation. This directly addresses the need for power semiconductor devices (IGBTs, SiC MOSFETs) in data center HVDC (High Voltage Direct Current) systems. 2 It has limited tolerance and requires millisecond-level interruption to prevent fire and explosion.
[0010] A first aspect of the present invention provides a smart fuse for a data center, comprising: a control module, a fuse module, and a heat dissipation module; The control module is used to generate a disconnection signal based on the main circuit current signal; The heat dissipation module is used to dissipate heat from the fuse module under normal current flow conditions, and to reduce heat dissipation from the fuse module in response to the disconnection signal. The fuse module is used to carry the main circuit current and dissipate heat in real time under the normal current-carrying state, and switches to the active fuse state in response to the heat dissipation module reducing heat dissipation, so as to cut off the main circuit current. The DC arc generated by the fuse is extinguished by quartz sand.
[0011] Preferably, it also includes an auxiliary heating module; The auxiliary heating module is thermally coupled to the fuse module and is used to provide auxiliary heat to the melt after the breaking signal is generated. The control module includes a thermal state estimation unit, which is used to acquire the real-time thermal state parameters of the fuse module, calculate the predicted value of the remaining fusing time based on the real-time thermal state parameters, and adjust the heating power of the auxiliary heating module according to the comparison result between the predicted value of the remaining fusing time and the target fusing time. The fuse module switches to active fuse state in response to the coordinated response of the heat dissipation module and the auxiliary heating module.
[0012] Preferably, the heat dissipation module includes a liquid cooling channel, a liquid cooling connector, and a liquid cooling substrate. The liquid cooling connector includes a plurality of connectors, one end of which is connected to the liquid cooling channel. The fuse module is thermally coupled to the heat dissipation module through the liquid cooling substrate.
[0013] Preferably, the control module includes a current detection module and a signal control module, and the heat dissipation module further includes a solenoid valve adapted to the number of connectors. The solenoid valve is disposed on the liquid cooling channel and is used to control the opening and closing of the liquid cooling channel. The current detection module and the signal control module are respectively disposed on the fuse module. The current detection module is used to collect main circuit current data. The signal control module is used to generate a disconnection signal based on the main circuit current data. The disconnection signal is used to control the solenoid valve to close the liquid cooling channel.
[0014] Preferably, the fuse module includes several fusible elements, insulating spacers, and quartz sand. The insulating spacers include, along their thickness direction, an upper copper layer, a ceramic layer, and a lower copper layer. The lower copper layer is connected to the liquid-cooled substrate by solder. The support feet of the fusible elements are fixedly connected to the upper copper layer. The fusible elements and the insulating spacers form a receiving space, and the quartz sand is filled in the receiving space.
[0015] Preferably, the fuse module further includes terminals, terminal connecting conductors, and a housing. The terminals are respectively located at the input and output ends of the fuse module. Each terminal and each fuse element is connected to the terminal connecting conductor. The terminals, housing, and insulating gasket form a sealed cavity.
[0016] A second aspect of the present invention provides a control method for a smart fuse in a data center, comprising: The main circuit current data is collected in real time through the current detection module; The heat accumulation of the main circuit is calculated by the signal control module based on the main circuit current data. A disconnection signal is generated based on the heat accumulation and / or main circuit current data; Based on the disconnection signal, the solenoid valve is controlled to close the liquid cooling channel, cutting off the cooling medium to reduce heat dissipation from the fuse module.
[0017] Preferably, the control module controls the solenoid valve to close the liquid cooling channel, thereby reducing the heat dissipation of the fuse module; The control module acquires the current temperature change rate of the melt and determines whether the current temperature change rate is lower than a preset rate threshold. If it is lower than the preset rate threshold, the control module starts the auxiliary heating module and repeats the following closed-loop adjustment steps at a preset cycle: acquire the current temperature parameters and the current temperature change rate of the melt. Based on the current temperature parameters and the current temperature change rate, update the remaining melting time prediction value; compare the updated remaining melting time prediction value with the target melting time; if the remaining melting time prediction value is greater than the target melting time, increase the heating power of the auxiliary heating module; if the remaining melting time prediction value is less than or equal to the target melting time, maintain or decrease the heating power of the auxiliary heating module. The melt is broken by the combined effect of reduced heat dissipation and auxiliary heating, thereby cutting off the main circuit current.
[0018] Preferably, the step of generating a disconnection signal based on the accumulated heat and / or main circuit current data further includes: When the instantaneous value of the main circuit current is lower than the preset current threshold and the heat accumulation is lower than the heat accumulation threshold, the control module keeps the solenoid valve open, and the cooling medium continues to flow through the liquid-cooled substrate to dissipate heat from the fuse module in real time. The signal control module generates the disconnection signal when any of the following conditions are met, specifically: the instantaneous value of the main circuit current reaches or exceeds the preset current threshold, the preset current threshold corresponds to a small multiple overload range of 2 to 5 times the rated current in the data center HVDC system, and the heat accumulation reaches the heat accumulation threshold within a preset time window. The disconnect signal is used to control the solenoid valve to close the liquid cooling channel, cut off the flow of cooling medium, and cause the fuse module to switch to active fuse state due to loss of heat dissipation.
[0019] Preferred options also include: The control module performs high-frequency sampling on the main circuit current data and extracts disturbance current components whose duration is less than a preset duration threshold and whose amplitude is less than a short-circuit current threshold. Based on the amplitude and duration of the disturbance current component, calculate the cumulative thermal contribution of the disturbance current component within a preset time window; If the cumulative heat contribution value reaches the preset disturbance heat accumulation warning threshold and the instantaneous value of the main circuit current is lower than the preset current threshold, the signal control module issues a heat accumulation warning signal; the heat accumulation warning signal is used to indicate that the fuse module is close to the heat accumulation fuse boundary. The control module uploads the heat accumulation early warning signal to the data center EMS / BMS, and the EMS / BMS (Energy Management System / Battery Management System) triggers the external command signal based on the heat accumulation early warning signal.
[0020] Preferably, the control module also includes receiving external command signals sent by the data center BMS / EMS, specifically including: The signal control module dynamically adjusts the preset current threshold and / or the heat accumulation threshold based on the system status information carried by the external command signal. When the external command signal is an active disconnection command, the signal control module directly generates the disconnection signal and controls the solenoid valve to close the liquid cooling channel; The system status information includes at least one of the following: battery cluster insulation resistance value, battery cell temperature, bus power fluctuation rate, and load-side temperature anomaly signal.
[0021] Because the present invention adopts the above technical solution, it has the following advantages and positive effects compared with the prior art: 1. The heat dissipation module of the fuse is integrated with the active disconnection control module. During normal flow, the liquid-cooled substrate provides efficient heat dissipation, which greatly improves the flow capacity and reduces the size of the equipment. At the same time, no additional mechanical or explosive actuators are required, which greatly simplifies the product structure.
[0022] 2. Eliminating the blind zone of low-current overload protection and achieving full-current range protection: The control module monitors the current in real time and generates a tripping signal, closing the liquid-cooled flow channel to disrupt the thermal balance of the fuse and drive it to actively melt and break. This mechanism fills the protection blind zone of traditional fuses in the range of 2 to 5 times the rated current, effectively preventing power semiconductor devices from being damaged by overheating due to overload. At the same time, an auxiliary heating module provides auxiliary heat to further accelerate the temperature rise of the fuse, enabling the fuse to reliably break under different overload multiples.
[0023] 3. Two-level tripping triggering mechanism for enhanced collaborative protection: It features a two-level mechanism of autonomous protection based on current detection and system collaborative protection based on external commands from the data center BMS / EMS. When an abnormal precursor is detected but has not yet reached its own tripping threshold, active tripping can be directly triggered by external commands to achieve proactive isolation before the fault escalates.
[0024] 4. By sampling the main circuit current at high frequency, extracting the disturbance current component and calculating the cumulative heat contribution value, a heat accumulation warning signal is sent to the EMS / BMS in advance when the circuit is close to the fuse boundary, so that the system has the ability to intervene proactively and adjust the dynamic threshold.
[0025] 5. The disconnection response time can be matched by preset solenoid valve response delay parameters and / or cooling medium flow parameters. The control module determines the target disconnection response time according to the overload multiple and configures the execution parameters accordingly to meet the differentiated requirements of the data center HVDC system for disconnection under different overload conditions. At the same time, the use of quartz sand to extinguish the DC arc generated by the fuse ensures the safety and reliability of the disconnection process in high voltage DC scenarios. Attached Figure Description
[0026] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, wherein: Figure 1 This is a first view of the intelligent fuse in this invention; Figure 2 This is a second view of the intelligent fuse in this invention; Figure 3 This is a third view of the intelligent fuse in this invention; Figure 4 This is a first detailed view of the intelligent fuse of the present invention; Figure 5 This is a structural diagram of the terminal connection conductor of the intelligent fuse of the present invention.
[0027] Explanation of reference numerals in the attached figures: 1: Control module; 2: Fuse module; 201: Fuse element; 202: Insulating spacer; 20201: Upper copper layer; 20202: Ceramic layer; 20203: Lower copper layer; 204: Terminal; 205: Terminal connecting conductor; 206: Housing; 3: Heat dissipation module; 301: Liquid cooling channel; 302: Liquid cooling connector; 303: Liquid cooling substrate; 304: Solenoid valve. Detailed Implementation
[0028] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become clearer from the following description and claims. It should be noted that the drawings are all in a very simplified form and use non-precise ratios, and are only used to facilitate and clarify the illustration of the embodiments of the present invention.
[0029] It should be noted that all directional indicators (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicator will also change accordingly.
[0030] First Embodiment See Figures 1-3 The first aspect of the present invention provides a smart fuse for a data center, comprising: a control module 1, a fuse module 2, and a heat dissipation module 3; Control module 1 is used to generate a disconnection signal based on the main circuit current signal; The heat dissipation module 3 is used to dissipate heat from the fuse module 2 under normal current-carrying conditions, and to reduce heat dissipation from the fuse module 2 in response to the disconnection signal. The fuse module 2 is used to carry the main circuit current and dissipate heat in real time under normal current-carrying conditions, and switches to active fuse state in response to the heat dissipation module 3 reducing heat dissipation, so as to cut off the main circuit current. The DC arc generated by the fuse is extinguished by quartz sand.
[0031] The active breaking state is triggered by the heat dissipation module reducing heat dissipation when the main circuit current exceeds the preset minimum breaking current. The fuse's cooling system serves as both a heat dissipation component and an active breaking actuator. The breaking process is driven by rapid switching of cooling capacity. Specifically: First, under normal flow conditions, the cooling medium continuously flows through the liquid-cooled substrate, carrying away the Joule heat generated by the fuse, achieving a dynamic thermal balance between heat generation and dissipation. When the control module determines that the active breaking condition is met, it first generates a breaking signal, controlling the solenoid valve to close the liquid-cooled flow channel, cutting off the flow of cooling medium and reducing the heat dissipation capacity of the fuse module. At this time, the fuse's heat dissipation power approaches zero, the thermal balance is forcibly broken, and the fuse switches from a thermal equilibrium state to a thermal imbalance state, with the input power (I) decreasing. 2 R) Continuously generating heat causes the melt temperature to rise. Second, the control module acquires the current rate of temperature change of the melt and determines whether it is lower than a preset rate threshold. If it is lower, the control module activates the auxiliary heating module to provide additional auxiliary heat to the melt. The auxiliary heating, combined with the Joule heating of the main circuit current, increases the melt temperature rise rate and shortens the time required for the melt to reach the melting point, suitable for situations where the temperature rise is slow after cooling is interrupted under small overload conditions. Third, the control module repeatedly executes the following closed-loop adjustment steps at a preset cycle: acquire the current temperature parameters and current rate of temperature change of the melt; update the predicted remaining melting time based on the current temperature parameters and current rate of temperature change; compare the updated predicted remaining melting time with the target melting time; if the predicted remaining melting time is greater than the target melting time, increase the heating power of the auxiliary heating module; if the predicted remaining melting time is less than or equal to the target melting time, maintain or decrease the heating power of the auxiliary heating module. Through the aforementioned closed-loop regulation, the melt, under the combined effects of reduced heat dissipation capacity and auxiliary heating, precisely reaches the melting temperature within the target melting time. Fourth, when the temperature in the narrow neck region of the melt reaches the material's melting point, a solid-liquid phase transition occurs, causing the melt to melt instantly, forming a physical fracture and completely cutting off the main circuit fault current at the conductive circuit level. At the instant the melt vaporizes and melts, the metal vapor is broken down under the action of DC voltage, forming a high-temperature DC arc. The quartz sand filling the space between the melt and the insulating spacer rapidly melts under the action of the high-temperature arc, forming a high-resistivity glassy substance. Through cooling, compression, and arc segmentation, it absorbs arc energy and restricts arc propagation, ultimately forcibly extinguishing the DC arc.
[0032] This application establishes a two-level tripping trigger mechanism. The first level is an early warning mechanism: based on high-frequency sampling to extract the disturbance current component, the cumulative heat contribution value is calculated. When the early warning threshold is reached, a heat accumulation early warning signal is issued and uploaded to the EMS / BMS, and the system side executes external intervention commands such as load reduction and load migration. The second level is the tripping function: based on the instantaneous value of the main circuit current reaching a preset threshold (corresponding to 2 to 5 times the rated current), or the heat accumulation reaching the fusing threshold, or receiving an active tripping command from the EMS / BMS, the solenoid valve is triggered to close the liquid cooling channel, and active fusing is executed.
[0033] This application establishes a three-level protection strategy. The first level is monitoring: when there are slight fluctuations or short-term disturbances in the main circuit current and the heat accumulation is low, the control module records relevant data and uploads it to the EMS / BMS without taking any action. The second level is early warning: when the accumulated heat contribution value of the disturbance current reaches the early warning threshold, or the heat accumulation is close to but has not yet reached the fusing threshold, a heat accumulation early warning signal is issued. The EMS / BMS then decides whether to implement system derating or issue an active tripping command based on the global status. The third level is the tripping function: when the instantaneous current value reaches the threshold, the heat accumulation reaches the fusing threshold, or an active tripping command is received, active fusing is executed.
[0034] This system integrates the fuse's heat dissipation module with the active tripping control module. During normal current flow, efficient heat dissipation via a liquid-cooled substrate enhances current carrying capacity and reduces device size. Simultaneously, thermal management controls trigger the fuse, eliminating the need for additional mechanical or explosive actuators and simplifying the product structure. The control module monitors the current in real-time and generates a tripping signal, closing the liquid-cooled flow channel to disrupt the fusible element's thermal balance and drive it to actively melt. This mechanism fills the protection blind zone of traditional fuses in the 2 to 5 times rated current range, preventing damage to power semiconductor devices (IGBTs, SiC MOSFETs) due to overload heating. Furthermore, an auxiliary heating module provides additional heat after the heat dissipation capacity decreases, further accelerating the fusible element's temperature rise. A thermal state estimation unit acquires the fusible element's temperature parameters and temperature change rate in real-time, calculates the predicted remaining melting time, compares it with the target melting time, and adjusts the heating power of the auxiliary heating module in a closed-loop manner. This ensures the fusible element melts within the target time under the combined effects of reduced heat dissipation capacity and auxiliary heating, solving the problems of uncontrollable melting time and incompatibility with power semiconductor I / O in traditional fuses. 2The system addresses the technical challenges of the t-resistance curve. It features a two-tiered mechanism: autonomous protection based on current detection and system-wide collaborative protection based on external commands from the data center BMS / EMS. When an anomaly is detected but before the system's own breaking threshold is reached, active breaking can be triggered directly via external commands, achieving proactive isolation before the fault escalates and realizing deep integration between the underlying power distribution protection and the upper-level dispatch management. By sampling the main circuit current, extracting the disturbance current component, and calculating the cumulative heat contribution value, a heat accumulation warning signal is sent to the EMS / BMS in advance when the system approaches the fuse boundary, enabling the system to have proactive intervention and dynamic threshold adjustment capabilities, avoiding unplanned power outages caused by unnecessary physical fuse failures. The response time can be matched by preset solenoid valve response delay parameters and / or cooling medium flow parameters. The control module determines the target breaking response time based on the overload multiple and configures the execution parameters accordingly, meeting the differentiated requirements of the data center HVDC system for rapid breaking under different overload conditions; simultaneously, quartz sand is used to extinguish the DC arc generated by the fuse, ensuring reliable breaking under high-voltage DC scenarios.
[0035] Preferably, it also includes an auxiliary heating module; The auxiliary heating module and the fuse module are thermally coupled together to provide auxiliary heat to the molten metal after the breaking signal is generated; The control module includes a thermal state estimation unit, which is used to acquire the real-time thermal state parameters of the fuse module, calculate the predicted value of the remaining fusing time based on the real-time thermal state parameters, and adjust the heating power of the auxiliary heating module according to the comparison result between the predicted value of the remaining fusing time and the target fusing time. The fuse module switches to active fuse state in response to the coordinated response of the heat dissipation module and the auxiliary heating module.
[0036] The thermal state estimation unit is used to acquire the thermal state parameters of the fuse element in the fuse module in real time after the breaking signal is generated, and to calculate the predicted value of the remaining breaking time based on the thermal state parameters.
[0037] Specifically, the thermal state estimation unit acquires the current temperature parameters and current temperature change rate of the melt, and determines the current heat accumulation state of the melt based on a heat balance model. This heat balance model uses the integral of the product of the square of the main circuit current and the melt resistance over time as the total heat generation, subtracts the heat carried away by the cooling medium and heat loss through other means, to obtain the heat accumulation in the melt at the current moment. The heat carried away by the cooling medium is determined based on the cooling medium flow rate, cooling medium inlet temperature, cooling medium outlet temperature, and the specific heat capacity of the cooling medium.
[0038] After acquiring the current heat accumulation, current temperature parameters, and current temperature change rate, the thermal state estimation unit calculates the predicted remaining melting time based on the following principle: It obtains the difference between the melting point of the melt and the current temperature, divides this difference by the current temperature change rate, and the resulting value is the predicted remaining melting time. This predicted value represents the time required for the melt to reach the melting point at the current heating rate.
[0039] After generating the interruption signal, the control module first controls the solenoid valve to close the liquid cooling channel, reducing the heat dissipation capacity of the fuse module. After the liquid cooling channel is closed, the control module acquires the current temperature change rate of the molten metal and determines whether the current temperature change rate is lower than a preset rate threshold. If it is lower than the preset rate threshold, it indicates that the Joule heat generated by the main circuit current alone is insufficient to bring the molten metal to the melting temperature within the target melting time. The control module then activates the auxiliary heating module and provides auxiliary heat to the molten metal.
[0040] After the auxiliary heating module is activated, the control module repeatedly executes the closed-loop predictive adjustment step at a preset cycle, namely: acquiring the current temperature parameters and current temperature change rate of the melt, updating the predicted value of the remaining melting time, comparing the updated predicted value of the remaining melting time with the target melting time, and adjusting the heating power of the auxiliary heating module according to the comparison result: if the predicted value of the remaining melting time is greater than the target melting time, it indicates that the current heating rate is too slow, and the heating power of the auxiliary heating module is increased; if the predicted value of the remaining melting time is less than or equal to the target melting time, it indicates that the current heating rate meets the requirements, and the heating power of the auxiliary heating module is maintained or reduced. The target melting time is based on the I0 of the power semiconductor devices (IGBT, SiCMOSFET) in the data center HVDC system. 2 The tolerance capability is determined. For a given overload factor, the control module will adjust the It of the power semiconductor device. 2 The maximum time that the power semiconductor device can withstand is calculated by dividing the withstand value by the square of the current overload current, and this value is taken as the target melting time.
[0041] Through the synergistic effect of breaking the thermal balance by interrupting cooling, accelerating the temperature rise by auxiliary heating, and accurately controlling the timing of melting by closed-loop prediction, the melt can accurately reach the melting temperature and melt within the target melting time.
[0042] As an optional implementation, the auxiliary heating module may employ at least one of the following: a resistance heating film, a thin-film heater, a PTC heating element, a heating resistance wire, an induction heating structure, or a controllable trigger conductor disposed adjacent to the melt. The auxiliary heating module does not operate under normal current flow conditions or operates in standby mode with extremely low power. Upon receiving a start command from the control module, it begins to provide auxiliary heat to the melt. The heating power of the auxiliary heating module is adjusted in real time by the control module according to the closed-loop regulation steps to achieve precise control of the melt temperature rise rate.
[0043] As an optional implementation, a thermal state estimation unit can be used to replace the analytical calculation model described above with a trained artificial neural network model. The artificial neural network model is a semi-linear feedforward network model, including an input layer, a hidden layer, and an output layer. The artificial neural network model takes at least two of the following parameters as input: the temperature parameter of the melt, the rate of temperature change, the main circuit current parameter, and the cooling medium parameter, and outputs the predicted remaining melting time. The artificial neural network model is trained based on experimental data of the thermal characteristics of the fuse module under different current multiples, different cooling medium flow rates, and different ambient temperatures. Through this artificial neural network model, accurate prediction of the remaining melting time can be achieved under nonlinear and multi-parameter coupling conditions.
[0044] Preferably, the heat dissipation module 3 includes a liquid cooling channel 301, a liquid cooling connector 302, and a liquid cooling substrate 303. The liquid cooling connector 302 includes a plurality of connectors, one end of which is connected to the liquid cooling channel 301. The fuse module 2 is thermally coupled to the heat dissipation module 3 through the liquid cooling substrate 303.
[0045] The heat generated by the fuse module when carrying a large current is conducted to the liquid-cooled substrate through the insulating spacer and solder, etc., as a heat-conducting medium. The liquid-cooled substrate, as the core heat exchange component, transfers heat to the cooling medium in the internal flow channel. The flowing liquid cooling medium carries away the heat, maintaining the dynamic thermal balance of the fuse. The liquid-cooled connector, as a standardized fluid inlet / outlet interface, connects the external pipeline to the internal flow channel. When the intelligent fuse of this application is connected to a data center liquid cooling system, it utilizes the system pressure difference provided by the data center CDU (cooling distribution unit) to drive a large-scale circulation of the liquid cooling medium. Alternatively, when the intelligent fuse of this application is used for independent cooling, it can form a local small-scale circulation within the liquid-cooled flow channel through its own micro-power or phase change principle, achieving continuous heat transfer. The liquid-cooled flow channel 301 structure, in conjunction with the solenoid valve, constitutes a controllable heat dissipation channel. During normal flow, the channel is open, and the liquid cooling medium carries away the heat; when a disconnection signal is received, the solenoid valve cuts off the flow channel (or bypass), the heat conduction link is interrupted, heat cannot be dissipated, thus breaking the thermal balance and driving the fuse to melt rapidly. Through the design of the liquid-cooled connector, the liquid cooling medium can be directly connected and shared with the existing large-scale liquid cooling system of the data center. This shared piping and shared medium design avoids configuring a separate cooling system for the fuse, reducing the initial investment and operational complexity of the data center. In the transitional phase from air cooling to liquid cooling in the data center, or in certain nodes lacking centralized liquid cooling capabilities, this application allows the liquid cooling medium to be directly encapsulated within the flow channel to achieve independent self-circulating cooling. This dual-mode compatibility greatly expands the application scenarios of smart fuses. The fuse module and the liquid-cooled substrate are directly connected via thermal coupling, resulting in extremely low thermal resistance. Whether using shared or independent cooling, it achieves heat dissipation efficiency far exceeding that of traditional air-cooled fuses. For HVDC buses in data centers, smaller fuses with smaller cross-sections can be used, saving power distribution cabinet space and adapting to the increasing power density of single racks.
[0046] Preferably, the control module 1 includes a current detection module and a signal control module, and the heat dissipation module 3 also includes a solenoid valve 304 adapted to the number of connectors. The solenoid valve 304 is disposed on the liquid cooling channel 301 and is used to control the opening and closing of the liquid cooling channel 301. The current detection module and the signal control module are respectively disposed on the fuse module 2. The current detection module is used to collect the main circuit current data. The signal control module is used to generate a disconnection signal based on the main circuit current data. The disconnection signal is used to control the solenoid valve 304 to close the liquid cooling channel 301.
[0047] A current detection module mounted on the fuse collects main circuit current data in real time and transmits it to the signal control module. The signal control module has a built-in preset current threshold and performs real-time comparison and logical operations on the current data. When the current data reaches or exceeds the preset threshold, the signal control module autonomously generates an electrical tripping signal. This signal drives the solenoid valve located in the liquid cooling channel. Upon receiving the signal, the solenoid valve quickly closes the liquid cooling channel, cutting off the flow of the liquid cooling medium. This action of cutting off the flow of the liquid cooling medium directly causes the heat dissipation channel of the fuse module to be physically interrupted, instantly switching it from a steady-state thermal equilibrium of liquid cooling to an adiabatic heat storage state, forcing the area to quickly accumulate heat and fuse. Through the closed-loop linkage of current detection and signal control, the fuse can autonomously determine and trigger tripping within extremely small overload ranges such as 2 to 5 times the rated current, without the need for external BMS / EMS intervention. This solves the protection blind zone problem of traditional fuses failing to trip in time under small overloads, effectively preventing damage to power semiconductor devices such as IGBTs and SiC MOSFETs due to prolonged overload. As a high-speed switching element, the solenoid valve significantly reduces the interruption delay. The interruption response time is determined by both the solenoid valve response time and the fusible element heat accumulation time, and can be matched by adjusting the solenoid valve specifications and control parameters to perfectly meet the ultra-fast interruption requirements of data center HVDC systems. By highly integrating the current detection module, signal control module, and actuating solenoid valve onto the fuse body, the originally passive thermal protection device is transformed into an intelligent fuse with sensing, decision-making, and execution capabilities. This design reduces reliance on complex external control loops, improving the independence and reliability of equipment deployment.
[0048] See Figure 4 and Figure 5 Preferably, the fuse module 2 includes several fuse elements 201, insulating spacers 202, and quartz sand. The insulating spacers 202 include, along the thickness direction, an upper copper layer 20201, a ceramic layer 20202, and a lower copper layer 20203. The lower copper layer 20203 is connected to the liquid-cooled substrate 303 by solder. The support feet of the fuse elements 201 are fixedly connected to the upper copper layer 20201. The fuse elements 201 and the insulating spacers 202 form a receiving space, and the quartz sand is filled in the receiving space.
[0049] The insulating spacer employs a structure consisting of an upper copper layer 20201, a ceramic layer 20202, and a lower copper layer 20203. The ceramic layer provides electrical insulation, while the upper and lower copper layers and the ceramic itself possess thermal conductivity. The lower copper layers 20201 and 20203 are seamlessly connected to the liquid-cooled substrate 303 via solder, allowing the Joule heat generated by the molten material to pass through the insulating layer with extremely low thermal resistance and be conducted to the liquid-cooled substrate 303 where it is carried away by the liquid cooling medium. The support feet of the molten material are directly fixed to the upper copper layer. The upper copper layer not only serves as a current inductor and outlet conductor but also as a mechanical support, suspending and fixing the molten material within the containment space. This ensures uniform heat dissipation from the molten material 201 and maintains structural stability under thermal expansion and contraction or electrodynamic forces.
[0050] When the melt 201 melts due to the disruption of thermal equilibrium, a high-temperature DC arc is generated at the neck. The quartz sand filling the space between the melt 201 and the insulating spacer melts rapidly under the action of the high-temperature arc and forms a high-resistivity glassy substance. By cooling, compressing and splitting the arc, it absorbs the arc energy and limits the arc spread, ultimately extinguishing the DC arc.
[0051] Addressing the high voltage and high current characteristics of data center HVDC systems, a ceramic layer (20202) is incorporated into the insulating spacer. This ensures absolute electrical isolation between the main circuit and the liquid cooling system while providing a heat conduction path. This guarantees that the liquid cooling medium efficiently removes heat during normal flow, supporting the system's high power density requirements. Data center DC arcs are extremely difficult to extinguish due to the lack of a natural zero-crossing point, posing a fire and explosion risk. By tightly filling the sealed enclosure with quartz sand, it is ensured that DC arcs generated during both rapid and overload interruptions are quickly and safely extinguished, preventing fault escalation and ensuring the physical safety of the data center. The molten metal is fixed by an overlying copper layer, ensuring uniform heat distribution and maintaining good thermal balance under normal operating conditions. When the solenoid valve cuts off the liquid cooling medium, the narrow neck area of the molten metal, which has lost heat dissipation, can quickly concentrate heat, accelerating the melting process and improving the accuracy and consistency of the active interruption response.
[0052] Preferably, the fuse module 2 further includes terminals 204, terminal connecting conductors 205, and housing 206. The terminals 204 are respectively located at the input end and the output end of the fuse module 2. The terminals 204 and each fuse element 201 are respectively connected to the terminal connecting conductors 205. The terminals 204, housing 206, and insulating spacer 202 form a sealed cavity.
[0053] Current is introduced through input terminal 204, evenly distributed to each molten element 201 via terminal connecting conductor 205, and finally converged to output terminal through terminal connecting conductor. The terminal connecting conductor acts as a busbar node, reducing contact resistance and achieving current balance among multiple molten elements, preventing premature melting of a single molten element due to uneven current distribution. Terminal 204, housing 206, and bottom insulating spacer 202 together form a rigid sealed cavity. This cavity completely physically isolates the internal molten elements 201 and the filled quartz sand from the external environment, not only locking the filling density and position of the quartz sand but also preventing the entry of external air, moisture, and dust. Simultaneously, the quartz sand, as the core arc-extinguishing medium, can rapidly absorb arc heat when the molten elements melt and generate a high-temperature DC arc, melting to form high-resistivity glassy crystals. By dividing, compressing, and cooling the arc, it forcibly increases the arc resistance, ultimately extinguishing the arc rapidly in a DC circuit without a natural zero-crossing point. To address the high current demands of data center HVDC buses, ranging from hundreds to thousands of amperes, a low-impedance busbar design with terminal connections effectively achieves parallel current sharing among multiple fuse elements. This enhances the overall current-carrying capacity of the fuse and prevents single-point overheating or premature blowout due to uneven current distribution, ensuring long-term stable operation under high-current conditions. Data center DC arcs are extremely difficult to extinguish due to the lack of a natural zero-crossing point. A rigid, sealed cavity formed by terminals, a housing, and insulating spacers locks in the dense filling state of the internal quartz sand. When the fuse element actively melts, the quartz sand divides, cools, and increases the arc resistance, forcibly extinguishing the DC arc and completely confining it within the cavity, eliminating the risk of arc leakage and fire. This sealed structure completely physically isolates the internal core current-carrying and arc-extinguishing components from the external environment, effectively preventing the intrusion of moisture, dust, and salt spray from the data center, preventing oxidation of the internal fuse elements or moisture absorption of the quartz sand, and ensuring that the fuse's protective characteristics do not shift.
[0054] Second Embodiment A second aspect of the present invention provides a control method for a smart fuse in a data center, comprising: The main circuit current data is collected in real time through the current detection module; The heat accumulation of the main circuit is calculated by the signal control module based on the main circuit current data. A disconnection signal is generated based on the heat accumulation and / or main circuit current data; Based on the disconnection signal, the solenoid valve 304 closes the liquid cooling channel 301, cutting off the cooling medium to reduce heat dissipation to the fuse module 2.
[0055] The control method first acquires the main circuit current data in real time through a current detection module and transmits it to the signal control module. The signal control module has a built-in thermal model that dynamically calculates the heat accumulation in the main circuit based on the real-time current data, thereby simulating the current accumulated thermal stress state of the system in real time. The formula for calculating the heat accumulation is I. 2 t represents the main circuit current data, and t represents time. The signal control module uses the heat accumulation of instantaneous current data as a dual criterion for fusion and comparison. When the overall state reaches a preset danger threshold, the system can more realistically assess the current overload danger level and autonomously generate a precise disconnection control signal. The signal control module converts the generated digital disconnection signal into a physical execution command and sends it to the solenoid valve on the liquid cooling channel. After receiving the command, the solenoid valve quickly acts, closing the liquid cooling channel, physically cutting off the circulation of the liquid cooling medium, and completely severing the heat dissipation path of the fuse module. After the heat dissipation path is cut off, the fuse instantly switches to an adiabatic heat storage state. At this time, the thermal balance is forcibly broken, heat accumulates rapidly in the narrow neck area of the fuse, and the temperature rises sharply to the melting point, ultimately triggering the fuse to melt rapidly and completing the active disconnection. By calculating the heat accumulation in real time, the control method can dynamically track and accurately match the I of power semiconductors such as IGBTs and SiC MOSFETs. 2 The fuse has a tolerance boundary. When a small-scale sustained overload occurs in a data center HVDC system, it can intervene at the critical point before the semiconductor is damaged by accumulated thermal stress, preventing device burnout and avoiding malfunctions due to normal load fluctuations. The fusion of instantaneous current and heat accumulation criteria enables the fuse to accurately identify overload ranges of 2 to 5 times the rated value, which are difficult for traditional protection schemes to cover. It overcomes the protection blind zone created by the slow heat accumulation due to passive heat dissipation in traditional fuses, achieving full protection coverage from small-scale overloads to short-current. This control method efficiently couples a complex thermal decision-making algorithm with a simplified current-cutting and heat-dissipating physical actuator. After the heat dissipation path is cut off, the adiabatic heat storage effect forces the heat of the fuse to concentrate instantaneously, shortening the time required for passive heat accumulation to reach the melting point, effectively preventing economic losses caused by large-scale IT load power outages due to fault escalation.
[0056] Preferably, it further includes: the control module controls the solenoid valve to close the liquid cooling channel, reducing the heat dissipation of the fuse module; the control module obtains the current temperature change rate of the melt and determines whether the current temperature change rate is lower than a preset rate threshold; if it is lower than the preset rate threshold, the control module starts the auxiliary heating module and repeats the following closed-loop adjustment steps at a preset cycle: obtaining the current temperature parameters and current temperature change rate of the melt; Based on the current temperature parameters and the current rate of temperature change, update the predicted remaining melting time; compare the updated predicted remaining melting time with the target melting time; if the predicted remaining melting time is greater than the target melting time, increase the heating power of the auxiliary heating module; if the predicted remaining melting time is less than or equal to the target melting time, maintain or reduce the heating power of the auxiliary heating module. The melt is broken by the combined effect of reduced heat dissipation and auxiliary heating, thereby cutting off the main circuit current.
[0057] When the control module generates the disconnection signal, it first controls the solenoid valve to close the liquid cooling channel, reducing the heat dissipation capacity of the fuse module, i.e., cooling interruption. After the liquid cooling medium is cut off, the heat dissipation of the fuse module decreases, and the fuse element switches from a thermal equilibrium state to a thermal imbalance state. At this time, the Joule heat continuously generated by the main circuit current cannot be carried away by the cooling medium and is converted into a temperature rise of the fuse element itself.
[0058] The control module acquires the current temperature change rate of the melt and determines whether the current temperature change rate is lower than a preset rate threshold. If the current temperature change rate is lower than the preset rate threshold, it indicates that the Joule heat generated by the main circuit current alone is insufficient to bring the melt to the melting temperature within the target melting time. The control module then activates the auxiliary heating module (auxiliary heat supply) to provide additional auxiliary heat to the melt, thereby increasing the temperature rise rate of the melt. If the current temperature change rate is not lower than the preset rate threshold, it indicates that Joule heat alone is sufficient to meet the heating rate requirement. The control module does not activate the auxiliary heating module to save energy.
[0059] After the auxiliary heating module is activated, the control module repeatedly executes the closed-loop predictive adjustment step at a preset cycle. This involves acquiring the current temperature parameters and current temperature change rate of the melt, updating the predicted remaining melting time, comparing the updated predicted remaining melting time with the target melting time, and adjusting the heating power of the auxiliary heating module in real time based on the comparison result. If the predicted remaining melting time is greater than the target melting time, it indicates that the current heating rate is too slow, and the heating power of the auxiliary heating module is increased. If the predicted remaining melting time is less than or equal to the target melting time, it indicates that the current heating rate meets the requirements, and the heating power of the auxiliary heating module is maintained or decreased. Through this closed-loop adjustment, the melt, under the combined effect of reduced heat dissipation capacity and auxiliary heating, accurately reaches the melting temperature within the target melting time.
[0060] When the temperature in the narrow neck region of the melt reaches its melting point, the melt instantly melts and breaks, forming a physical fracture that completely cuts off the fault current in the main circuit from the conductive circuit level. At the moment the melt vaporizes and melts, a direct current arc is formed. Since direct current has no natural zero-crossing point, it must be forcibly extinguished. At this time, the quartz sand filling the area around the melt rapidly melts under the action of the arc, ultimately extinguishing the direct current arc.
[0061] Through a collaborative control mechanism of cooling interruption, auxiliary heating, and closed-loop prediction, the digital decisions of the control system are transformed into a controllable physical fusing process, enabling the fuse to actively disconnect within the target fusing time. The physical air gap formed by the melting of the fusible element ensures that even if the control circuit fails under extreme fault conditions, the main circuit current can be completely cut off, providing a final hard safety line for the data center power supply network. Data center HVDC systems have high voltage and large short-circuit capacity, and DC arcs are extremely difficult to extinguish and are destructive. The arc-extinguishing effect of quartz sand ensures that the fuse melts within a preset time and cools and extinguishes within the sealed space inside the fuse, preventing arc propagation and ensuring the safety of physical assets. The above process describes a complete closed loop from normal current flow to active fusing and finally to arc extinguishing. The initial liquid cooling maintains the reliability of normal flow, the mid-term cooling interruption, auxiliary heating, and closed-loop predictive coordinated control achieve precise control of the melting time, and the late-term quartz sand arc extinguishing ensures the thoroughness and safety of fault clearing. The seamless connection of the three enables the smart fuse to achieve full-scenario adaptive protection capabilities, including heat dissipation and current carrying, precise timing melting, and safe arc extinguishing.
[0062] Preferably, the step of generating a disconnection signal based on heat accumulation and / or main circuit current data further includes: When the instantaneous value of the main circuit current is lower than the preset current threshold and the heat accumulation is lower than the heat accumulation threshold, the control module 1 keeps the solenoid valve 304 open, and the cooling medium continues to flow through the liquid-cooled substrate 303 to dissipate heat from the fuse module 2 in real time. The signal control module generates a disconnection signal when any of the following conditions are met: the instantaneous value of the main circuit current reaches or exceeds the preset current threshold, the preset current threshold corresponds to the small multiple overload range of 2 to 5 times the rated current in the data center HVDC system, and the heat accumulation reaches the heat accumulation threshold within the preset time window. The disconnect signal is used to control the solenoid valve 304 to close the liquid cooling channel 301, cut off the flow of cooling medium, and cause the fuse module 2 to switch to active fuse state due to loss of heat dissipation.
[0063] Under normal flow conditions, the solenoid valve remains open only if both the instantaneous current and the accumulated heat are below a threshold. This ensures that the liquid cooling system provides efficient heat dissipation only within the system's absolutely safe thermal equilibrium range, precisely maintaining the fuse's normal flow capacity. Two independent trigger paths are set: one is when the instantaneous current exceeds a preset threshold (directly targeting sudden overloads of 2-5 times the normal value); the other is when the accumulated heat exceeds the limit within a specific time window. Triggering either path indicates that the system has entered the thermal runaway danger zone, and the control module immediately generates a decision signal. The disconnect signal directly acts on the actuator, controlling the solenoid valve to close the liquid cooling channel, physically cutting off the liquid cooling medium. The fuse breaks the thermal equilibrium, using the Joule heat of the molten element to heat up and melt. The strict dual low-limit judgment under normal conditions ensures that the liquid cooling will not be accidentally disconnected or the fuse will not be triggered when the data center load fluctuates normally; while under fault conditions, it ensures decisive action as soon as a dangerous trend appears, eliminating protection failure due to a single judgment. This achieves perfect decoupling between daily operation and fault protection. Normally, liquid cooling is kept on to support high-current, high-density current carrying capacity. Once a fault is detected, the efficient action of cutting off the heat dissipation medium forces the fuse to actively melt, effectively preventing thermal runaway damage to expensive power devices such as IGBTs and SiC. This application includes a two-stage triggering mechanism to enhance collaborative protection. The first stage triggering is a system-level early warning and soft intervention. Specifically, based on the high-frequency disturbance heat accumulation reaching the warning threshold, a heat accumulation warning signal is triggered and uploaded to the EMS / BMS, where the system side executes external intervention commands such as load reduction and external circuit breaker tripping. The second stage triggering is a local physical hard interruption. Based on the main circuit's large-multiple overload current or total heat accumulation reaching the melting threshold, a breaking signal is triggered, controlling the solenoid valve to open, causing the fuse to melt.
[0064] Preferred options also include: Control module 1 performs high-frequency sampling of the main circuit current data and extracts disturbance current components whose duration is less than a preset duration threshold and whose amplitude is less than a short-circuit current threshold. Based on the amplitude and duration of the disturbance current component, calculate the cumulative thermal contribution of the disturbance current component within a preset time window; If the cumulative heat contribution value reaches the preset disturbance heat accumulation warning threshold and the instantaneous value of the main circuit current is lower than the preset current threshold, the signal control module issues a heat accumulation warning signal; the heat accumulation warning signal is used to indicate that the fuse module 2 is close to the heat accumulation fuse boundary. Control module 1 uploads the heat accumulation early warning signal to the data center EMS / BMS, and the EMS / BMS triggers an external command signal based on the heat accumulation early warning signal.
[0065] The control module performs high-frequency sampling of the main circuit current and filters out disturbance current components with short durations and amplitudes below the short-circuit threshold using algorithms. Based on the amplitude and duration of these disturbances, it calculates their cumulative heat contribution within a time window, accurately quantifying the microscopic impact of complex load fluctuations in the data center on the thermal balance of the fuse. Below the trigger threshold for fuse operation, the system adds a disturbance heat accumulation warning threshold. When the heat contribution of a disturbance reaches the warning threshold, and the instantaneous current of the system is still within the normal range, the control module only issues a warning signal without executing the thermal break fuse action. The intelligent fuse, acting as an edge node, uploads local thermal status warning signals to the data center's EMS / BMS (Energy / Battery Management System) in real time. After receiving the signal, the EMS / BMS is no longer limited to local physical disconnection but performs logical judgments and issues external commands based on a global perspective, achieving deep integration of underlying power distribution protection and upper-level scheduling management. It pioneers a two-level disconnection triggering mechanism of system-level soft intervention warning and local-level physical hard fuse, achieving deep collaboration between edge perception and global scheduling, and constructing a multi-layered protection closed loop. The system proactively identifies heat accumulation caused by high-frequency disturbances and coordinates with the EMS / BMS to reduce or transfer load, prioritizing soft scheduling to mitigate potential risks and avoid unplanned power outages and component damage caused by direct physical fuse failure. In the event of ineffective system-level soft intervention or a sudden high-multiplier short circuit, local thermal shutdown and rapid fuse failure serve as the ultimate hard defense. This two-tiered mechanism works in tandem to perfectly balance the high continuity of power supply to the data center with absolute safety under extreme conditions.
[0066] Preferably, it also includes a control module 1 that receives external command signals sent by the data center BMS / EMS, specifically including: The signal control module dynamically adjusts the preset current threshold and / or heat accumulation threshold based on the system status information carried by the external command signal; When the external command signal is an active disconnect command, the signal control module directly generates a disconnect signal to control the solenoid valve 304 to close the liquid cooling channel 301; The system status information includes at least one of the following: battery cluster insulation resistance value, battery cell temperature, bus power fluctuation rate, and load-side temperature anomaly signal.
[0067] Based on the global system status issued by BMS / EMS, local current and heat accumulation thresholds are dynamically adjusted to ensure that the protection strategy accurately matches the real-time operating conditions of the system, avoiding false trips or failures to trip caused by rigid settings. Breaking the limitations of unidirectional protection, it supports the upper-level system in directly issuing active disconnect commands when a serious non-electrical fault is detected, achieving emergency physical isolation. Multi-dimensional status information from the battery side, bus side, and load side is incorporated into the fuse tripping decision-making process, enabling cross-system linkage, preventing fault propagation from the source, and improving the overall safety margin of the data center.
[0068] In the description of this application, it should be noted that the terms "inner" and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product is in use. They are used only for the convenience of describing this application and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.
[0069] It should also be noted that, unless otherwise explicitly specified and limited, the terms "setup" and "connection" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.
[0070] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific identification content executed by the system and device described above can be referred to the corresponding process in the foregoing method embodiments.
[0071] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments. Even if various changes are made to the present invention, if these changes fall within the scope of the claims of the present invention and their equivalents, they shall still fall within the protection scope of the present invention.
Claims
1. A smart fuse for data centers, characterized in that, include: Control module, fuse module, heat dissipation module; The control module is used to generate a disconnection signal based on the main circuit current signal; The heat dissipation module is used to dissipate heat from the fuse module under normal current flow conditions, and to reduce heat dissipation from the fuse module in response to the disconnection signal. The fuse module is used to carry the main circuit current and dissipate heat in real time under the normal current-carrying state, and switches to the active fuse state in response to the heat dissipation module reducing heat dissipation, so as to cut off the main circuit current. The DC arc generated by the fuse is extinguished by quartz sand.
2. The smart fuse for data centers according to claim 1, characterized in that, It also includes an auxiliary heating module; The auxiliary heating module is thermally coupled to the fuse module and is used to provide auxiliary heat to the melt after the breaking signal is generated. The control module includes a thermal state estimation unit, which is used to acquire the real-time thermal state parameters of the fuse module, calculate the predicted value of the remaining fusing time based on the real-time thermal state parameters, and adjust the heating power of the auxiliary heating module according to the comparison result between the predicted value of the remaining fusing time and the target fusing time. The fuse module switches to active fuse state in response to the coordinated response of the heat dissipation module and the auxiliary heating module.
3. The smart fuse for data centers according to claim 1, characterized in that, The heat dissipation module includes a liquid cooling channel, a liquid cooling connector, and a liquid cooling substrate. The liquid cooling connector includes several connectors, one end of which is connected to the liquid cooling channel. The fuse module is thermally coupled to the heat dissipation module through the liquid cooling substrate.
4. The smart fuse for data centers according to claim 3, characterized in that, The control module includes a current detection module and a signal control module. The heat dissipation module also includes a solenoid valve adapted to the number of connectors. The solenoid valve is located on the liquid cooling channel and is used to control the opening and closing of the liquid cooling channel. The current detection module and the signal control module are respectively located in the fuse module. The current detection module is used to collect main circuit current data. The signal control module is used to generate a disconnection signal based on the main circuit current data. The disconnection signal is used to control the solenoid valve to close the liquid cooling channel.
5. The smart fuse for data centers according to claim 3, characterized in that, The fuse module includes several fusible elements, insulating spacers, and quartz sand. The insulating spacers include an upper copper layer, a ceramic layer, and a lower copper layer in sequence along the thickness direction. The lower copper layer is connected to the liquid-cooled substrate by solder. The support feet of the fusible elements are fixedly connected to the upper copper layer. The fusible elements and the insulating spacers form a receiving space, and the quartz sand is filled in the receiving space.
6. The smart fuse for data centers according to claim 5, characterized in that, The fuse module further includes terminals, terminal connecting conductors, and a housing. The terminals are respectively located at the input and output ends of the fuse module. Each terminal and each fuse element is connected to the terminal connecting conductor. The terminals, housing, and insulating gasket form a sealed cavity.
7. A control method for a smart fuse used in a data center, characterized in that, include: The main circuit current data is collected in real time through the current detection module; The heat accumulation of the main circuit is calculated by the signal control module based on the main circuit current data. A disconnection signal is generated based on the heat accumulation and / or main circuit current data; Based on the disconnection signal, the solenoid valve is controlled to close the liquid cooling channel, cutting off the cooling medium to reduce heat dissipation from the fuse module.
8. The control method for a smart fuse for a data center according to claim 7, characterized in that, Also includes: The control module controls the solenoid valve to close the liquid cooling channel, thereby reducing the heat dissipation of the fuse module; The control module acquires the current temperature change rate of the melt and determines whether the current temperature change rate is lower than a preset rate threshold. If it is lower than the preset rate threshold, the control module starts the auxiliary heating module and repeats the following closed-loop adjustment steps at a preset cycle: acquire the current temperature parameters and the current temperature change rate of the melt. Based on the current temperature parameter and the current temperature change rate, update the predicted remaining melting time; then compare the updated predicted remaining melting time with the target melting time. If the predicted remaining melting time is greater than the target melting time, the heating power of the auxiliary heating module is increased; if the predicted remaining melting time is less than or equal to the target melting time, the heating power of the auxiliary heating module is maintained or decreased. The melt is broken by the combined effect of reduced heat dissipation and auxiliary heating, thereby cutting off the main circuit current.
9. The control method for a smart fuse for a data center according to claim 7, characterized in that, The step of generating a disconnection signal based on the heat accumulation and / or main circuit current data further includes: When the instantaneous value of the main circuit current is lower than the preset current threshold and the heat accumulation is lower than the heat accumulation threshold, the control module keeps the solenoid valve open, and the cooling medium continues to flow through the liquid-cooled substrate to dissipate heat from the fuse module in real time. The signal control module generates the disconnection signal when any of the following conditions are met, specifically: the instantaneous value of the main circuit current reaches or exceeds the preset current threshold, the preset current threshold corresponds to a small multiple overload range of 2 to 5 times the rated current in the data center HVDC system, and the heat accumulation reaches the heat accumulation threshold within a preset time window. The disconnect signal is used to control the solenoid valve to close the liquid cooling channel, cut off the flow of cooling medium, and cause the fuse module to switch to active fuse state due to loss of heat dissipation.
10. The control method for a smart fuse for a data center according to claim 7, characterized in that, Also includes: The control module performs high-frequency sampling on the main circuit current data and extracts disturbance current components whose duration is less than a preset duration threshold and whose amplitude is less than a short-circuit current threshold. Based on the amplitude and duration of the disturbance current component, calculate the cumulative thermal contribution of the disturbance current component within a preset time window; If the cumulative heat contribution value reaches the preset disturbance heat accumulation warning threshold and the instantaneous value of the main circuit current is lower than the preset current threshold, the signal control module issues a heat accumulation warning signal. The heat accumulation warning signal is used to indicate that the fuse module is approaching the heat accumulation and melting boundary; The control module uploads the heat accumulation early warning signal to the data center EMS / BMS, and the EMS / BMS triggers an external command signal based on the heat accumulation early warning signal.
11. The control method for a smart fuse for a data center according to claim 9, characterized in that, It also includes the control module receiving external command signals sent by the data center BMS / EMS, specifically including: The signal control module dynamically adjusts the preset current threshold and / or the heat accumulation threshold based on the system status information carried by the external command signal. When the external command signal is an active disconnection command, the signal control module directly generates the disconnection signal and controls the solenoid valve to close the liquid cooling channel; The system status information includes at least one of the following: battery cluster insulation resistance value, battery cell temperature, bus power fluctuation rate, and load-side temperature anomaly signal.
