Substation battery failure removal device based on bidirectional mosfet structure

Abstract

This invention discloses a substation battery fault isolation device based on a bidirectional MOSFET structure, belonging to the field of substation battery safety technology. It includes a fault diagnosis unit, a multi-dimensional data acquisition unit, a bypass connection unit, and a first bidirectional MOSFET structure. The fault diagnosis unit is electrically connected to the bypass connection unit and the first bidirectional MOSFET structure. One end of the second bidirectional MOSFET structure in the bypass connection unit is electrically connected to the negative terminal of the battery, and the other end of the second bidirectional MOSFET structure is electrically connected to one end of an inductor. The other end of the inductor is electrically connected to one end of the first bidirectional MOSFET structure, and the other end of the first bidirectional MOSFET structure is electrically connected to the positive terminal of the battery. The fault diagnosis unit is also communicatively connected to the multi-dimensional data acquisition unit, which is electrically connected to the battery. This invention solves the technical problem of the difficulty in safely, accurately, and promptly isolating faulty batteries in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of substation battery safety technology, specifically a substation battery fault isolation device based on a bidirectional MOSFET structure. Background Technology

[0002] In substations, the DC system is considered the heart of the substation, providing uninterrupted and highly reliable power to critical primary and secondary equipment such as relay protection devices, automatic control equipment, emergency lighting, communication systems, and circuit breaker tripping and closing circuits. Battery banks, as the core energy storage unit of the DC system, play an irreplaceable role: when AC power fails or malfunctions, the battery bank is the sole backup power source for the safe and orderly operation of the entire station or for implementing fault isolation protection. Its operational reliability directly affects whether grid faults can be correctly isolated, preventing the escalation of accidents, and even impacting the safety and stability of the entire regional power grid. Currently, substation DC systems commonly use multiple batteries connected in series in groups. However, as chemical energy storage devices, batteries inevitably exhibit performance degradation differences between individuals due to factors such as sulfation, water loss, long-term float charging, shallow charging and discharging, and ambient temperature. When individual batteries experience voltage drops, capacity decay, or increased internal resistance due to long-term aging, a series of problems arise. Existing technologies have significant shortcomings in addressing these issues: 1. Passive maintenance mode with insufficient prevention: Currently, maintenance relies mainly on regular inspections by operators and manual inspections such as individual cell voltage measurements and verification discharges during annual DC system inspections. This method is severely delayed and cannot capture the degree of battery degradation in real time. There is a time window of several weeks between the start of battery performance degradation and manual fault detection, during which the system operates with defects, posing a significant safety hazard. 2. Outdated handling methods affecting power supply continuity: Once a faulty battery is discovered, a power outage plan must be arranged for manual replacement by technicians. This process not only requires interrupting the power supply to that section of the DC bus but is also complex and carries safety risks. 3. Incomplete status assessment with high risk of misjudgment: Existing monitoring mostly focuses only on float charge voltage, while battery aging often manifests in more subtle parameters such as increased internal resistance and temperature changes. The lack of multi-parameter fusion-based intelligent diagnosis easily leads to misjudgments of battery status or failure to detect potential faults in advance. Therefore, how to safely, accurately, and promptly isolate faulty batteries is a technical challenge that existing technologies struggle to solve. Summary of the Invention

[0003] To address the technical problem of existing technologies' inability to safely, accurately, and promptly isolate faulty batteries, this invention provides a substation battery fault isolation device based on a bidirectional MOSFET structure. Through the coordinated operation of a fault diagnosis unit, a multi-dimensional data acquisition unit, a first bidirectional MOSFET structure, and a bypass connection unit, it achieves safe, timely, and accurate isolation of faulty batteries. This solves the technical problem of existing technologies' inability to safely, accurately, and promptly isolate faulty batteries.

[0004] To solve the above-mentioned technical problems, the present invention provides a substation battery fault isolation device based on a bidirectional MOSFET structure, including a fault diagnosis unit, a multi-dimensional data acquisition unit, a bypass connection unit, and a first bidirectional MOSFET structure. The fault diagnosis unit is electrically connected to the bypass connection unit and the first bidirectional MOSFET structure. One end of the second bidirectional MOSFET structure in the bypass connection unit is electrically connected to the negative terminal of the battery, and the other end of the second bidirectional MOSFET structure in the bypass connection unit is electrically connected to one end of the inductor. The other end of the inductor is electrically connected to one end of the first bidirectional MOSFET structure, and the other end of the first bidirectional MOSFET structure is electrically connected to the positive terminal of the battery. The fault diagnosis unit is also communicatively connected to the multidimensional data acquisition unit, and the multidimensional data acquisition unit is electrically connected to the battery.

[0005] Preferably, the fault diagnosis unit is used to determine whether the battery is faulty based on the real-time multi-dimensional data collected by the multi-dimensional data acquisition unit. When the battery is determined to be faulty, a first control signal is generated and sent to drive the second bidirectional MOSFET structure in the bypass connection unit to turn on, and the real-time multi-dimensional data is continuously monitored. When the current data in the real-time multi-dimensional data meets the preset requirements, a second control signal is generated and sent to drive the first bidirectional MOSFET structure to turn off.

[0006] Preferably, the step of determining whether the battery is faulty based on real-time multi-dimensional data collected by the multi-dimensional data acquisition unit includes: Based on the battery safety requirements, preset conditions are obtained, and it is determined whether the real-time multi-dimensional data meets the preset conditions. When the conditions are met, it indicates a fault. If the conditions are not met, the historical operating stage of the battery is obtained from the resistance changes in the historical multi-dimensional data collected by the multi-dimensional data acquisition unit. A mapping relationship between the historical operating stage and the historical multi-dimensional data within the historical operating stage is established to obtain stage training data. The initial fault judgment model based on random forest is trained using the stage training data to obtain the final fault judgment model. The final fault judgment model and real-time multi-dimensional data are used to determine whether the battery is faulty.

[0007] Preferably, the step of determining whether the battery is faulty based on real-time multi-dimensional data collected by the multi-dimensional data acquisition unit further includes: When the resistance in the real-time multi-dimensional data is greater than the preset resistance, it means that the preset condition is met. When the voltage in the real-time multi-dimensional data is less than the preset voltage, and the duration of the voltage being less than the preset voltage exceeds the preset time, it indicates that the preset condition is met. If the temperature in the real-time multi-dimensional data is greater than the preset temperature, it means that the preset condition is met.

[0008] Preferably, the step of training the initial fault determination model based on random forest using stage training data to obtain the final fault determination model includes: The initial fault determination model is trained using the initial stable data in the stage training data to obtain the first final fault determination model. The initial fault determination model is trained using the performance degradation data in the stage training data to obtain the second final fault determination model. The initial fault determination model is trained using the fault latent data in the stage training data to obtain the third final fault determination model. The initial fault determination model is trained using the fault failure data in the stage training data to obtain the fourth final fault determination model. Among them, the first final fault determination model, the second final fault determination model, the third final fault determination model and the fourth final fault determination model constitute the final fault determination model.

[0009] Preferably, the step of determining whether the battery is faulty by using the final fault determination model and real-time multi-dimensional data includes: The current operating stage of the battery is obtained by measuring the resistance change in real-time multi-dimensional data, and then matched with the final fault determination model corresponding to the current operating stage. The real-time multi-dimensional data is input into the final fault determination model corresponding to the current operating stage to obtain the weights of the resistance change, voltage change and temperature change in the real-time multi-dimensional data. The battery fault is determined based on the weights of the resistance change, voltage change and temperature change.

[0010] Preferably, it also includes a voltage compensation unit and a current acquisition unit. The voltage compensation unit is electrically connected to the fault diagnosis unit, and the current acquisition unit is communicatively connected to the fault diagnosis unit. One end of the voltage compensation unit is electrically connected to the DC positive bus, and the other end of the voltage compensation unit is connected to one end of the current acquisition unit. The other end of the current acquisition unit is electrically connected to one end of the first bidirectional MOSFET structure.

[0011] Preferably, the fault diagnosis unit is further configured to calculate voltage compensation based on the voltage loss of the battery pack while issuing the second control signal, generate and issue a voltage adjustment signal to drive the voltage compensation unit to adjust the voltage, and also to receive the adjustment success signal output by the voltage compensation unit.

[0012] Preferably, the fault diagnosis unit is also used to receive a successful conduction signal output by the bypass connection unit.

[0013] Preferably, the fault diagnosis unit is also used to receive a disconnection success signal output by the first bidirectional MOSFET structure.

[0014] By adopting the above technical solution, the present invention has the following advantages: By comprehensively assessing and identifying battery faults based on real-time multi-dimensional data, and combining a bypass connection unit constructed with a second bidirectional MOSFET structure and inductor components with a first bidirectional MOSFET structure, the faulty battery can be safely, promptly, and accurately isolated. This solves the technical problem of safely, accurately, and promptly isolating faulty batteries, which is difficult in existing technologies. Considering that the dominant factors and manifestations of battery failures differ at different operating stages, this invention matches and calls a dedicated fault judgment model corresponding to the current stage. After being specifically trained with data specific to the corresponding stage, this model has deeply learned and mastered the fault evolution law and characteristic parameter correlation logic of that stage. It can autonomously output the dynamic weights of multi-dimensional parameters such as resistance change, voltage change, and temperature change under that stage. Finally, through the dynamic fusion of multiple parameters, it achieves accurate judgment of battery fault status, fundamentally solving the problem of fault misjudgment and missed judgment caused by the inability of fixed weights to adapt to the fault characteristics of the entire life cycle, and significantly improving the accuracy of battery fault judgment. Attached Figure Description

[0015] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings.

[0016] Figure 1 This is a schematic diagram of the first structure of the substation battery fault isolation device based on a bidirectional MOSFET structure according to the present invention. Figure 2 This is a schematic diagram of the bypass connection unit and the first bidirectional MOSFET structure in the substation battery fault clearing device based on the bidirectional MOSFET structure of the present invention. Figure 3 This is a schematic diagram of the second structure of the substation battery fault isolation device based on a bidirectional MOSFET structure according to the present invention. Figure 4 This is a schematic diagram of the substation battery fault isolation device based on bidirectional MOSFET structure of the present invention applied to multiple batteries. The components include: 1. Fault diagnosis unit; 2. Multidimensional data acquisition unit; 3. Battery; 4. Bypass connection unit; 41. Second bidirectional MOSFET structure; 42. Inductor element; 5. First bidirectional MOSFET structure; 6. Voltage compensation unit; 61. Third bidirectional MOSFET structure; 62. Fourth bidirectional MOSFET structure; 63. First inductor element; 7. Current acquisition unit; and 8. Backup battery. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only one preferred embodiment of this invention and are only used to explain this invention. They do not limit the scope of protection of this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it may also have additional steps not included in the figures; the process may correspond to a method, function, procedure, subroutine, subroutine, etc.

[0019] Example 1: like Figure 1 As shown, a substation battery fault isolation device based on a bidirectional MOSFET structure includes a fault diagnosis unit 1, a multi-dimensional data acquisition unit 2, a bypass connection unit 4, and a first bidirectional MOSFET structure 5. The fault diagnosis unit 1 is electrically connected to the bypass connection unit 4 and the first bidirectional MOSFET structure 5. One end of the second bidirectional MOSFET structure 41 in the bypass connection unit 4 is electrically connected to the negative terminal of the battery 3, and the other end of the second bidirectional MOSFET structure 41 in the bypass connection unit 4 is electrically connected to one end of the inductor element 42. The other end of the inductor element 42 is electrically connected to one end of the first bidirectional MOSFET structure 5, and the other end of the first bidirectional MOSFET structure 5 is electrically connected to the positive terminal of the battery 3. The fault diagnosis unit 1 is also communicatively connected to the multidimensional data acquisition unit 2, and the multidimensional data acquisition unit 2 is electrically connected to the battery 3.

[0020] like Figure 2As shown, both the second bidirectional MOSFET structure 41 and the first bidirectional MOSFET structure 5 in the bypass connection unit 4 are composed of two common-drain N-channel power MOSFETs connected back-to-back. The bidirectional MOSFET structure forms a bidirectional controllable switch. Specifically, the first bidirectional MOSFET structure 5 is composed of two common-drain N-channel power MOSFETs K1 and K2 connected back-to-back, and the first bidirectional MOSFET structure 5 serves as a switch to isolate the faulty battery. The second bidirectional MOSFET structure 41 in the bypass connection unit 4 is composed of two common-drain N-channel power MOSFETs K3 and K4 connected back-to-back, forming a bidirectional controllable switch.

[0021] The fault diagnosis unit 1 is used to determine whether the battery 3 is faulty based on the real-time multi-dimensional data collected by the multi-dimensional data acquisition unit 2. When the battery 3 is determined to be faulty, a first control signal is generated and sent to drive the bidirectional MOSFET structure 41 in the bypass connection unit 4 to turn on, and the real-time multi-dimensional data is continuously monitored. When the current data in the real-time multi-dimensional data meets the preset requirements, a second control signal is generated and sent to drive the first bidirectional MOSFET structure 5 to turn off.

[0022] In this embodiment, when the gate signals of the first bidirectional MOSFET structure 5 are both low, MOSFET K1 and MOSFET K2 in the first bidirectional MOSFET structure 5 are both turned off. Since the body diodes of the two MOSFETs are in opposite directions, current cannot be conducted in either direction, thus forming a bidirectional blocking state. The gate drive signals of all MOSFETs are isolated and driven by the control processor in the fault diagnosis unit 1 through a dedicated driver chip. Control processor (battery control unit, BCU): The core is an industrial-grade ARM processor, running an embedded Linux system and battery fault detection algorithm. It has multiple isolated CAN interfaces for connecting all battery management units (BMUs, including the first bidirectional MOSFET structure 5 and bypass connection unit 4). It has an Ethernet port that supports the IEC 61850 protocol and can upload data to the station control layer. It controls the bypass connection unit 4 and the first bidirectional MOSFET structure 5 of each battery 3, as well as the voltage compensation unit 6 of the total output circuit (such as a bidirectional DC / DC module or a set of backup battery pack switching switches). In bidirectional conduction mode, both MOSFETs K3 and K4 need to be turned on to allow current to flow in either direction. When sufficient drive voltage is applied to the gates of the bidirectional MOSFET structure 41 in the bypass connection unit 4, both MOSFETs K3 and K4 enter the conducting state, achieving bidirectional conduction. Forward current (from left to right): MOSFET K1's drain → source conducts, and K2 can also conduct, thus allowing current flow. Reverse current (from right to left): MOSFET K2's drain → source conducts, and K1 conducts, with the current direction opposite to the forward direction. All MOSFET gate drive signals are isolated and driven by the control processor through a dedicated driver chip.

[0023] In some embodiments, determining whether the battery is faulty based on real-time multi-dimensional data collected by the multi-dimensional data acquisition unit includes: Based on the battery safety requirements, preset conditions are obtained, and it is determined whether the real-time multi-dimensional data meets the preset conditions. When the conditions are met, it indicates a fault. If the conditions are not met, the historical operating stage of the battery is obtained from the resistance changes in the historical multi-dimensional data collected by the multi-dimensional data acquisition unit. A mapping relationship between the historical operating stage and the historical multi-dimensional data within the historical operating stage is established to obtain stage training data. The initial fault judgment model based on random forest is trained using the stage training data to obtain the final fault judgment model. The final fault judgment model and real-time multi-dimensional data are used to determine whether the battery is faulty.

[0024] Specifically, the step of determining whether the battery is faulty based on real-time multi-dimensional data collected by the multi-dimensional data acquisition unit also includes: When the resistance in the real-time multi-dimensional data is greater than the preset resistance, it means that the preset condition is met. When the voltage in the real-time multi-dimensional data is less than the preset voltage, and the duration of the voltage being less than the preset voltage exceeds the preset time, it indicates that the preset condition is met. If the temperature in the real-time multi-dimensional data is greater than the preset temperature, it means that the preset condition is met.

[0025] In this embodiment, the preset resistance is 1.5 times the initial resistance of the battery, the preset voltage is 90% of the standard voltage, the preset time is 30 seconds, and the preset temperature is 45 degrees Celsius. Furthermore, the preset resistance, preset voltage, preset time, and preset temperature can be flexibly set according to user needs.

[0026] As one embodiment, the step of training the initial fault determination model based on random forest using staged training data to obtain the final fault determination model includes: The initial fault determination model is trained using the initial stable data in the stage training data to obtain the first final fault determination model. The initial fault determination model is trained using the performance degradation data in the stage training data to obtain the second final fault determination model. The initial fault determination model is trained using the fault latent data in the stage training data to obtain the third final fault determination model. The initial fault determination model is trained using the fault failure data in the stage training data to obtain the fourth final fault determination model. Among them, the first final fault determination model, the second final fault determination model, the third final fault determination model and the fourth final fault determination model constitute the final fault determination model.

[0027] Understandably, during the initial stable phase, the resistance of a battery remains stable with minimal fluctuations. During the performance degradation phase, the resistance begins to show a slow but continuous upward trend. In the fault latency phase, the resistance rises rapidly, reaching a clear inflection point. During battery failure, the resistance rises sharply, exhibiting a sudden change. By dividing the entire battery lifecycle into different stages—initial stability, performance degradation, fault latency, and failure—and training multiple fault detection models using training data corresponding to each stage, the final fault detection model can fully learn and match the fault characteristics and evolution patterns of the battery at different operating stages. Since the resistance change trends, dominant fault factors, and manifestations differ significantly across stages, a staged training approach effectively improves the model's accuracy and sensitivity in identifying fault modes at each stage. Particularly during the fault latency and failure phases, the model can more accurately capture key features such as accelerated resistance increases, inflection points, and sudden changes, thereby significantly improving early fault warning capabilities and the reliability of fault detection. Furthermore, the multi-stage model fusion approach enhances the robustness and adaptability of the overall model, enabling it to maintain stable judgment performance under different battery health conditions. This effectively reduces the false positive and false negative rates of fault diagnosis, and improves the comprehensiveness and accuracy of battery fault diagnosis.

[0028] As one embodiment, the step of determining whether the battery is faulty by using a final fault determination model and real-time multi-dimensional data includes: The current operating stage of the battery is obtained by measuring the resistance change in real-time multi-dimensional data, and then matched with the final fault determination model corresponding to the current operating stage. The real-time multi-dimensional data is input into the final fault determination model corresponding to the current operating stage to obtain the weights of the resistance change, voltage change and temperature change in the real-time multi-dimensional data. The battery fault is determined based on the weights of the resistance change, voltage change and temperature change.

[0029] In this embodiment, a comprehensive score is used to determine whether the battery is faulty, based on changes in resistance, voltage, and temperature. , , , These represent the weights of voltage change, resistance change, and temperature change, respectively. Indicates voltage change. Indicates the change in resistance. This indicates temperature change. When the overall score exceeds the preset score, a fault is indicated. The preset score can be flexibly set according to requirements.

[0030] In some embodiments, such as Figure 3As shown, it also includes a voltage compensation unit 6 and a current acquisition unit 7. The voltage compensation unit 6 is electrically connected to the fault diagnosis unit 1, and the current acquisition unit 7 is communicatively connected to the fault diagnosis unit 1. One end of the voltage compensation unit 6 is electrically connected to the DC positive bus, and the other end of the voltage compensation unit 6 is connected to one end of the current acquisition unit 7. The other end of the current acquisition unit 7 is electrically connected to one end of the first bidirectional MOSFET structure 5.

[0031] In this embodiment, the voltage compensation unit 6 is specifically an automatic switching unit. The voltage compensation unit 6 is composed of a third bidirectional MOSFET structure 61, a fourth bidirectional MOSFET structure 62, and a first inductor element 63. The third bidirectional MOSFET structure 61 is connected in series with the first inductor element 63. The third bidirectional MOSFET structure 61 is also electrically connected to the DC positive bus. The first inductor element 63 is also electrically connected to the positive terminal of the backup battery 8. The fourth bidirectional MOSFET structure 62 is electrically connected to the DC positive bus, the fourth bidirectional MOSFET structure 62 is electrically connected to the current acquisition unit 7, and the fourth bidirectional MOSFET structure 62 is also electrically connected to the negative terminal of the backup battery 8.

[0032] Specifically, the fault diagnosis unit 1 is also used to calculate voltage compensation based on the voltage loss of the battery pack while issuing the second control signal, generate and issue a voltage adjustment signal to drive the voltage compensation unit 6 to adjust the voltage, and is also used to receive the adjustment success signal output by the voltage compensation unit 6.

[0033] Specifically, the fault diagnosis unit 1 is also used to receive the successful conduction signal output by the bypass connection unit 4.

[0034] Specifically, the fault diagnosis unit 1 is also used to receive the disconnection success signal output by the first bidirectional MOSFET structure 5.

[0035] Understandably, such as Figure 4 As shown, by using the device of this invention, fault detection can be performed on each battery in the battery pack, thereby identifying whether any battery is faulty and actively isolating the faulty battery, thus improving the reliability of the battery pack power supply and solving the problems of passive fault isolation, easy power interruption, and high operation and maintenance costs of existing substation battery packs. Specifically, the inductor 42 and the first bidirectional MOSFET structure 5 are both electrically connected to the negative terminal of the adjacent battery. The second bidirectional MOSFET structure 41 in the bypass connection unit 4 of the last group of devices and the negative terminal of the battery 3 in the last group of devices are both electrically connected to the DC negative bus. Under normal battery monitoring, K1 and K2 of all battery management units are turned on, and K3 and K4 are turned off, allowing the batteries to supply power normally in series. Real-time multi-dimensional data is periodically (e.g., every second) reported to the battery control processor. The battery control unit continuously runs the fault detection algorithm in the final fault determination model.

[0036] The battery fault isolation process is as follows: At time t0 (initial): K1 and K2 of battery Bat-n are turned on, and K3 and K4 are turned off. At time t1 (bypass establishment): After receiving the isolation command, the battery management unit first sends a signal to control K3 and K4 in the bypass connection unit 4 to slowly turn on (soft-start PWM can be used). Due to the presence of inductor 42, the bypass current slowly rises from 0. At this time, the current gradually shifts from "via Bat-n" to "via bypass branch"; During the t1-t2 phase (current transfer): as the bypass current rises to near the load current, the current flowing through Bat-n gradually decreases to near zero. This process lasts approximately several milliseconds to tens of milliseconds, determined by the inductance value and drive speed, achieving "zero-current switching." At time t2 (fault clearing): Once the battery control unit detects or determines that the bypass current has stably carried the full load current, it immediately sends a signal to turn off K1 and K2. At this time, Bat-n is completely disconnected from the main circuit, and its voltage becomes floating. Since the current in K1 and K2 was almost zero before the switch, there is no arc during the turn-off; After t2 (stable bypass): K3 and K4 remain on, while K1 and K2 remain off. Bat-n is reliably isolated, and the battery pack continues to form a current path through bypass connection unit 4, with the total output voltage dropping by only about 2V; 4) System-level voltage compensation and status reporting: Simultaneously with the issuance of the isolation command, the battery control unit sends a voltage control command to the voltage compensation unit 6, increasing its output voltage by approximately 2V (or activating a spare battery) to maintain a constant DC bus voltage. The battery management unit of Bat-n reports an "isolated" status. The battery control unit updates the system topology, records the event log, and reports alarm information and maintenance recommendations (such as "Please replace battery n when appropriate") via the station control network.

[0037] The substantial effects of this invention are: By comprehensively assessing and identifying battery faults based on real-time multi-dimensional data, and combining a bypass connection unit constructed with a second bidirectional MOSFET structure and inductor components with a first bidirectional MOSFET structure, the faulty battery can be safely, promptly, and accurately isolated. This solves the technical problem of safely, accurately, and promptly isolating faulty batteries, which is difficult in existing technologies. This system achieves safe isolation of faulty batteries, reducing the risk of short circuits: By controlling the timing of first connecting the bypass and then disconnecting the faulty battery, and combining this with a series inductor to suppress sudden current changes, it fundamentally avoids direct short circuit accidents caused by the potential difference between the positive and negative terminals of the faulty battery at the moment of switching. The isolation process is arc-free and impact-free, and its safety level is significantly higher than that of the traditional mechanical switch solution using relays and contactors. It ensures the stability of DC bus voltage: while isolating faulty batteries, by putting in compensation batteries or DC / DC voltage regulator modules for voltage regulation, the fluctuation of DC bus voltage can be strictly controlled within ±1.5%, ensuring uninterrupted power supply to critical loads such as downstream relay protection and control systems, and realizing true online maintenance; Extremely fast fault response and handling speed: The solid-state switch based on MOSFET has an action time in the microsecond range. The entire process from algorithm confirmation of fault to completion of isolation and voltage compensation can be completed within 100 milliseconds, which far exceeds the speed of manual response and can effectively suppress the impact range of faults. Significantly improved fault accuracy: By comprehensively analyzing voltage, internal resistance, and temperature, early warning of potential battery faults can be achieved, shortening the discovery time of latent faults from "months" to "days". The early warning accuracy is expected to reach over 95%, significantly reducing the risk of sudden battery failure.

[0038] The specific embodiments described above are preferred embodiments of the substation battery fault isolation device based on the bidirectional MOSFET structure of the present invention, and are not intended to limit the specific scope of the present invention. The scope of the present invention includes, but is not limited to, these specific embodiments. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A substation battery fault isolation device based on a bidirectional MOSFET structure, characterized in that, It includes a fault diagnosis unit, a multi-dimensional data acquisition unit, a bypass connection unit, and a first bidirectional MOSFET structure; The fault diagnosis unit is electrically connected to the bypass connection unit and the first bidirectional MOSFET structure. One end of the second bidirectional MOSFET structure in the bypass connection unit is electrically connected to the negative terminal of the battery, and the other end of the second bidirectional MOSFET structure in the bypass connection unit is electrically connected to one end of the inductor. The other end of the inductor is electrically connected to one end of the first bidirectional MOSFET structure, and the other end of the first bidirectional MOSFET structure is electrically connected to the positive terminal of the battery. The fault diagnosis unit is also communicatively connected to the multidimensional data acquisition unit, and the multidimensional data acquisition unit is electrically connected to the battery.

2. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 1, characterized in that, The fault diagnosis unit is used to determine whether the battery is faulty based on the real-time multi-dimensional data collected by the multi-dimensional data acquisition unit. When the battery is determined to be faulty, a first control signal is generated and sent to drive the second bidirectional MOSFET structure in the bypass connection unit to turn on, and the real-time multi-dimensional data is continuously monitored. When the current data in the real-time multi-dimensional data meets the preset requirements, a second control signal is generated and sent to drive the first bidirectional MOSFET structure to turn off.

3. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 2, characterized in that, The method of determining whether the battery is faulty based on real-time multi-dimensional data collected by the multi-dimensional data acquisition unit includes: Based on the battery safety requirements, preset conditions are obtained, and it is determined whether the real-time multi-dimensional data meets the preset conditions. When the conditions are met, it indicates a fault. If the conditions are not met, the historical operating stage of the battery is obtained from the resistance changes in the historical multi-dimensional data collected by the multi-dimensional data acquisition unit. A mapping relationship between the historical operating stage and the historical multi-dimensional data within the historical operating stage is established to obtain stage training data. The initial fault judgment model based on random forest is trained using the stage training data to obtain the final fault judgment model. The final fault judgment model and real-time multi-dimensional data are used to determine whether the battery is faulty.

4. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 3, characterized in that, The method of determining whether the battery is faulty based on real-time multi-dimensional data collected by the multi-dimensional data acquisition unit also includes: When the resistance in the real-time multi-dimensional data is greater than the preset resistance, it means that the preset condition is met. When the voltage in the real-time multi-dimensional data is less than the preset voltage, and the duration of the voltage being less than the preset voltage exceeds the preset time, it indicates that the preset condition is met. If the temperature in the real-time multi-dimensional data is greater than the preset temperature, it means that the preset condition is met.

5. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 3, characterized in that, The step of training the initial fault determination model based on random forest using phased training data to obtain the final fault determination model includes: The initial fault determination model is trained using the initial stable data in the stage training data to obtain the first final fault determination model. The initial fault determination model is trained using the performance degradation data in the stage training data to obtain the second final fault determination model. The initial fault determination model is trained using the fault latent data in the stage training data to obtain the third final fault determination model. The initial fault determination model is trained using the fault failure data in the stage training data to obtain the fourth final fault determination model. Among them, the first final fault determination model, the second final fault determination model, the third final fault determination model and the fourth final fault determination model constitute the final fault determination model.

6. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 3, characterized in that, The process of determining whether a battery is faulty using a final fault determination model and real-time multi-dimensional data includes: The current operating stage of the battery is obtained by measuring the resistance change in real-time multi-dimensional data, and then matched with the final fault determination model corresponding to the current operating stage. The real-time multi-dimensional data is input into the final fault determination model corresponding to the current operating stage to obtain the weights of the resistance change, voltage change and temperature change in the real-time multi-dimensional data. The battery fault is determined based on the weights of the resistance change, voltage change and temperature change.

7. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 2, characterized in that, It also includes a voltage compensation unit and a current acquisition unit. The voltage compensation unit is electrically connected to the fault diagnosis unit, and the current acquisition unit is communicatively connected to the fault diagnosis unit. One end of the voltage compensation unit is electrically connected to the DC positive bus, and the other end of the voltage compensation unit is connected to one end of the current acquisition unit. The other end of the current acquisition unit is electrically connected to one end of the first bidirectional MOSFET structure.

8. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 7, characterized in that, The fault diagnosis unit is also used to calculate voltage compensation based on the voltage loss of the battery pack while issuing the second control signal, generate and issue a voltage adjustment signal to drive the voltage compensation unit to adjust the voltage, and receive the adjustment success signal output by the voltage compensation unit.

9. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 2, characterized in that, The fault diagnosis unit is also used to receive the successful conduction signal output by the bypass connection unit.

10. The substation battery fault clearing device based on a bidirectional MOSFET structure according to claim 2, characterized in that, The fault diagnosis unit is also used to receive the disconnection success signal output by the first bidirectional MOSFET structure.

Images