AI-based EV rapid charger power module health management system

Abstract

The present invention relates to a Prognostics and Health Management (PHM) system for an EV fast charger power module, comprising: a multi-sensor data integration unit (100) capable of measuring all performance and degradation data of the EV fast charger power module, vulnerable component status data of the power converter, and cooling system health data; a predictive correction unit (200) using a hybrid digital twin that predicts and maintains the remaining useful life (RUL) of the EV fast charger power module through a hybrid digital twin model based on the data of the multi-sensor data integration unit (100); and an AI-based adaptive charging control system (300) having an embedded AI platform for optimizing the control of the output power and active cooling system of the EV fast charger according to the predicted remaining useful life (RUL) of the EV fast charger power module and criteria for operational abnormal signs. By implementing an AI-based EV fast charger power module health management system characterized by including a fault self-diagnosis and external notification system (400) that performs self-operation, fault removal, and emergency notification functions for abnormal data phenomena detected by the multi-sensor data integration unit (100) of the AI ​​Platform of the above-described AI-based adaptive charging control system (300), the lifespan of the EV fast charger module can be maximized and the operating rate of the charging station can be increased, and the health of management can be guaranteed by self-diagnosing and recovering from detected faults and notifying relevant operators and manufacturers of prompt and accurate information regarding faults that cannot be resolved.

Description

Technology Field

[0001] The present invention relates to an electric vehicle (EV) rapid charging system, and more specifically, to an AI-based EV rapid charger power module health management system that monitors the operating status of a high-power power module (SiC / GaN) in real time, manages the remaining lifespan of the power module by utilizing an AI-based prediction model and criteria for detecting operational anomalies, and ensures the soundness of management through self-diagnosis and external notification in the event of a failure. Background Technology

[0003] With the recent popularization of electric vehicles, there is a demand for ultra-fast charging technology (over 350kW), and accordingly, wide band gap (WBG) semiconductor modules based on SiC (silicon carbide) and GaN (gallium nitride) are being used as essential components in fast charger power conversion devices.

[0004] However, rapid thermal change (△Tj) stress occurring in high-power environments causes critical failure modes such as solder fatigue and wire bonding failure inside power modules, thereby reducing the lifespan of power modules and the operating rate of charging stations.

[0005] Conventional power module protection methods rely on passive approaches, such as simply halting operation when temperature or current thresholds are exceeded or derating the output to reduce the load below the rated value in order to improve reliability. However, this derating method had drawbacks, including the inability to predict failures based on accumulated degradation damage and the failure to fully utilize component performance. Furthermore, there was a lack of systems capable of early detection and proactive response to abnormal signs in critical peripheral components, such as electrolytic capacitors and cooling fans.

[0006] Meanwhile, Korean Registered Patent No. 10-2816308 (June 5, 2025) discloses a fault prediction system, device, and method for an artificial intelligence-based charger.

[0007] However, while the aforementioned patented technology predicts the state of an electric vehicle charger through an AI-based fault prediction model that uses the state information collected by the fault prediction device as an input variable and predicted values ​​regarding the timing and status of the electric vehicle charger's failure as output variables, it lacks the technical concept of attempting self-diagnosis and recovery for detected failures or notifying relevant operators and manufacturers of prompt and accurate information regarding unresolved failures. The problem to be solved

[0009] The objective of the present invention is to provide an AI-based EV fast charger power module health management system that resolves the problems of the prior art, precisely predicts the degradation status of the EV fast charger's power module in real time, and actively controls the charger's power output and cooling system based on AI using the predicted Remaining Useful Life (RUL) and specific criteria for operational abnormalities, thereby maximizing the lifespan of the EV fast charger module and increasing the operating rate of the charging station.

[0010] Another objective is to provide an AI-based EV fast charger power module health management system that attempts self-diagnosis and recovery for detected failures and ensures management soundness by promptly and accurately notifying relevant operators and manufacturers of unresolved failures. means of solving the problem

[0012] According to a feature of the present invention for achieving the aforementioned purpose, in a Prognostics and Health Management (PHM) system for EV fast charger power modules, the system comprises: a multi-sensor data integration unit (100) capable of measuring all of the following: power module performance and degradation data (110) having a measured value of a Temperature Sensitive Electrical Parameter (TSEP) (Vce_on or Rds_on) for accurately estimating the module current, output voltage, heat sink temperature, coolant flow rate, and instantaneous junction temperature (Tj) of the EV fast charger; vulnerable component status data (120) having temperature information of an electrolytic capacitor, which is a vulnerable component of a power converter; and cooling system health data (130) having fan reference speed and vibration information of 10,000 to 15,000 rpm according to the noise, current, and bearing life of the cooling fan; The present invention provides an AI-based EV fast charger power module health management system characterized by including: a predictive correction unit (200) using a hybrid digital twin that predicts and maintains the remaining lifespan (RUL) of the power module of the EV fast charger through a hybrid digital twin model based on data from the multi-sensor data integration unit (100); an AI-based adaptive charging control system (300) in which an AI Platform is embedded to optimize the control of the output power and active cooling system of the EV fast charger according to the predicted and maintained remaining lifespan (RUL) of the power module of the EV fast charger and criteria for operational abnormal signs; and a fault self-diagnosis and external notification system (400) in which the AI ​​Platform of the AI-based adaptive charging control system (300) performs self-operation, fault removal, and emergency notification functions regarding abnormal data phenomena detected by the multi-sensor data integration unit (100).

[0014] According to another embodiment of the present invention, the cooling system health data (130) of the multi-sensor data integration unit (100) further includes a reference value that is pre-set to detect normal operation (Normal) when the cooling water inlet temperature (Tin) is -25℃ to +55℃, warning and output limit (Derating) when it is +55℃ to +65℃, and abnormal heating (Fault / Trip) when it is +65℃ to +70℃.

[0015] According to another embodiment of the present invention, the hybrid digital twin model (210) of the prediction correction unit (200) using the hybrid digital twin is characterized by utilizing a Physics-Informed Neural Network (PINN) and combining thermal and electrical equations of a power module package with real-time operation data to quantify the cumulative damage factors of the module and continuously predict the remaining lifespan (RUL).

[0016] According to another embodiment of the present invention, the AI-based adaptive charging control system (300) further comprises: an RUL-based power modulator (310) that actively controls the operation of a charger based on the predicted remaining lifespan (RUL) and an abnormal data phenomenon judgment criterion, and when the RUL prediction falls below a dynamic threshold, the system preemptively reduces the maximum output current and / or duty cycle of the charger to mitigate a thermal change (△Tj) that causes damage; and a dynamic cooling optimization means (320) in which an AI Platform embedded in the AI-based adaptive charging control system (300) predicts future thermal load based on the vehicle's state of charge (SoC) profile and dynamically controls the pump speed and fan speed of the cooling system according to a fan speed abnormality judgment criterion to maintain the junction temperature (Tj) of the power module within the RUL optimization band.

[0017] The fault self-diagnosis and external notification system (400) according to another embodiment of the present invention is further characterized by including: a self-operation and fault removal means (410) in which an AI Platform embedded in the AI-based adaptive charging control system (300) performs fault removal based on self-operation such as a communication signal reboot for an abnormal data phenomenon detected; an emergency notification and preliminary inspection request means (420) that automatically generates a preliminary inspection request for a serious temperature overshoot or a fan speed difference of 10% or more that causes a fault such as a Shut Down that cannot be resolved by self-operation; and an integrated information notification means (430) that performs the function of notifying a Charge Point Operator (CPO) business operator and a charger manufacturer responsible for after-sales service of abnormal information regarding a fault that cannot be resolved by self-operation such as a communication signal reboot of an EV fast charger generated from the emergency notification and preliminary inspection request means (420).

[0018] The fault self-diagnosis and external notification system (400) according to another embodiment of the present invention is further characterized by dynamically modulating the maximum output power or current profile of the charger when at least one operational abnormality is detected among a current deviation of 1% or more per module and a difference of 1% or more between the output command value and the output voltage.

[0019] The fault self-diagnosis and external notification system (400) according to another embodiment of the present invention is characterized by generating a warning signal when the speed of the cooling fan shows a deviation of 5% from the rated speed (e.g., 15,000 rpm), and stopping the operation of the corresponding power module when it shows a deviation of 10% or more. Effects of the invention

[0021] The AI-based EV fast charger power module health management system according to a preferred embodiment of the present invention can be expected to have the following effects.

[0022] (1) The present invention can maximize the lifespan of the EV fast charger module and increase the operating rate of the charging station by accurately predicting the degradation state of the power module of the EV fast charger in real time and actively controlling the power output and cooling system of the charger based on AI with the predicted remaining lifespan (RUL) and specific operational abnormality indicators.

[0023] (2) The present invention can ensure the soundness of power module management of EV fast chargers by attempting self-diagnosis recovery for faults detected through a fault self-diagnosis and external notification system, and by notifying relevant operators and manufacturers of prompt and accurate information regarding faults that are not resolved. Brief explanation of the drawing

[0025] FIG. 1 is a drawing showing the prior art. FIG. 2 is a block diagram showing the overall technical configuration of an AI-based EV fast charger power module health management system according to a preferred embodiment of the present invention. Specific details for implementing the invention

[0026] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, it should be noted that in assigning reference numerals to the components of each drawing, the same components are given the same reference numeral whenever possible, even if they are shown in different drawings. Furthermore, in describing the present invention, if it is determined that a detailed description of related known components or functions could obscure the essence of the invention, such detailed description is omitted.

[0027] First, referring to FIG. 2, the technical configuration means of the AI-based EV fast charger power module health management system according to a preferred embodiment of the present invention is broadly composed of a multi-sensor data integration unit (100), a predictive correction unit (200) using a hybrid digital twin, an AI-based adaptive charging control system (300), and a fault self-diagnosis and external notification system (400).

[0028] Referring to FIG. 2, the multi-sensor data integration unit (100) according to an embodiment of the present invention is a means for integrating multi-sensing data including data for health management (PHM) of an EV (Electric Vehicles) fast charger power module to monitor or detect in real time, and is a means capable of measuring all of the power module performance and degradation data of the EV fast charger, the vulnerable component status data of the power conversion device, and the health data of the cooling system.

[0029] Regarding the power module performance and degradation data of the EV fast charger mentioned above, unlike slow chargers that charge using an alternating current (AC) method of 3 to 11 kW, EV (Electric Vehicles) fast chargers charge using a direct current (DC) method of 50 to 400 kW. Since these fast chargers can reduce the burden on the battery and high charging costs due to their fast speed and high current, monitoring the power module performance and degradation data of EV fast chargers is very important. In particular, fast chargers generate high-voltage DC power directly from an external source and supply it directly to the battery without passing through the vehicle's onboard charger (OBC). For this reason, charging is possible at significantly faster speeds ranging from 50 kW to ultra-fast charging of 400 kW, and using high voltage can significantly reduce power loss by reducing resistance in the charger cable and contact resistance at the connection points between terminals. Accordingly, the embodiment of the present invention is characterized by configuring a multi-sensor data integration unit (100) that enables the AI ​​Platform to quickly and accurately determine and take countermeasures from multiple detections of performance and degradation data of power modules such as a power module for DC supply of an EV fast charger, vulnerable component status data of power conversion devices such as an AC-DC converter, MC (Magnetic Contact) or VC (Vacuum, Contactor), and electrolytic capacitor, and cooling system health data such as the speed and vibration of a fan and bearing noise.

[0030] In addition, the multi-sensor data integration unit (100) according to an embodiment of the present invention further includes power module performance and degradation data (110) having a temperature-sensitive electrical parameter (TSEP) (Vce_on or Rds_on) measurement value for accurately estimating the module current, output voltage, heat sink temperature, cooling water flow rate, and instantaneous junction temperature (Tj) of the power die, vulnerable component status data (120) having temperature information of an electrolytic capacitor which is a vulnerable component of the power conversion device, and cooling system health data (130) having noise, current, and vibration information of the fan at 15,000 rpm based on bearing life.

[0031] In addition, the cooling system health data (130) of the multi-sensor data integration unit (100) according to an embodiment of the present invention further includes a reference value pre-set to detect normal operation (Normal) if the cooling water inlet temperature (Tin) is -25℃ to +55℃, warning and output limit (Derating) if it is +55℃ to +65℃, and abnormal overheating (Fault / Trip) if it is +65℃ to +70℃.

[0032] Here, the cooling system health data (130) of the multi-sensor data integration unit (100) according to the embodiment of the present invention refers to the temperature standard of a liquid-cooled power module used in an electric vehicle ultra-fast charger (100kW class or higher). Although there may be some differences depending on the manufacturer and internal components (Si MOSFET vs. SiC MOSFET), the multi-sensor data integration unit (100) detects the liquid-cooled power module temperature operating range by setting it in advance as shown in [Table 1] below, and the operation is intermittently controlled by the AI-based adaptive charging control system (300) described later.

[0033]

[0034] Generally, the set temperature standard for liquid-cooled power modules used in electric vehicle ultra-fast chargers of 100kW or higher is controlled to ensure optimal power conversion efficiency between 25°C and 45°C, thereby preventing condensation. Additionally, at temperatures below -25°C, the viscosity of the coolant (or antifreeze mixture) increases, which increases the pump load; therefore, it must be operated after preheating with a heater or the initial flow rate must be adjusted. Furthermore, the inlet water temperature must not exceed 50°C to 55°C to maintain the temperature of the internal semiconductor junction within a safety margin (approximately 20°C to 30°C). Accordingly, in the embodiment of the present invention, as shown in [Table 1] above, the state based on the set temperature standard of the liquid-cooled power module is classified into three categories: normal operation, warning and output limitation (Derating), and abnormal heating (Fault / Trip).

[0035] Referring to FIG. 2, the predictive correction unit (200) using the hybrid digital twin according to an embodiment of the present invention is a means for receiving data from the multi-sensor data integration unit (100) and predicting and preserving the lifespan of a power module through a hybrid digital twin model, and is a means for predicting and preserving the remaining useful life (RUL) of the power module of the EV fast charger through a hybrid digital twin model based on the data of the multi-sensor data integration unit (100).

[0036] In addition, the hybrid digital twin model (210) of the predictive correction unit (200) using the hybrid digital twin according to an embodiment of the present invention further includes utilizing a Physics-Informed Neural Network (PINN) and combining thermal and electrical equations of a power module package with real-time operation data to quantify the cumulative damage factors of the module and continuously predict the remaining lifespan (RUL).

[0037] Here, the Digital Twin refers to creating a three-dimensional model identical to the real world and connecting the real world and a virtual digital environment based on data. In other words, information regarding the current state of the actual power module package lifespan can be obtained through a virtualized Digital Twin instead of the actual physical asset. This has a unique feature that allows for the minimization of sudden accidents in EV fast chargers through the AI ​​Platform embedded in the multi-sensor data integration unit (100) according to the embodiment of the present invention and the AI-based adaptive charging control system (300) described later. Furthermore, the Physics-Informed Neural Network (PINN) refers to a neural network that numerically models physical phenomena in addition to the functions of an Artificial Neural Network (ANN). Through this, it repeatedly learns data patterns to reduce the amount of training data and further enhance the interpretability of the model. Accordingly, in an embodiment of the present invention, a hybrid digital twin model (210) is specifically configured within a prediction correction unit (200) using the hybrid digital twin to quantify the cumulative damage factors of the module and continuously predict the remaining lifespan (RUL) by integrating and combining the thermal and electrical equations and real-time operation data of the power module package. This has a unique feature that enables more accurate and reliable predictions by utilizing physical constraints, unlike traditional neural networks.

[0038] Referring to FIG. 2, the AI-based adaptive charging control system (300) according to an embodiment of the present invention is a means for the AI ​​Platform to optimally control the output and cooling system of the EV fast charger according to an abnormal sign of the predictive correction unit (200) using the hybrid digital twin, and is a means for optimally controlling the output power and active cooling system of the EV fast charger according to the remaining life (RUL) of the power module of the EV fast charger predicted and the criteria for operational abnormal signs.

[0039] In addition, the AI-based adaptive charging control system (300) according to an embodiment of the present invention further includes an RUL-based power modulator (310) that actively controls the operation of a charger based on the predicted remaining lifespan (RUL) and an abnormal data phenomenon judgment criterion, and when the RUL prediction falls below a dynamic threshold, the system preemptively reduces the maximum output current and / or duty cycle of the charger to mitigate a thermal change (△Tj) that causes damage; and a dynamic cooling optimization means (320) in which an AI Platform embedded in the AI-based adaptive charging control system (300) predicts future thermal load based on the vehicle's state of charge (SoC) profile and dynamically controls the pump speed and fan speed of the cooling system according to a fan speed abnormality judgment criterion to maintain the junction temperature (Tj) of the power module within the RUL optimization band.

[0040] Here, the criteria for determining abnormal data phenomena and corresponding measures according to the embodiment of the present invention are as shown in [Table 2], and the operation is intermittently controlled in the AI-based adaptive charging control system (300), and the signal generation, etc. is processed in the fault self-diagnosis and external notification system (400) described later. Derating refers to reducing the load below the rated value in order to improve reliability.

[0041]

[0042] Referring to FIG. 2, the fault self-diagnosis and external notification system (400) according to an embodiment of the present invention is a means for performing self-operation, fault removal, and emergency notification functions regarding abnormal data phenomena, wherein the AI ​​Platform of the AI-based adaptive charging control system (300) performs self-operation, fault removal, and emergency notification functions regarding abnormal data phenomena detected by the multi-sensor data integration unit (100).

[0043] In addition, the fault self-diagnosis and external notification system (400) according to an embodiment of the present invention further includes: a self-operation and fault removal means (410) that performs fault removal based on self-operation, such as a communication signal reboot, for an abnormal data phenomenon detected by an AI Platform embedded in the AI-based adaptive charging control system (300); an emergency notification and preliminary inspection request means (420) that automatically generates a preliminary inspection request for a serious temperature exceedance or a fan speed difference of 10% or more that causes a fault such as a Shut Down that cannot be resolved by self-operation; and an integrated information notification means (430) that performs the function of notifying a CPO business operator and a charger manufacturer responsible for after-sales service of abnormal information regarding a fault that cannot be resolved by self-operation, such as a communication signal reboot, generated from the emergency notification and preliminary inspection request means (420).

[0044] Here, the fault self-diagnosis and external notification system (400) according to an embodiment of the present invention has a unique feature that can resolve the problems pointed out in the prior art by not only predicting the status of the electric vehicle charger based on an artificial intelligence (AI Platform) but also attempting self-diagnosis recovery for detected faults or notifying relevant operators and manufacturers of rapid and accurate information through a server and wired / wireless communication network for unresolved faults.

[0045] In addition, the fault self-diagnosis and external notification system (400) according to an embodiment of the present invention further includes dynamically modulating the maximum output power or current profile of the charger when at least one operational abnormality is detected among a current deviation of 1% or more per module and a difference of 1% or more between the output command value and the output voltage.

[0046] In addition, the fault self-diagnosis and external notification system (400) according to an embodiment of the present invention further includes generating a warning signal when the speed of the cooling fan shows a deviation of 5% from the rated speed (e.g., 15,000 rpm), and stopping the operation of the corresponding power module when it shows a deviation of 10% or more.

[0047] In this example, the fault self-diagnosis and external notification system (400) according to the embodiment of the present invention performs fault diagnosis by an AI Platform embedded in the fault self-diagnosis and external notification device (400) according to the “criteria for determining abnormal data phenomena” of [Table 1]. At this time, for faults that are not resolved by self-operation such as a communication signal reboot, it performs a function of automatically notifying the CPO business operator and the charger manufacturer of charger abnormality information and a request for preliminary inspection, thereby having a unique feature that allows for more rapid and effective management of the health of the power module for the AI-based EV fast charger.

[0048] The foregoing description is merely an illustrative explanation of the technical concept of the present invention, and those skilled in the art to which the present invention pertains will be able to make various modifications and variations within the scope of the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are intended to explain, not limit, the technical concept of the present invention, and the scope of the technical concept of the present invention is not limited by such embodiments. The scope of protection of the present invention shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of the present invention. Explanation of the symbols

[0050] 100: Multi-sensor data integration unit 110: Power Module Performance and Degradation Data 120: Vulnerable component status data 130: Cooling system health data 200: Predictive correction unit using a hybrid digital twin 210: Hybrid Digital Twin Model 300: AI-based Adaptive Charging Control System 310: RUL-based power modulator 320: Dynamic Cooling Optimization Means 400: Fault Self-Diagnosis and External Notification System 410: Fault operation and fault removal means 420: Means for Emergency Notification and Preliminary Inspection Request 430: Integrated Information Notification Means

Claims

Claim 1 In a Prognostics and Health Management (PHM) system for an EV fast charger power module, the system comprises: a multi-sensor data integration unit (100) capable of measuring all of the following: power module performance and degradation data (110) having a measured value of a Temperature Sensitive Electrical Parameter (TSEP) (Vce_on or Rds_on) for accurately estimating the module current, output voltage, heat sink temperature, coolant flow rate, and instantaneous junction temperature (Tj) of the EV fast charger; vulnerable component status data (120) having temperature information of an electrolytic capacitor, which is a vulnerable component of a power converter; and cooling system health data (130) having fan reference speed of 10,000 to 15,000 rpm and vibration information according to the noise, current, and bearing life of the cooling fan; and a predictive system using a hybrid digital twin that predicts and maintains the remaining useful life (RUL) of the EV fast charger power module through a hybrid digital twin model based on the data of the multi-sensor data integration unit (100). An AI-based EV fast charger power module health management system characterized by including: a correction unit (200); an AI-based adaptive charging control system (300) in which an AI Platform is embedded to optimize the control of the output power and active cooling system of the EV fast charger according to the remaining lifespan (RUL) of the power module of the EV fast charger and criteria for operational abnormal signs; and a fault self-diagnosis and external notification system (400) in which the AI ​​Platform of the AI-based adaptive charging control system (300) performs self-operation, fault removal, and emergency notification functions for abnormal data phenomena detected by the multi-sensor data integration unit (100). Claim 2 delete Claim 3 In claim 1, the cooling system health data (130) of the multi-sensor data integration unit (100) further includes a reference value pre-set to detect normal operation (Normal) if the cooling water inlet temperature (Tin) is -25℃ to +55℃, warning and output limit (Derating) if it is +55℃ to +65℃, and abnormal heating (Fault / Trip) if it is +65℃ to +70℃, characterized by an AI-based EV fast charger power module health management system. Claim 4 In claim 1, the hybrid digital twin model (210) of the prediction correction unit (200) using the hybrid digital twin utilizes a Physics-Informed Neural Network (PINN) and combines thermal and electrical equations of the power module package with real-time operation data to quantify the cumulative damage factors of the module and continuously predict the remaining lifespan (RUL), thereby comprising an AI-based EV fast charger power module health management system. Claim 5 In claim 1, the AI-based adaptive charging control system (300) actively controls the operation of the charger based on the predicted remaining lifespan (RUL) and an abnormal data phenomenon judgment criterion, and when the RUL prediction falls below a dynamic threshold, the system preemptively reduces the maximum output current and / or duty cycle of the charger to mitigate the thermal cycle (△Tj) that causes damage, and further includes a dynamic cooling optimization means (320) in which an AI Platform embedded in the AI-based adaptive charging control system (300) predicts future thermal load based on the vehicle's state of charge (SoC) profile and dynamically controls the pump speed and fan speed of the cooling system according to a fan speed abnormal judgment criterion to maintain the junction temperature (Tj) of the power module within the RUL optimization band. Claim 6 In claim 1, the fault self-diagnosis and external notification system (400) further comprises: a self-operation and fault removal means (410) that performs fault removal based on self-operation, such as a communication signal reboot, for abnormal data phenomena detected by an AI Platform embedded in the AI-based adaptive charging control system (300); an emergency notification and preliminary inspection request means (420) that automatically generates a preliminary inspection request for a serious temperature overshoot or a fan speed difference of 10% or more that causes a fault such as a Shut Down that cannot be resolved by self-operation; and an integrated information notification means (430) that performs the function of notifying a CPO business operator and a charger manufacturer responsible for after-sales service of abnormal information regarding a fault that cannot be resolved by self-operation, such as a communication signal reboot, generated from the emergency notification and preliminary inspection request means (420). Claim 7 The AI-based EV fast charger power module health management system according to claim 1, wherein the fault self-diagnosis and external notification system (400) further includes dynamically modulating the maximum output power or current profile of the charger when detecting at least one operational abnormality sign among a current deviation of 1% or more per module and a difference of 1% or more between the output command value and the output voltage. Claim 8 In claim 1, the fault self-diagnosis and external notification system (400) is characterized by generating a warning signal when the speed of the cooling fan shows a deviation of 5% from the rated speed, and stopping the operation of the corresponding power module when it shows a deviation of 10% or more, in an AI-based EV fast charger power module health management system.

Images