Method for globally monitoring junction temperature of power device of motor drive in robot joint module

Abstract

The application discloses a method for globally monitoring junction temperature of power devices of motor driving in a robot joint module, and belongs to the technical field of thermal management of power electronic devices. The method comprises the following steps: collecting the on-voltage drop and on-current of the lower bridge arm MOSFET, calculating the on-resistance, combining the calibrated linear relationship, and estimating the junction temperature of the lower MOSFET in real time; fusing the NTC temperature data and the thermal coupling model, predicting the junction temperature of the upper MOSFET, and realizing the global monitoring of the power devices of the full-bridge inverter. On the hardware, the ADC channel of three-phase current sampling is multiplexed, and the voltage drop acquisition circuit containing a clamping-compensation diode is adopted to solve the high-voltage measurement problem. Experiments show that the junction temperature monitoring error of the method is ≤±2 ℃, the response time is <10 ms, no additional sensor is needed, and the system reliability and thermal management efficiency are significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of thermal management technology for power electronic devices, and relates to a method and system for global monitoring of junction temperature of power devices driven by motors in robot joint modules. Specifically, it relates to a method and system for global monitoring of junction temperature of power devices based on on-resistance, which is suitable for thermal status monitoring of high power density motor drivers. Background Technology

[0002] The internal components of a motor driver include MOSFETs (Metal-Oxide-Semiconductor Field-Effect Transistors), chips, diodes, and other components. With the increasing integration and power density of robot motor drivers, thermal management has become a more serious issue. Moore's Law states that the number of transistors that can be placed on an integrated circuit doubles every 18 months. Increased component density leads to a continuous reduction in the space occupied by individual components, a gradual shrinking of chip physical size, and a continuous improvement in circuit integration. Furthermore, as components continue to evolve towards thinner and higher-performance designs, PCB layouts and wiring are becoming increasingly dense. In addition, to meet high-performance requirements, chip operating frequencies are constantly increasing. These factors all lead to a significant increase in heat generation per unit volume, resulting in a continuous increase in heat flux density generated by the motor driver during operation, thus causing overheating problems.

[0003] Electronic components are extremely sensitive to temperature; high temperatures can have numerous negative effects, directly impacting device performance. Furthermore, sustained increases in temperature can lead to significant thermal stress, causing thermal expansion of materials, which may not only damage components but also cause irreversible damage to chips. According to the Allene equation, for every 10°C increase in ambient temperature, the lifespan of a component is halved—this is known as the "10°C rule."

[0004] Junction temperature is the highest temperature of the actual semiconductor chip (wafer, die) in an electronic device; it is typically higher than the case temperature and device surface temperature. Junction temperature measures the time required for heat dissipation from the semiconductor wafer to the packaged device casing, as well as the thermal resistance, and is a key indicator of the operating status and health of power devices. Therefore, monitoring junction temperature... T jJunction temperature is crucial for improving system performance and reliability. During driver operation, real-time monitoring of junction temperature provides data support for improving system reliability, optimizing power density, and managing available capacity. It also forms the technological foundation for device lifetime prediction and health management. Traditional junction temperature measurement methods mainly include four categories: physical contact methods, optical methods, electrothermal coupling methods, and temperature-sensitive electrical parameter (TSEP) methods.

[0005] (1) Physical contact measurement method Physical contact measurement is the most direct and convenient method, but it requires penetrating the device package, removing the thermally conductive silicone from the chip, and installing a thermistor inside the power module to directly measure the internal temperature. Thermistors are made of semiconductor materials, and their resistance changes with temperature. Thermocouple temperature measurement is based on the thermoelectric effect, using two different conductors forming a closed circuit. When a temperature difference exists between the two ends of the thermocouple, a thermoelectric electromotive force (EMF) is generated in the circuit; the corresponding temperature information can be obtained by measuring this EMF. While the physical contact method is simple, its measurement accuracy is low, it takes a long time, and it has strict requirements for the placement of the thermistor, resulting in poor practicality and operability.

[0006] (2) Optical method Optical methods use infrared detection and optical imaging techniques to receive the infrared radiation energy emitted by the power device under test (DUT) and convert it into an infrared thermogram. While accurate, this method requires access to the device's interior, opening the encapsulation and removing thermal insulation silicone to achieve optical contact. Furthermore, optical methods for measuring junction temperature require expensive testing equipment and are not suitable for online real-time junction temperature estimation.

[0007] (3) Electrothermal coupling model method The electrothermal coupling model is a non-invasive measurement method, with the advantage of not requiring direct contact with the device and having a relatively simple testing method. This method is based on heat loss and thermal impedance network models to calculate the junction temperature data of the device. Its measurement process uses an equivalent thermal path model to describe thermal characteristics, and typically uses an RC thermal impedance network for modeling and optimization. The literature (Cui H, Hu F, Zhang Y, et al. Heat spreading path optimization of IGBT thermal network model[J]. Microelectronics reliability, 2019, 103: 113511.) adjusts the thermal resistance and thermal capacity parameters by controlling the heat dissipation angle of each material layer, and introduces the interaction mechanism of heat conduction and heat convection in the Caure model, fully considering the influence of the heat dissipation effect of the base plate, thereby improving the heat transfer path and making the model more accurate than the traditional Caure model. The literature (Sang Yalei. Study on fatigue life and thermal stress reproduction of power devices in electric vehicle converters [D]. Guangzhou: South China University of Technology, 2019.) established a coupled thermal network of multi-chip packaged modules based on transient thermal impedance response and quantified the service life of power devices under actual operating conditions. The literature (Wang Zhaoping, Xin Jinlei, Du Mingxing. Construction of hybrid thermal network model and its junction temperature estimation method [J]. Journal of Power Supply, 2024, 22(03): 30-37.) combined the advantages of the two traditional thermal network models, Cauer and Foster, studied their interface methods and integrated them, and at the same time considered the aging effect of chip solder layer, proposed an improved hybrid thermal network model. The literature (Sun Xiepeng, Du Mingxing, Hu Jingwei. IGBT module junction temperature estimation and aging monitoring method based on transfer function [J]. Journal of Tianjin University of Technology, 2025, 41(01): 36-43.) takes the third-order Cauer model as an example and studies the method of extracting thermal network model parameters based on the base plate temperature transfer function, and realizes accurate junction temperature estimation.

[0008] However, the establishment of relevant thermal network models often depends heavily on the actual physical dimensions and material parameters of the chip, and relevant material and dimensional data are difficult to obtain in most applications. Based on this, the literature (Liu Ping, Li Haipeng, Miao Yiru, et al. Online extraction method of junction temperature of silicon carbide power module based on built-in temperature sensor [J]. Journal of Electrical Engineering, 2021, 36(12): 2522-2534.) proposes an online extraction method of junction temperature based on built-in negative temperature coefficient (NTC) temperature sensor, establishes a multi-chip thermally coupled thermal network model, extracts thermal network impedance parameters through finite element simulation, and verifies its stability under different boundary conditions. However, using a built-in NTC temperature sensor increases the manufacturing complexity and cost of the power module, and may also affect the monitoring results due to sensor accuracy and long-term reliability issues.

[0009] (4) Thermistor electrical parameter method Compared to other junction temperature monitoring methods, the TSEP method uses the chip itself as a temperature sensor, enabling non-invasive measurement and making it ideal for online monitoring of chip junction temperatures. Depending on the sensitive parameters, the TSEP method can be categorized into different types. Its core principle lies in utilizing the temperature-dependent characteristics of certain electrical parameters of semiconductor devices to estimate the device's junction temperature.

[0010] In power electronic devices, the short-circuit current of switching devices usually exhibits a good linear relationship with junction temperature. However, fatigue damage and aging of the bonding wires inside the chip can reduce the short-circuit current test results. To address this, the literature (Yang Shumeng, Sun Pengju, Du Xiong, et al. A method for measuring IGBT module junction temperature based on combined short-circuit current unaffected by aging[J]. Proceedings of the CSEE, 2020, 40(18): 5770-5779.) proposes a method for monitoring IGBT module junction temperature based on combined short-circuit current. Through analysis and simulation of wire cutting experiments, the influence of aging on the junction temperature measurement method is studied, and the feasibility of the method in online monitoring in converters is verified based on an experimental platform. Based on theoretical analysis and experimental results, it is found that this method can not only eliminate the influence of bonding wire aging on junction temperature measurement, but also has good linearity, high sensitivity, and online monitoring capability. However, in practical applications, this method requires the design of specific dual-level special drive circuits for different switching modules, increasing the difficulty of hardware implementation and limiting its widespread application in engineering.

[0011] Furthermore, although the threshold voltage in the temperature-sensitive parameter ( V th While exhibiting good temperature sensitivity, direct measurement is susceptible to current oscillations. Therefore, the literature (Guan Wei, Dong Lei. Threshold voltage junction temperature estimation method based on IGBT Miller platform [J]. Electrical Transmission, 2024, 54(06): 22-28.) proposes an indirect calculation method. Vth The method involves indirectly calculating the Miller plateau voltage value based on a description of the switching transient process. V th Furthermore, the literature (Zhang Qinghao, Zhang Pinjia. An online monitoring method for SiC MOSFET junction temperature based on gate voltage threshold detection [J]. Proceedings of the CSEE, 2020, 40(18): 5742-5751.) proposes a method based on... V th A novel method for monitoring junction temperature was developed, and the relationship between junction temperature and... V th The linear relationship between them was determined, and a method based on polynomial fitting was proposed. V th Measurement methods. However, this method requires an accurate description of the device's switching transient process, involving complex mathematical models, and its accuracy is highly susceptible to noise and electromagnetic interference. Furthermore, in practical applications, accurately measuring and calculating the Miller plateau voltage requires specialized test circuits and related equipment, which increases the difficulty and cost of application to some extent.

[0012] In addition to the threshold voltage ( V th Besides the junction temperature (TSEP) method, the on-state voltage drop of power devices is also an important thermistor parameter, and its value changes with junction temperature, thus it can also be used for junction temperature estimation. In practical applications, other electrical parameters closely related to junction temperature, such as turn-off voltage overshoot and turn-on transient current overshoot, can also be used as thermistor parameters for junction temperature estimation. It should be noted that although the TSEP method is simple, fast, and allows for non-invasive measurement, its estimation accuracy is easily affected by measurement errors and device aging. Furthermore, due to hardware resource limitations, the TSEP method is usually only suitable for estimating the junction temperature of a single power device, making it difficult to meet the practical needs of complex systems such as motor drives. Therefore, when applying the TSEP method for junction temperature estimation, relevant influencing factors need to be fully considered, and corresponding optimization measures should be taken to improve the applicability and reliability of the monitoring method. Summary of the Invention

[0013] The purpose of this invention is to provide a method for global monitoring of the junction temperature of power devices driven by motors in robot joint modules. This method involves extracting the on-resistance variation characteristics of MOSFETs online and establishing a data-driven thermal coupling model to estimate the junction temperature of all devices. Compared to traditional methods, this method offers higher real-time performance and insulator compatibility, and can adapt to changes in thermal characteristics under different operating conditions. The objective of this invention is achieved through the following specific technical solutions.

[0014] A method for global monitoring of junction temperature of power devices driven by motors in a robot joint module includes the following steps: S1 collects the on-state voltage drop and on-state current of the MOSFETs in the lower bridge arm of the three-phase inverter in real time. S2 calculates the real-time on-resistance of the lower bridge arm MOSFET by the ratio of on-voltage drop to on-current. S3 establishes a linear calibration model of on-resistance and junction temperature, and calculates the real-time junction temperature of the lower bridge arm MOSFET based on the real-time on-resistance. S4 uses the junction temperature data of the lower arm MOSFET and the measurement data of the negative temperature coefficient (NTC) temperature sensor to build a thermally coupled model, predict the junction temperature of the upper arm MOSFET, and realize global monitoring of the junction temperature of power devices in the motor driver.

[0015] The present invention provides a method for global monitoring of junction temperature of power devices driven by motors in robot joint modules. This method can achieve global monitoring of junction temperature of power devices driven by motors without the need for additional sensors, with the error controlled within ±2℃.

[0016] Furthermore, the on-resistance is extracted online by multiplexing the three-phase current sampling ADC (Analog-Digital Converter) channel, and current and voltage drop data are acquired synchronously through a time-division sampling strategy. By multiplexing the ADC channel, hardware resources can be saved and the addition of dedicated sampling circuitry can be avoided.

[0017] Furthermore, a low-end three-resistor sampling circuit is used to measure the three-phase current, and the on-state voltage drop is acquired during the MOSFET conduction window, solving the measurement problem of the high-voltage upper bridge arm and improving sampling safety. Even further, the on-state voltage drop acquisition circuit includes a bias current source, a clamping diode D1, and a compensation diode D2, where: D1 is used to isolate the high voltage of the bus; D2 compensates for the measurement error of D1; achieving high-precision measurement of minute voltage drops (mV level).

[0018] Furthermore, the linear calibration model in step S3 was established through calibration experiments, with an operating temperature range of 0℃ to 150℃. This linear calibration model is suitable for medium and high voltage MOSFETs and exhibits high stability of the temperature sensitivity coefficient.

[0019] Furthermore, the thermal coupling model in step S4 is generated through data-driven training, with the inputs being the lower tube junction temperature, the measurement data of the NTC temperature sensor, and the load conditions, and the output being the predicted value of the upper tube junction temperature.

[0020] This invention also provides a junction temperature monitoring system for implementing the above-described global monitoring method, comprising: a main control chip; a three-phase current sampling circuit with multiplexed ADC channels; a forward voltage drop acquisition circuit including clamping and compensation diodes; an NTC temperature sensor; and a thermally coupled model processing module. The junction temperature monitoring system provided by this invention has low hardware cost and is easily integrated into existing motor drivers.

[0021] Furthermore, the main control chip executes a time-division sampling algorithm, prioritizing the acquisition of three-phase current data for FOC control within a PWM cycle; switching to on-state voltage drop sampling during the conduction window. This ensures real-time control synchronization with temperature monitoring.

[0022] Furthermore, the NTC temperature sensor is mounted on the power module substrate to monitor changes in heat distribution in real time, thereby enhancing the dynamic response capability of the thermal coupling model.

[0023] The present invention also provides a motor driver, characterized in that it integrates the aforementioned junction temperature monitoring system. The driver can have built-in junction temperature protection logic, which, based on the monitoring results of the junction temperature monitoring system, triggers frequency reduction or shutdown when the junction temperature exceeds a threshold, thereby preventing overheating damage and improving system reliability.

[0024] The present invention offers the following beneficial technical effects: The junction temperature monitoring method and system provided by this invention not only possess high accuracy and stability but also exhibit excellent dynamic response capabilities. Compared to traditional junction temperature measurement methods, this method combines actual on-resistance circuit measurement with a data-driven model, enabling real-time acquisition of the junction temperature distribution information of the entire power device. This effectively eliminates the limitations of traditional single-point temperature measurement, avoids blind spots in junction temperature monitoring, and thus significantly improves monitoring reliability. Experimental results show that the deviation between the monitored junction temperature and the actual temperature under different load conditions is consistently controlled within ±2℃, reliably reflecting the thermal state of the MOSFET under complex operating environments. Attached Figure Description

[0025] Figure 1 It is the cell structure of a power MOSFET.

[0026] Figure 2 It relates to the junction temperature, on-state voltage drop, and current.

[0027] Figure 3 It relates to the junction temperature and on-resistance.

[0028] Figure 4 This is a schematic diagram of the conduction voltage drop acquisition principle of a three-phase inverter.

[0029] Figure 5 It is a sampling process for three-phase current and on-state voltage drop within a sampling period.

[0030] Figure 6 This is a block diagram of the hardware design structure of the motor driver in the robot joint module.

[0031] Figure 7 This is a front view of the motor driver in the robot joint module.

[0032] Figure 8 This is a reverse view of the motor driver in the robot joint module.

[0033] Figure 9 This is the junction temperature monitoring result under no-load conditions.

[0034] Figure 10 This is the junction temperature monitoring result under 50% rated load conditions.

[0035] Figure 11 This is the junction temperature monitoring result under 100% rated load conditions.

[0036] Figure 12 This is the junction temperature monitoring result under sudden load conditions.

[0037] Figure 13 This is the junction temperature monitoring result under sudden load reduction conditions. Detailed Implementation

[0038] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0039] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention 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 the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance, quantity, or position.

[0040] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" 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 mechanical connection or an electrical 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 invention according to the specific circumstances.

[0041] The cell structure of a power MOSFET is as follows Figure 1As shown. When the MOSFET is turned on, electrons start from the source (S) terminal, flow sequentially through the P-type base region, channel, JFET region, N-drift region and substrate, and finally reach the drain (D) terminal. At this time, the MOSFET exhibits resistive behavior.

[0042] MOSFET on-resistance R ds(on) It mainly consists of the following parts: leakage contact resistance R CD N + Substrate resistance R Sub Drift layer resistance R Drift JFET region resistance R JFET Accumulation layer resistance R A Channel resistance R CH Source region resistance R n+ Contact resistance of source R CS Among them, the on-resistance is mainly the channel resistance. R CH Accumulation layer resistance R A and drift layer resistance R Drift Dominance can be represented as:

[0043] In the formula, K A For current injection into the JFET region from the accumulation layer; W G This refers to the gate width; χ JP The depth of the P-region junction; u nA The electron mobility of the accumulation layer; p D The resistivity of the drift region; W cell The width of the cell; a is the width of the current cross-section in the drift region; t is the width of the drift region. Wherein, u nA and p D It will increase as the temperature rises, and V th It decreases as the temperature rises, as can be seen from equations (2-1) and (2-3). R CH It has a negative temperature sensitivity coefficient, and RDrift It has a positive temperature coefficient. For medium- and high-voltage MOSFETs, within the operating temperature range of 0°C to 150°C, R ds(on) Primarily influenced by the resistance in the drift region, it exhibits a positive temperature sensitivity coefficient, which can be approximated as... R ds(on) There is a linear positive correlation between temperature and [the relationship between temperature and temperature].

[0044] like Figure 2 The diagram shows the three-dimensional relationship between junction temperature, on-state voltage drop, and current. Through simplification, this is transformed into a two-dimensional relationship between junction temperature and on-resistance. (See diagram for details.) Figure 3 As shown, there is a clear linear correlation between junction temperature and on-resistance. Measuring on-resistance can accurately determine the junction temperature state, providing effective technical support for junction temperature monitoring and management, and is suitable for thermal management systems lacking direct temperature measurement methods.

[0045] R ds(on) Online extraction can be achieved by measuring the on-state voltage drop of the power MOSFET. V ds(on) With conduction current I ds(on) The ratio is obtained and can be expressed as:

[0046] Generally speaking, MOSFETs have a higher switching frequency and shorter conduction time compared to devices such as IGBTs, therefore, during the conduction period... I ds(on) Not much has changed. R ds(on) Online extraction V ds(on) and I ds(on) The sampling synchronization requirement is relatively low. However, compared to the transient process of a switch, the conduction time is relatively long. V ds(on) and I ds(on) Online extraction does not require high temporal resolution.

[0047] The vector control system of a PMSM motor samples and transforms the three-phase current, comparing it with a reference current to achieve closed-loop control. This system uses resistance sampling for current detection, a method also applicable to... I ds(on) Measurement.

[0048] This invention integrates on-state voltage drop sampling into the system, multiplexing it with the three-phase current sampling via an ADC sampling channel. Considering hardware resources, sampling accuracy, and time window limitations, this invention employs a low-end three-resistor sampling method to measure the three-phase current. Because the MOSFET on-state voltage drop is extremely small and difficult to measure directly, such as... Figure 4 The circuit shown provides a relatively simple and effective testing scheme. It isolates the drain-source voltage when the MOSFET is off and performs measurements when the MOSFET is on. The measurement circuit consists of a bias current source and diodes D1 and D2 connected in series. The current source provides a stable current to ensure the diodes are forward-biased; diode D1 clamps the voltage, isolating the high voltage on the bus and preventing damage to subsequent circuits due to overvoltage; diode D2 compensates for measurement errors introduced by D1, improving the accuracy and precision of the measurement.

[0049] The on-resistance of MOSFETs in a three-phase inverter can be calculated by separately collecting their on-current and voltage drop. However, this method is mainly applicable to the measurement of the lower-arm MOSFETs, while the junction temperature information of the upper-arm MOSFETs is difficult to obtain. Relying solely on traditional junction temperature measurement methods based on on-resistance makes it difficult to achieve global monitoring of the power devices in the entire inverter system.

[0050] To address this technical challenge, this invention combines historical temperature data from multiple power devices with measurement results from negative temperature coefficient (NTC) resistors (i.e., NTC temperature sensors). This method fully leverages thermal coupling relationships, allowing for the prediction of the junction temperature of the upper MOSFET by directly measuring the temperature data of the lower MOSFET and the NTC resistor, thus achieving comprehensive monitoring of the entire power module. Due to the thermal coupling effect between MOSFETs, the temperature of the lower MOSFET directly affects not only the junction temperature of the upper MOSFET but also alters the overall thermal distribution of the inverter. The NTC resistor, with its negative temperature coefficient, can rapidly respond to temperature changes, accurately capturing temperature variations in power devices and playing a crucial role in MOSFET junction temperature monitoring. By combining historical experimental data, a mathematical model can be established between the junction temperatures of the upper and lower MOSFETs and the NTC resistor data, enabling real-time monitoring of the global power device junction temperature and timely detection of potential over-temperature issues.

[0051] The junction temperature monitoring and control algorithm calculates the on-resistance based on real-time detection of the MOSFET's on-state voltage drop and current, thereby estimating the junction temperature. To avoid increasing hardware costs and implement this function with a single MCU, the analog-to-digital converter ADC_123 channel is reused to multiplex the sampling of the three-phase on-state voltage drop and current, effectively calculating the on-resistance to estimate the MOSFET's junction temperature information, thus ensuring the accuracy of the overall sampling circuit and its real-time feedback capability.

[0052] like Figure 5 The diagram illustrates the sampling process for three-phase current and on-state voltage drop within a single sampling cycle. First, by setting Sample_Flag=1, three-phase current data is prioritized for acquisition, a key parameter for efficient FOC control. After current sampling, the system resets the sampling flag to Sample_Flag=0 and begins acquiring on-state voltage drop data. This involves a series of signal processing steps, including compensation, isolation, amplification, conditioning, and noise filtering. The processed sampled data is transmitted to the MCU. Based on the established linear model of on-resistance and junction temperature, the sampled data is used for real-time MOSFET temperature monitoring, and global junction temperature monitoring is performed based on the aforementioned thermal coupling model. In this sampling algorithm, the acquisition and processing of current and on-state voltage drop data are closely integrated, ensuring not only the accuracy of motor control but also providing reliable data support for MOSFET temperature monitoring. This is crucial for achieving efficient motor control and effective MOSFET temperature monitoring.

[0053] The overall hardware design structure of the motor driver in the robot joint module is as follows: Figure 6 As shown, its core component is a minimal system built with an STM32F446RET6 main control chip, responsible for the central control of the entire drive system. In addition, the driver includes a gate driver, a three-phase inverter power module, a communication module, a position sensor, a sampling circuit, and corresponding protection circuits. The pulse width modulation (PWM) signal generated by the main control chip is amplified by the drive circuit and can effectively control the switching of the transistors in the power module. The power module obtains three-phase current data through current sampling and combines it with the FOC algorithm to adjust the output of the PWM signal to achieve precise motor control. The encoder is responsible for collecting and monitoring the motor's position information in real time, and transmitting it back to the main control chip for processing via an RS-485 communication interface. After receiving the data, the main control chip uses it to calculate and adjust the motor's speed and torque in real time to adapt to different loads and operating conditions.

[0054] The actual motor driver in the robot joint module provided by this invention is shown below. Figure 7 and Figure 8 The image shows a high-performance MOSFET driver. This driver features a wide operating voltage range and excellent heat dissipation. It also incorporates a monitoring function that can monitor voltage, current, and critical junction temperature parameters in real time to prevent overheating and overload, ensuring long-term stable operation of the motor driver and thus improving the overall system stability.

[0055] To fully simulate the various load conditions that MOSFETs may encounter in practical applications, experiments were conducted under different operating conditions to monitor the junction temperature of the power devices in the motor driver. By operating the motor driver under different load conditions, including constant operating conditions from low to high, as well as environments with sudden load increases and decreases, the applicable conditions that it might encounter in practical applications were simulated, thereby verifying the applicability of the proposed method. The core objective of the experiments was to verify whether the junction temperature data obtained by the junction temperature monitoring method based on on-resistance of this invention matches the actual junction temperature data. The junction temperature data obtained by the junction temperature monitoring method provided by this invention was compared with the maximum junction temperature data of six MOSFETs obtained directly through thermocouple measurements to evaluate the accuracy and reliability of this improved junction temperature monitoring method.

[0056] Experiments were conducted under different operating conditions: no load, 50% and 100% rated load. The experimental results show that the steady-state junction temperature of the MOSFET exhibits a significant upward trend as the power load gradually increases. The experimental results are as follows: Figures 9-11 As shown, under different load conditions, the junction temperature monitoring results are highly consistent with the actual junction temperature change curves, and the error in estimating the junction temperature is always controlled within ±2℃, indicating that the junction temperature monitoring method provided by this invention can effectively reflect the actual junction temperature of the MOSFET.

[0057] Experiments were conducted under two different operating conditions: sudden load increase and sudden load decrease. The experimental results are as follows: Figure 12 , Figure 13 As shown in the figure. Experimental results show that under conditions of sudden load increase and decrease, the junction temperature monitoring value can accurately track the actual junction temperature change and respond quickly at the moment of load change. By comparing the junction temperature fluctuation of MOSFET monitoring in real time, the applicability and high accuracy of this junction temperature monitoring method under dynamic load conditions are verified, ensuring that the junction temperature monitoring results under different operating conditions are not distorted.

[0058] Although embodiments of the present invention have been shown and described above, it is understood that these embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and alterations to the above embodiments within the scope of the present invention without departing from its principles and spirit. The scope of protection of the present invention is defined by the claims and their equivalents.

Claims

1. A method for global monitoring of junction temperature of power devices driven by motors in a robot joint module, characterized in that, Includes the following steps: S1 collects the on-state voltage drop and on-state current of the MOSFETs in the lower bridge arm of the three-phase inverter in real time. S2 calculates the real-time on-resistance of the lower bridge arm MOSFET by the ratio of on-voltage drop to on-current. S3 establishes a linear calibration model of on-resistance and junction temperature, and calculates the real-time junction temperature of the lower bridge arm MOSFET based on the real-time on-resistance. S4 utilizes the junction temperature data of the lower bridge arm MOSFET and the measurement data of the NTC temperature sensor, which is mounted on the power module substrate, to build a thermal coupling model and predict the junction temperature of the upper bridge arm MOSFET, thereby achieving global monitoring of the junction temperature of power devices in the motor driver.

2. The global monitoring method according to claim 1, characterized in that, The on-resistance is extracted online by multiplexing the ADC channel of the three-phase current sampling, and the current and voltage drop data are obtained within one PWM cycle through a time-division sampling strategy.

3. The global monitoring method according to claim 1, characterized in that, A low-end three-resistor sampling circuit is used to measure the three-phase current, and the on-state voltage drop is collected during the MOSFET conduction window.

4. The global monitoring method according to claim 3, characterized in that, The on-state voltage drop acquisition circuit includes a bias current source, a clamping diode D1, and a compensation diode D2, wherein: D1 is used to isolate the high voltage of the bus; D2 is used to compensate for the measurement error of D1.

5. The global monitoring method according to claim 1, characterized in that, The linear calibration model in step S3 is established through calibration experiments, with an operating temperature range of 0℃ to 150℃.

6. The global monitoring method according to claim 1, characterized in that, The thermal coupling model in step S4 is generated through data-driven training. The inputs are the junction temperature data of the lower bridge arm MOSFET, the measurement data of the NTC temperature sensor, and the load conditions. The output is the predicted junction temperature value of the upper bridge arm MOSFET.

7. A junction temperature monitoring system for implementing the global monitoring method according to any one of claims 1-6, characterized in that, include: Main control chip; A three-phase current sampling circuit that reuses ADC channels; A forward voltage drop acquisition circuit including clamping and compensation diodes; NTC temperature sensor; Thermally coupled model processing module.

8. The junction temperature monitoring system according to claim 7, characterized in that, The main control chip executes a time-division sampling algorithm within one PWM cycle: Prioritize the acquisition of three-phase current data for FOC control; During the conduction window, switch to conduction voltage drop sampling.

9. The junction temperature monitoring system according to claim 7, characterized in that, The NTC temperature sensor is mounted on the power module substrate to monitor changes in heat distribution in real time.

10. A motor driver, characterized in that, Integrate the junction temperature monitoring system according to any one of claims 7-9.

Images