A control method for dynamically balancing temperature difference between upper and lower bridge arms of a power module

Abstract

This invention belongs to the field of temperature difference control technology for upper and lower bridge arms of a power module, and particularly relates to a method for dynamically balancing the temperature difference between the upper and lower bridge arms. The method includes: S1, acquiring the temperatures of the upper and lower bridge arms; S2, calculating the temperature difference and the rate of change of the temperature difference between the upper and lower bridge arms; wherein, T1 equals the upper bridge arm temperature minus the lower bridge arm temperature; if T1 < T2, entering monitoring mode and recording only data; if T1 ≤ < T2, entering a mild adjustment mode and proceeding to S3; if T1 ≥ T2, entering an aggressive adjustment mode and proceeding to S4; wherein T1 and T2 are preset first and second temperature difference thresholds, respectively; S3, reducing the temperature difference by adjusting the PWM duty cycle and dead time; S4, reducing the temperature difference by switching the operating mode and adjusting the switching frequency. This method can achieve active, rapid, and accurate dynamic balancing of the temperature difference between the upper and lower bridge arms without affecting system performance.

Description

Technical Field

[0001] This invention belongs to the field of temperature difference control technology for upper and lower bridge arms of a module, and particularly relates to a method for controlling the temperature difference between the upper and lower bridge arms of a dynamic equalization power module. Background Technology

[0002] In modern power electronic systems, bridge power conversion circuits (such as inverters and motor drivers) are widely used in industrial drives, new energy power generation, and electric vehicles. These circuits typically use insulated-gate bipolar transistors (IGBTs) as the main switching devices, forming upper and lower bridge arm structures to achieve bidirectional energy flow and efficient conversion. However, due to significant differences in operating modes, current paths, voltage stress, and heat dissipation conditions between the upper and lower bridge arms, the problem of uneven temperature distribution during operation is becoming increasingly prominent.

[0003] Specifically, the upper IGBT experiences a higher bus voltage during switching and faces significant voltage spikes and switching losses at turn-off. The lower IGBT, on the other hand, may experience additional conduction losses during freewheeling due to diode reverse recovery current, particularly under high-frequency, heavy-load conditions. Furthermore, differences in packaging structure, heat dissipation channel layout, and thermal resistance characteristics between the upper and lower IGBTs lead to varying rates of heat accumulation, further exacerbating the temperature difference. This persistent temperature difference not only causes inconsistent device aging rates but can also trigger localized thermal failures, reducing the overall reliability and lifespan of the system.

[0004] To address the aforementioned issues, existing technologies mostly employ passive or semi-active strategies. For example, they improve overall thermal management performance by optimizing heat dissipation design (such as increasing heatsink area and improving airflow structure); or they use fixed dead-time compensation methods to mitigate the impact of uneven switching losses. Some advanced solutions introduce temperature sensors for real-time monitoring and combine them with over-temperature protection mechanisms to implement derating control. However, these methods generally have the following limitations:

[0005] 1. Passive Derating and Fixed Compensation: Traditional methods rely on power limiting or preset duty cycle offsets triggered by high temperatures, which is a typical "post-event response" strategy. This static adjustment method cannot dynamically adjust according to load changes, ambient temperature fluctuations, or transient conditions, often sacrificing system efficiency for stability, making it difficult to adapt to complex and ever-changing application scenarios.

[0006] 2. Limitations of Single-Parameter Adjustment: Most solutions only affect the temperature difference by adjusting a single control parameter (such as PWM duty cycle or dead time), lacking the ability to coordinate and optimize multi-dimensional control variables. This can easily lead to problems such as output voltage waveform distortion, increased harmonic content, and enhanced electromagnetic interference (EMI), and may even affect the power quality at the load end.

[0007] 3. Insufficient response speed: Although some linkage controls based on heat dissipation systems (such as fan speed regulation and liquid cooling flow regulation) can alleviate temperature rise to a certain extent, their mechanical response delay is generally at the level of hundreds of milliseconds, which is far higher than the dynamic process of high-frequency switching devices (such as kHz to MHz level), making it difficult to effectively suppress instantaneous temperature rise events.

[0008] 4. Limited Monitoring Accuracy: Current temperature sensing methods mainly rely on external sensors or estimation methods based on junction temperature models. The former is greatly affected by factors such as package thermal resistance drift and installation position deviation, while the latter is easily affected by parasitic parameter changes, aging effects, and modeling errors, resulting in temperature measurement errors generally exceeding ±5°C, which seriously affects control accuracy.

[0009] Therefore, how to achieve active, rapid, and accurate dynamic balancing of the temperature difference between the upper and lower bridge arms without affecting system performance has become an urgent problem to be solved. Summary of the Invention

[0010] To address the shortcomings of the existing technology, this invention provides a method for controlling the temperature difference between the upper and lower bridge arms of a power module, which can achieve active, rapid, and accurate dynamic balancing of the temperature difference between the upper and lower bridge arms without affecting system performance.

[0011] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0012] A method for controlling the temperature difference between the upper and lower bridge arms of a dynamic equalization power module includes the following steps:

[0013] S1. Obtain the temperatures of the upper and lower bridge arms;

[0014] S2. Calculate the temperature difference between the upper and lower bridge arms. and the rate of change of temperature difference ;in, It equals the upper arm temperature minus the lower arm temperature; if If T1 < T1, then enter monitoring mode and only record data; if T1 ≤ T1 If <T2, then enter the gentle adjustment mode and switch to S3; if If T1 is greater than or equal to T2, then enter the aggressive adjustment mode and switch to S4; where T1 and T2 are the preset first temperature difference threshold and second temperature difference threshold, respectively.

[0015] S3. Reduce temperature difference by adjusting PWM duty cycle and dead time;

[0016] When adjusting the PWM duty cycle, the total duty cycle of the upper and lower bridge arms remains unchanged, and the single-sided offset does not exceed ±3%.

[0017] When adjusting the dead time, the adjusted dead time should be greater than or equal to the preset typical value;

[0018] S4. Reduce temperature difference by switching working modes and adjusting switching frequency;

[0019] When switching operating modes, the active switching roles of the upper and lower bridge arms are periodically interchanged; when adjusting the switching frequency, the switching frequency on the high-temperature side is reduced according to a preset method.

[0020] Compared with the prior art, the present invention has the following advantages:

[0021] 1. A graded response mechanism based on temperature difference. This method first acquires the real-time temperatures of the upper and lower bridge arms and calculates the temperature difference ΔT and its rate of change d(ΔT) / dt. Based on two preset temperature difference thresholds T1 and T2, the system is divided into three operating modes: monitoring mode, mild adjustment mode, and aggressive adjustment mode. This graded response mechanism can distinguish the severity of temperature differences, avoiding excessive intervention when the temperature difference is slight, while rapidly initiating strong regulation when the temperature difference increases significantly. Compared with traditional fixed thresholds or single protection strategies, it has stronger adaptability and robustness.

[0022] 2. Dual-parameter coordinated adjustment to reduce temperature difference. In mild adjustment mode, the loss distribution of the upper and lower bridge arms is optimized by simultaneously adjusting the PWM duty cycle and dead time. The duty cycle adjustment keeps the total duty cycle constant, allowing only a single-sided deviation of no more than ±3%, effectively avoiding output voltage distortion. The dead time adjustment ensures it is not lower than the typical value, preventing shoot-through risk. This dual-parameter coordinated adjustment method overcomes the limitations of traditional methods that only adjust the duty cycle or only adjust the dead time, achieving fine-grained allocation of switching losses without compromising output waveform quality, significantly improving control flexibility and system stability.

[0023] 3. Operating Mode Switching and Dynamic Switching Frequency Adjustment. When the temperature difference reaches a high level, an aggressive adjustment mode is entered, employing deeper control methods: periodically switching the active switching roles of the upper and lower bridge arms (i.e., operating mode switching), allowing devices that have been continuously subjected to high stress to "rest"; simultaneously, reducing the switching frequency of the high-temperature side to decrease its switching losses. This strategy reconstructs the heat load distribution at the device usage logic level, avoiding long-term unidirectional heat accumulation and overcoming the passive response methods of traditional approaches that rely solely on heat dissipation or derating, achieving active thermal equilibrium without sacrificing output capability.

[0024] 4. Proactive control based on temperature difference trends. This method not only focuses on the current magnitude of the temperature difference but also introduces the rate of temperature difference change as a judgment criterion, enabling intervention and regulation before the temperature difference reaches a dangerous level. This gives the system a certain degree of predictability and feedforward capability, significantly improving response speed. Compared with traditional solutions that rely on cooling system linkage or delayed protection, it can suppress the temperature rise trend earlier, making it particularly suitable for transient thermal events in high-frequency switching scenarios.

[0025] In summary, this method constructs a closed-loop, adaptive temperature difference balancing system through multiple means, including multi-level state judgment, dual-parameter coordinated adjustment, operating mode switching, and dynamic frequency adjustment. It can not only actively identify and respond to temperature difference changes, but also achieve precise control without affecting output voltage quality, electromagnetic compatibility, or system efficiency. Therefore, this method can achieve proactive, rapid, and precise dynamic balancing of the temperature difference between the upper and lower bridge arms without impacting system performance.

[0026] Preferably, in S1, the temperatures of the upper and lower bridge arms are obtained through hybrid temperature sensing technology.

[0027] Preferably, the process of obtaining the temperature of the upper arm / lower arm includes:

[0028] First, obtain direct measurements of the bridge arm temperature. ;

[0029] Then, calculate the indirect estimate of the bridge arm temperature: ;

[0030] in, The substrate temperature; Thermal resistance to the junction shell; Total power loss;

[0031] Finally, the directly measured values Compared with indirect estimates Input the Kalman filter and output the fused temperature value, which will be used as the temperature in subsequent steps.

[0032] This setup combines direct measurement and indirect estimation to obtain bridge arm temperature information. Specifically, the direct measurement values... Typically derived from sensors mounted on a substrate or housing (such as thermistors, infrared sensors, etc.), these sensors offer advantages such as fast response and high real-time performance; while indirect estimation values... Based on the thermal resistance model and total power loss The calculations show that the temperature changes within the device's junction are reflected. This "multi-source heterogeneous" sensing strategy overcomes the limitations of single sensors, which are susceptible to environmental interference, positional deviations, or aging, and achieves complementarity and enhancement of temperature information.

[0033] 2. This model fully utilizes the physical characteristics of power devices, enabling dynamic estimation of critical thermal node temperatures without the need for direct junction temperature sensors. Compared to rough estimations relying solely on surface temperature, this method more closely approximates actual thermal stress conditions, making it particularly suitable for early warning and protection in high-temperature, high-power-density scenarios.

[0034] 3. Directly measured values Compared with indirect estimates The input is a Kalman filter, and the two parameters are weighted and fused using optimal estimation theory to output a more accurate, smoother, and noise-resistant temperature result. The Kalman filter possesses excellent dynamic tracking capabilities and noise suppression performance, continuously optimizing temperature estimation accuracy even under conditions of measurement error, model uncertainty, and external disturbances. Compared to simple static fusion methods such as averaging or taking the maximum value, this method significantly improves the reliability and stability of the temperature signal.

[0035] Preferably, the total power loss The formula for calculation is:

[0036] ;

[0037] In the formula, Phase current; For on-resistance; The switching frequency; , For switching energy.

[0038] This setup allows for: 1. Distinguishing between conduction losses and switching losses to achieve refined modeling. The formula decomposes the total power loss into two parts: one is the conduction loss determined by the square of the current and the on-resistance (…). The second is the switching loss, which is the product of the switching frequency and the energy of a single switch. This classification modeling approach accurately reflects the energy consumption mechanism of power devices at different operating stages, avoiding errors caused by traditional empirical formulas or simplified models. Especially in high-switching-frequency, high-current applications, it can more realistically reflect the heat load distribution of the devices.

[0039] 2. Introducing the effective value of phase current improves the accuracy of conduction loss calculation. Instead of using peak or average current to calculate conduction losses, this design fully considers the current waveform distortion and harmonic components under AC loads, making the conduction loss estimation closer to actual operating conditions. Compared to methods that only use DC components or peak current, this design significantly improves the calculation accuracy under complex load conditions such as variable frequency speed regulation and PWM control, thus providing a more reliable basis for subsequent temperature rise prediction and thermal management strategies.

[0040] 3. The calculation is simplified while maintaining accuracy by using switching energy instead of voltage-current integration. Switching losses are accounted for... , This indicates that the energy consumed during a single turn-on and turn-off process is represented, respectively. These parameters can usually be obtained from device datasheets or experimental calibration, eliminating the need for real-time integration of voltage-current curves, thus reducing algorithm complexity and computational resource requirements. Furthermore, because... and It already takes into account the effects of factors such as voltage, current change rate, and parasitic inductance, so it can still reflect the energy loss in the actual switching process well, balancing accuracy and efficiency.

[0041] 4. Supports online power loss estimation under dynamic operating conditions. All parameters can be acquired in real time through system sensors or controllers. and It can be pre-calibrated or corrected for temperature. This makes the entire formula applicable to dynamic power loss calculations under online operating conditions, and can be used in multiple aspects such as closed-loop thermal management, efficiency optimization, and fault diagnosis. Compared with static or offline estimation methods, this model has stronger real-time performance and adaptability.

[0042] Preferably, in S3, the process of adjusting the PWM duty cycle includes:

[0043] The duty cycle offset is calculated using an anti-saturation PID algorithm. , It consists of proportional terms, integral terms, and differential terms;

[0044] ;

[0045] in, , , These are proportional gain, integral gain, and derivative gain, respectively.

[0046] If the temperature of the upper bridge arm is higher, the adjusted duty cycle is:

[0047] ;

[0048] In the formula, , These are the upper bridge boom before and after adjustment, respectively. , These are the lower bridge arm before and after adjustment, respectively.

[0049] If the temperature of the lower bridge arm is higher, the adjusted duty cycle is:

[0050] .

[0051] This setup, 1) introduces an anti-saturation PID control algorithm to calculate the duty cycle offset ΔD during PWM duty cycle adjustment. This algorithm comprehensively considers the current temperature difference, historical cumulative temperature difference, and temperature difference trend, enabling rapid response to temperature deviations and suppression of overshoot. Specifically, the "anti-saturation" design prevents the integral term from accumulating excessively over long-term deviations, thus avoiding controller output distortion and improving system stability and long-term regulation performance. Compared to traditional proportional control or simple feedback mechanisms, this method possesses stronger dynamic tracking and anti-interference capabilities, making it particularly suitable for scenarios with frequent load fluctuations or rapid temperature rise rates.

[0052] 2. Directional duty cycle allocation achieves directional thermal equilibrium. Depending on whether the upper or lower bridge arm has a higher temperature, ΔD is added to the other side or subtracted from the higher-temperature side. That is, if the upper bridge arm is hotter, its duty cycle is reduced, and the lower bridge arm's duty cycle is increased; conversely, the lower arm's duty cycle is reduced. This directional adjustment strategy achieves dynamic load transfer, where "the hotter component works less, and the colder component bears more," directly affecting the device's conduction time and effectively altering its conduction loss distribution. Compared to methods relying solely on heat dissipation or passive temperature equalization, this method intervenes at the source of power distribution, resulting in higher regulation efficiency and proactivity.

[0053] Preferably, in S3, when adjusting the dead time, the dead time is adjusted according to the following formula:

[0054] ;

[0055] In the formula, and These are the dead time before and after the adjustment, respectively. This is for adjusting the coefficient.

[0056] This design dynamically adjusts the dead time according to the temperature difference: when a bridge arm has a higher temperature, its corresponding dead time is shortened, thereby increasing its on-time window, reducing the switching frequency or turn-off losses, and indirectly achieving "compensatory" excitation of the devices on that side. Conversely, the dead time is extended to limit its turn-on opportunities. This directional adjustment mechanism achieves dynamic "load transfer" or "operational throttling" of the high-temperature side devices, effectively alleviating local heat accumulation.

[0057] Preferably, in S4, when switching operating modes, the switching timing is also managed by a state machine to avoid output voltage jumps.

[0058] This configuration, by introducing a state machine to precisely manage the switching sequence during the operating mode transition, ensures that the output voltage does not suddenly change or jump when the roles of the upper and lower bridge arms are reversed. This solves the voltage oscillation or current surge problems that may be caused by traditional switching methods, and improves the electromagnetic compatibility and operational stability of the system.

[0059] Preferably, in S4, when adjusting the switching frequency, the switching frequency of the high-temperature side bridge arm is updated according to the following calculation formula. :

[0060] ;

[0061] in, This is the default value.

[0062] This setup directly links the switching frequency to the temperature difference in a non-linear manner—the greater the temperature difference, the more significant the frequency drop, creating an adaptive control mechanism that slows down as the temperature rises. Compared to traditional fixed-frequency reduction or step-wise adjustment, this method can more accurately match the temperature rise trend, achieving smoother and more efficient heat dissipation control, while avoiding harmonic disturbances or increased electromagnetic interference caused by sudden frequency changes.

[0063] Preferably, in S1, after obtaining the temperatures of the upper and lower bridge arms, it is also monitored whether the temperature of the upper or lower bridge arm exceeds the warning threshold. If either bridge arm exceeds the threshold, the derating protection is triggered and the fault is recorded.

[0064] This setup, through a real-time monitoring mechanism, can intervene promptly before devices enter hazardous operating areas, preventing permanent damage or sudden failure due to overheating. Compared to traditional methods that rely solely on final shutdown protection, this solution provides a tiered safety guarantee of "early warning—derating—protection," significantly improving the system's ability to cope with sudden operating conditions such as transient overloads and abnormal heat dissipation.

[0065] Preferably, in S1, it is also monitored whether the Vds voltage spike of the upper or lower bridge arm exceeds a preset spike threshold. If either bridge arm exceeds the threshold, the PWM output of the entire bridge arm is turned off.

[0066] Such a setup, due to factors such as parasitic inductance, sudden load changes, or drive mismatch, may cause significant Vds voltage spikes in power devices during turn-off. Although these transient overvoltages are short-lived, they can easily lead to avalanche breakdown or long-term insulation degradation. This solution, by monitoring Vds voltage in real time, can accurately capture these transient high-voltage events and react within microseconds, preventing devices from entering hazardous operating areas. Compared to traditional methods that rely on post-fault diagnosis or periodic maintenance, this approach offers stronger foresight and protection capabilities. Attached Figure Description

[0067] To make the objectives, technical solutions, and advantages of the invention clearer, the invention will now be described in further detail with reference to the accompanying drawings, wherein:

[0068] Figure 1 This is a schematic diagram of the architecture when this method is specifically applied;

[0069] Figure 2 This is a flowchart of the method. Detailed Implementation

[0070] The following detailed explanation illustrates the specific implementation methods:

[0071] Example 1

[0072] It should be noted that, in practice, this method can be fully automated through a temperature sensing layer, a temperature difference decision layer, and a dynamic execution layer, as shown in the following architecture: Figure 1 As shown.

[0073] like Figure 1 , Figure 2 As shown, this embodiment discloses a method for controlling the temperature difference between the upper and lower bridge arms of a dynamic equalization power module, including the following steps:

[0074] S1. Obtain the temperature of the upper and lower bridge arms through hybrid temperature sensing technology.

[0075] The process of obtaining the temperature of the upper arm / lower arm includes:

[0076] First, obtain direct measurements of the bridge arm temperature. ;

[0077] Then, calculate the indirect estimate of the bridge arm temperature: ;

[0078] in, The substrate temperature; Thermal resistance to the junction shell; Total power loss;

[0079] Finally, the directly measured values Compared with indirect estimates Input the Kalman filter and output the fused temperature value, which will be used as the temperature in subsequent steps.

[0080] In this way, the scheme combines direct measurement and indirect estimation to obtain bridge arm temperature information. Among them, the direct measurement values... Typically derived from sensors mounted on a substrate or housing (such as thermistors, infrared sensors, etc.), these sensors offer advantages such as fast response and high real-time performance; while indirect estimation values... Based on the thermal resistance model and total power loss The calculations show that the temperature can reflect the changing trend of the junction temperature inside the device. This "multi-source heterogeneous" sensing strategy overcomes the limitations of single sensors, which are susceptible to environmental interference, positional deviations, or aging, and achieves complementarity and enhancement of temperature information. Furthermore, this model fully utilizes the physical characteristics of power devices, enabling dynamic estimation of key thermal node temperatures even without a direct junction temperature sensor. Compared to a rough judgment relying solely on surface temperature, this method is closer to the actual thermal stress state, and is particularly suitable for early warning and protection in high-temperature, high-power-density scenarios. Moreover, the directly measured values... Compared with indirect estimates The input is a Kalman filter, and the two parameters are weighted and fused using optimal estimation theory to output a more accurate, smoother, and noise-resistant temperature result. The Kalman filter possesses excellent dynamic tracking capabilities and noise suppression performance, continuously optimizing temperature estimation accuracy even under conditions of measurement error, model uncertainty, and external disturbances. Compared to simple static fusion methods such as averaging or taking the maximum value, this method significantly improves the reliability and stability of the temperature signal.

[0081] In practical implementation, total power loss The formula for calculation is:

[0082] ;

[0083] In the formula, Phase current; For on-resistance; The switching frequency; , For switching energy.

[0084] This formula decomposes the total power loss into two parts: one is the conduction loss determined by the square of the current and the on-resistance (…). The second is the switching loss, which is the product of the switching frequency and the energy of a single switch. This classification modeling approach accurately reflects the energy consumption mechanism of power devices at different operating stages, avoiding errors caused by traditional empirical formulas or simplified models. Especially in high-switching-frequency, high-current applications, it can more realistically reflect the heat load distribution of the devices. Furthermore, using... Instead of using peak or average current to calculate conduction losses, this design fully considers the current waveform distortion and harmonic components under AC loads, making the conduction loss estimation closer to actual operating conditions. Compared to methods that only use DC components or peak current, this design significantly improves the calculation accuracy under complex load conditions such as variable frequency speed regulation and PWM control, thus providing a more reliable basis for subsequent temperature rise prediction and thermal management strategies.

[0085] Furthermore, the switching loss portion is... , This indicates that the energy consumed during a single turn-on and turn-off process is represented, respectively. These parameters can usually be obtained from device datasheets or experimental calibration, eliminating the need for real-time integration of voltage-current curves, thus reducing algorithm complexity and computational resource requirements. Furthermore, because... and It already incorporates the effects of factors such as voltage, current change rate, and parasitic inductance, thus still reflecting energy losses in actual switching processes relatively well, balancing accuracy and efficiency. Furthermore, it supports online estimation of power losses under dynamic operating conditions. All parameters can be acquired in real time through system sensors or controllers. and It can be pre-calibrated or corrected for temperature. This makes the entire formula applicable to dynamic power loss calculations under online operating conditions, and can be used in multiple aspects such as closed-loop thermal management, efficiency optimization, and fault diagnosis. Compared with static or offline estimation methods, this model has stronger real-time performance and adaptability.

[0086] To facilitate understanding, a diagram is provided.

[0087] Initialize hardware and configure control parameters: After power-on, initialize the main control chip, configure the PWM module, ADC sampling channel (corresponding to NTC sensor), and dead time; set the protection parameters of the IGBT driver chip: desaturation detection threshold. Fault response time ;

[0088] Temperature model loading: Read preset IGBT temperature estimation parameters from RAM, thermal resistance On resistance Switch energy , .

[0089] Initialize Kalman filter parameters: process noise covariance Q = 0.01, measurement noise covariance R = 0.1.

[0090] Real-time temperature monitoring and fusion: ADC sampling is triggered every 50μs to read the NTC resistance values ​​of the upper and lower bridge arms (corresponding to the temperature). , The substrate temperature was obtained after linearization calibration. Real-time acquisition of phase current (via current sensor) and switching frequency Calculate the total loss: Calculate the junction temperature based on the thermal resistance model. Directly measured values ​​( ) and indirect estimates ( Input to a Kalman filter, output the fused temperature value. , The error is <1°C.

[0091] S2. Calculate the temperature difference between the upper and lower bridge arms. and the rate of change of temperature difference ;in, It equals the upper arm temperature minus the lower arm temperature; if If T1 < T1, then enter monitoring mode and only record data; if T1 ≤ T1 If <T2, then enter the gentle adjustment mode and switch to S3; if If T1 is greater than or equal to T2, then enter the aggressive adjustment mode and switch to S4; where T1 and T2 are the preset first temperature difference threshold and second temperature difference threshold, respectively.

[0092] In practice, the temperature difference can be calculated every 100 μs. and rate of change .

[0093] S3. Reduce temperature difference by adjusting PWM duty cycle and dead time;

[0094] When adjusting the PWM duty cycle, the total duty cycle of the upper and lower bridge arms remains unchanged, and the single-sided offset does not exceed ±3%; when adjusting the dead time, the adjusted dead time is greater than or equal to the preset typical value.

[0095] In practice, T1 is 2℃ and T2 is 5℃. For ease of understanding, the hierarchical control strategy is illustrated in Table 1.

[0096] Table 1

[0097]

[0098] In practice, the process of adjusting the PWM duty cycle includes:

[0099] The duty cycle offset is calculated using an anti-saturation PID algorithm. , It consists of proportional terms, integral terms, and differential terms;

[0100] ;

[0101] in, , , These are proportional gain, integral gain, and derivative gain, respectively.

[0102] In practical implementation, typical parameters are: .

[0103] In practice, restrictions are also imposed. To prevent over-debugging.

[0104] If the temperature of the upper bridge arm is higher, the adjusted duty cycle is:

[0105] ;

[0106] In the formula, , These are the upper bridge boom before and after adjustment, respectively. , These are the lower bridge arm before and after adjustment, respectively.

[0107] If the temperature of the lower bridge arm is higher, the adjusted duty cycle is:

[0108] .

[0109] For specific adjustments: adjust according to the temperature difference between the upper and lower bridge arms. Calculate the offset Then, the conduction time of each arm is finely adjusted to balance the temperature of the upper and lower arms while keeping the total duty cycle constant.

[0110] During the adjustment of the PWM duty cycle, an anti-saturation PID control algorithm is introduced to calculate the duty cycle offset ΔD. This algorithm comprehensively considers the current temperature difference, historical cumulative temperature difference, and temperature difference change trend, enabling rapid response to temperature deviations and suppression of overshoot. Specifically, the "anti-saturation" design prevents the integral term from accumulating excessively over long-term deviations, thus preventing controller output distortion and improving system stability and long-term regulation performance. Compared to traditional proportional regulation or simple feedback mechanisms, this method possesses stronger dynamic tracking and anti-interference capabilities, making it particularly suitable for scenarios with frequent load fluctuations or rapid temperature rise rates. Furthermore, directional duty cycle allocation achieves directional thermal equilibrium. Depending on whether the upper or lower bridge arm has a higher temperature, ΔD is added to the other side or subtracted from the higher-temperature side; that is, if the upper bridge arm is hotter, its duty cycle is reduced, and the lower bridge arm's duty cycle is increased; conversely, the opposite is also true. This directional adjustment strategy achieves dynamic load transfer—"the hotter component works less, the colder component bears more"—directly affecting the device's conduction time and effectively altering its conduction loss distribution. Compared to methods that rely solely on heat dissipation or passive temperature equalization, this method intervenes at the source of power distribution, resulting in higher regulation efficiency and proactivity.

[0111] In practice, when adjusting the dead time, the dead time is adjusted according to the following calculation formula:

[0112] ;

[0113] In the formula, and These are the dead time before and after the adjustment, respectively. To adjust the coefficients. In specific implementation, It is 0.2.

[0114] This design dynamically adjusts the dead time according to the temperature difference: when a bridge arm has a higher temperature, its corresponding dead time is shortened, thereby increasing its on-time window, reducing the switching frequency or reducing turn-off losses, and indirectly achieving "compensatory" excitation of the devices on that side. Conversely, the dead time is extended to limit its turn-on opportunities. This directional adjustment mechanism achieves dynamic "load transfer" or "operational throttling" of the high-temperature side devices, effectively alleviating local heat accumulation.

[0115] S4. Reduce temperature difference by switching working modes and adjusting switching frequency;

[0116] When switching operating modes, the active switching roles of the upper and lower bridge arms are periodically interchanged; when adjusting the switching frequency, the switching frequency on the high-temperature side is reduced according to a preset method.

[0117] The main switching and freewheeling functions of the upper and lower bridge arms are periodically switched. For example, in motor drive, different bridge arms are used alternately to undertake the active switching task. Every N switching cycles (N=100~1000), the active switching roles of the upper and lower bridge arms are interchanged, and the PWM phase compensation voltage offset is adjusted to avoid output voltage jumps.

[0118] In practical implementation, a state machine is used to manage the switching timing during operating mode switching to avoid output voltage jumps. During operating mode switching, a state machine is introduced to precisely manage the switching timing, ensuring that the output voltage does not suddenly change or jump when the upper and lower bridge arms switch roles. This solves the voltage oscillation or current surge problems that may be caused by traditional switching methods, improving the system's electromagnetic compatibility and operational stability.

[0119] In practice, when adjusting the switching frequency, the switching frequency of the high-temperature side bridge arm is updated according to the following calculation formula. :

[0120] ;

[0121] in, This is a preset value. In actual implementation, .

[0122] For example, reducing the frequency from 10kHz to 8kHz can decrease switching losses. The frequency adjustment range is constrained by system efficiency and EMI.

[0123] This method directly links the switching frequency to the temperature difference in a non-linear manner—the greater the temperature difference, the more significant the frequency drop, creating an adaptive control mechanism that slows down the process as the temperature rises. Compared to traditional fixed-frequency reduction or step-wise adjustment, this method can more accurately match the temperature rise trend, achieving smoother and more efficient heat dissipation control, while avoiding harmonic disturbances or increased electromagnetic interference caused by sudden frequency changes.

[0124] This method first acquires the real-time temperatures of the upper and lower bridge arms and calculates the temperature difference ΔT and its rate of change d(ΔT) / dt. Based on two preset temperature difference thresholds T1 and T2, the system is divided into three operating modes: monitoring mode, mild adjustment mode, and aggressive adjustment mode. This graded response mechanism can distinguish the severity of temperature differences, avoiding excessive intervention when the temperature difference is slight, while rapidly initiating powerful regulation when the temperature difference increases significantly. Compared with traditional fixed thresholds or single protection strategies, it has stronger adaptability and robustness. In the mild adjustment mode, the loss distribution of the upper and lower bridge arms is optimized by simultaneously adjusting the PWM duty cycle and dead time. Specifically, the duty cycle adjustment keeps the total duty cycle constant, allowing only a single-sided deviation of no more than ±3%, effectively avoiding output voltage distortion; while the dead time adjustment ensures that it is not lower than the typical value, preventing shoot-through risk. This dual-parameter coordinated adjustment method breaks through the limitations of traditional methods that only adjust the duty cycle or only adjust the dead time, realizing fine allocation of switching losses without compromising the output waveform quality, significantly improving control flexibility and system stability.

[0125] When the temperature difference reaches a high level, an aggressive adjustment mode is activated, employing deeper control methods: periodically switching the active switching roles of the upper and lower bridge arms (i.e., operating mode switching), allowing devices that have been continuously subjected to high stress to "rest"; simultaneously, reducing the switching frequency of the high-temperature side to decrease its switching losses. This strategy reconstructs the heat load distribution at the device usage logic level, avoiding long-term unidirectional heat accumulation and overcoming the passive response methods of traditional approaches that rely solely on heat dissipation or derating, achieving active thermal balancing without sacrificing output capability. Furthermore, this method not only focuses on the current temperature difference magnitude but also introduces the rate of temperature difference change as a judgment criterion, enabling early intervention and regulation before the temperature difference reaches a dangerous level. This gives the system a certain degree of predictability and feedforward capability, significantly improving response speed. Compared to traditional solutions that rely on heat dissipation system linkage or delayed protection, it can suppress temperature rise trends earlier, making it particularly suitable for transient thermal events in high-frequency switching scenarios.

[0126] This method can achieve active, rapid, and accurate dynamic balancing of the temperature difference between the upper and lower bridge arms without affecting system performance.

[0127] Example 2

[0128] Unlike Embodiment 1, in S1 of this embodiment, after acquiring the temperatures of the upper and lower bridge arms, it also monitors whether the temperature of either the upper or lower bridge arm exceeds a warning threshold. If either bridge arm exceeds the threshold, derating protection is triggered, and the fault is recorded. This real-time monitoring mechanism allows for timely intervention before the device enters a hazardous operating area, preventing permanent damage or sudden failure due to overheating. Compared to traditional methods that rely solely on final shutdown protection, this solution provides a tiered safety guarantee of "warning-derating-protection," significantly improving the system's ability to cope with sudden conditions such as transient overload and abnormal heat dissipation.

[0129] In S1, the Vds voltage spike of either the upper or lower bridge arm is monitored to see if it exceeds a preset spike threshold. If either bridge arm exceeds this threshold, the PWM output of the entire bridge arm is shut down. Due to factors such as parasitic inductance, sudden load changes, or drive mismatch, power devices may generate significant Vds voltage spikes at the moment of turn-off. Although such transient overvoltages are short in duration, they can easily lead to avalanche breakdown or long-term insulation degradation of the device. This solution, by monitoring the Vds voltage in real time, can accurately capture these transient high-voltage events and react within microseconds, preventing the device from entering a dangerous operating area. Compared to traditional methods that rely on post-fault diagnosis or periodic maintenance, this method has stronger foresight and protective capabilities.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the technical solutions. Those skilled in the art should understand that any modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. A method for controlling the temperature difference between the upper and lower bridge arms of a dynamic equalization power module, characterized in that, Includes the following steps: S1. Obtain the temperatures of the upper and lower bridge arms; S2. Calculate the temperature difference between the upper and lower bridge arms. and the rate of change of temperature difference ;in, It equals the upper arm temperature minus the lower arm temperature; if If T1 < T1, then enter monitoring mode and only record data; if T1 ≤ T1 If <T2, then enter the gentle adjustment mode and switch to S3; if If T1 is greater than or equal to T2, then enter the aggressive adjustment mode and switch to S4; where T1 and T2 are the preset first temperature difference threshold and second temperature difference threshold, respectively. S3. Reduce temperature difference by adjusting PWM duty cycle and dead time; When adjusting the PWM duty cycle, the total duty cycle of the upper and lower bridge arms remains unchanged, and the single-sided offset does not exceed ±3%; when adjusting the dead time, the adjusted dead time is greater than or equal to the preset typical value. S4. Reduce temperature difference by switching working modes and adjusting switching frequency; When switching operating modes, the active switching roles of the upper and lower bridge arms are periodically interchanged; when adjusting the switching frequency, the switching frequency on the high-temperature side is reduced according to a preset method.

2. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 1, characterized in that: In S1, the temperatures of the upper and lower bridge arms are obtained through hybrid temperature sensing technology.

3. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 2, characterized in that: The process of obtaining the temperature of the upper arm / lower arm includes: First, obtain direct measurements of the bridge arm temperature. ; Then, calculate the indirect estimate of the bridge arm temperature: ; in, The substrate temperature; Thermal resistance to the junction shell; Total power loss; Finally, the directly measured values Compared with indirect estimates Input the Kalman filter and output the fused temperature value, which will be used as the temperature in subsequent steps.

4. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 3, characterized in that: Total power loss The formula for calculation is: ； In the formula, Phase current; For on-resistance; The switching frequency; , For switching energy.

5. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 1, characterized in that: In S3, the process of adjusting the PWM duty cycle includes: The duty cycle offset is calculated using an anti-saturation PID algorithm. , It consists of proportional terms, integral terms, and differential terms; ； in, , , These are proportional gain, integral gain, and derivative gain, respectively. If the temperature of the upper bridge arm is higher, the adjusted duty cycle is: ； In the formula, , These are the upper bridge boom before and after adjustment, respectively. , These are the lower bridge arm before and after adjustment, respectively. If the temperature of the lower bridge arm is higher, the adjusted duty cycle is: 。 6. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 5, characterized in that: In S3, when adjusting the dead time, the dead time is adjusted according to the following formula: ； In the formula, and These are the dead time before and after the adjustment, respectively. This is for adjusting the coefficient.

7. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 1, characterized in that: In S4, when switching operating modes, the switching timing is also managed through a state machine to avoid output voltage jumps.

8. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 1, characterized in that: In S4, when adjusting the switching frequency, the switching frequency of the high-temperature side bridge arm is updated according to the following calculation formula. : ； in, This is the default value.

9. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 1, characterized in that: In S1, after acquiring the temperatures of the upper and lower bridge arms, it also monitors whether the temperature of the upper or lower bridge arm exceeds the warning threshold. If either bridge arm exceeds the threshold, the derating protection is triggered and the fault is recorded.

10. The method for controlling the temperature difference between the upper and lower bridge arms of the dynamic equalization power module as described in claim 9, characterized in that: In S1, it also monitors whether the Vds voltage spike of the upper or lower bridge arm exceeds the preset spike threshold. If either bridge arm exceeds the threshold, the PWM output of the entire bridge arm is turned off.

Images