High-efficiency IGBT parallel topology optimization method
The IGBT parallel topology optimization method addresses uneven current distribution and thermal coupling by using real-time monitoring and adaptive control, enhancing stability and efficiency in IGBT systems.
Patent Information
- Application Number
- CN202411556732.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-04
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional IGBT parallel technology faces challenges in achieving dynamic current sharing, stability, and adapting to complex working conditions due to uneven current distribution and thermal coupling, leading to reduced reliability and service life.
A high-efficiency IGBT parallel topology optimization method involving real-time monitoring by sensors, adaptive control algorithms, and a distributed feedback network with deep learning models to adjust drive signals and thermal management for optimal current sharing and reduced interference.
Enhances stability and power capacity by optimizing current distribution, reducing thermal coupling and electromagnetic interference, and improving efficiency and reliability of IGBT systems.
Smart Images

Figure CN119070596A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of power electronics, and in particular to a high-efficiency IGBT parallel topology optimization method. Background Art
[0002] In power electronics and industrial control systems, IGBT, as a power electronic device, is widely used in frequency converters, electric vehicles, solar inverters and wind power converters to achieve power conversion and control. In order to improve the power capacity and reliability of the system, IGBT parallel technology has become an important means.
[0003] However, there are some challenges in traditional IGBT parallel technology. For example, when multiple IGBTs are used in parallel, the current distribution is often uneven due to the inconsistency of device parameters, which not only affects the stability of the system but also may shorten the service life of the device. In addition, under dynamic working conditions, such as changes in load or temperature, traditional parallel topologies have difficulty in quickly and accurately adjusting current distribution. Traditional control algorithms may not be able to adapt to complex working conditions, making it difficult to achieve real-time optimization and adjustment of IGBT parallel systems. Summary of the invention
[0004] 1. Technical issues to be resolved In view of the shortcomings of the prior art, the present invention provides a high-efficiency IGBT parallel topology optimization method, which solves the problem of how to achieve dynamic current sharing control in IGBT parallel connection, improve the stability and power capacity of IGBT during use, and reduce thermal coupling and electromagnetic interference.
[0005] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: A high-efficiency IGBT parallel topology optimization method, comprising: S1. Deploy sensors in each IGBT to monitor and generate comprehensive status data in real time; S2. Based on the comprehensive status data collected by the sensor, an adaptive control algorithm is used to generate a real-time dynamic drive signal, and the amplitude and duration of the drive signal are adjusted according to the status of each IGBT to achieve dynamic current sharing control of the IGBT under different working conditions; S3. Input the generated drive signal into the gate drive circuit of each IGBT, adjust the on-time and on-voltage of the IGBT, and then regulate its current output, so that the IGBT can automatically share the load current when working in parallel, ensuring the stability of dynamic current sharing; S4. Establish a distributed feedback network connecting all IGBTs, so that each IGBT can update its status information in real time after receiving its own regulation feedback signal, and transmit it to the distributed feedback network, so as to promote the coordinated optimization of the distributed feedback network and improve the current sharing effect; S5. Based on the status information in the distributed feedback network, the layout of the IGBT parallel connection is dynamically adjusted, and a multi-layer topology structure is constructed using new insulation materials and heat dissipation design to reduce thermal coupling and electromagnetic interference and enhance the overall stability of the IGBT parallel connection; S6. Combined with the data and status information collected by the distributed feedback network, the deep learning model is used to predict future load changes, and the switching frequency, duty cycle and drive signal parameters of the IGBT are dynamically adjusted according to the prediction results to improve the current sharing accuracy and the response speed of the distributed feedback network.
[0006] Preferably, a temperature sensor, a current sensor and a switch state sensor are integrated in each IGBT to monitor the operating parameters of the IGBT, including junction temperature, collector current and gate voltage; the analog signal collected by the sensor is converted into a digital signal through an analog-to-digital converter, and the converted digital signal is transmitted to the central processing unit, and the data is denoised to eliminate noise interference in the measurement, calibrated to ensure the accuracy of the data, and fused to integrate data from different sensors; after data processing, the central processing unit generates a comprehensive status data, and the comprehensive status data reflects the working status of the IGBT.
[0007] Preferably, after receiving the digital signal transmitted by the sensor, the central processing unit uses its built-in data processing unit to conduct in-depth analysis of these data to determine the real-time working status of each IGBT; through a comprehensive evaluation of the junction temperature, collector current and gate voltage, it has a comprehensive understanding of the working status of the IGBT; based on this analysis result, the central processing unit runs an adaptive control algorithm to dynamically adjust the parameters of the IGBT drive signal according to the real-time monitored working status data to optimize its performance and efficiency; one of the key calculations of the adaptive control algorithm is the drive power P drv , P drv It is calculated based on the gate charge Qg, the switching frequency fin and the actual output voltage swing ΔVg of the driver. The formula is P drv = Qg *fin * ΔVg; The gate charge Qg is obtained by integrating the gate charge curve in the region where the gate voltage is turned off and the gate voltage is turned on; The central processing unit calculates the gate charge according to the drive power P drv , adjust the amplitude and duration of the driving signal.
[0008] Preferably, the drive signal is transmitted to the gate of the IGBT, and its on-time and on-voltage are adjusted, thereby controlling the switching behavior of the IGBT; the working state is adjusted by monitoring the current output of each IGBT in real time and comparing it with a preset threshold; assuming that a parallel system contains four IGBT modules and the total current load of the system is 200 amperes, in order to achieve current sharing, a threshold of 50 amperes is preset for each IGBT module; if the current output of any IGBT exceeds this threshold, for example, reaching 55 amperes, the system will determine that the current output of the module exceeds the threshold; when it is monitored that the current output of a certain IGBT exceeds the set threshold, the IGBT's off time is extended, its on time is reduced, and the current output is reduced to prevent damage caused by overheating or excessive current; on the contrary, if the current output is lower than the threshold, the on-time of the IGBT is increased to increase its current output to ensure that it bears a reasonable current load in parallel operation.
[0009] Preferably, the distributed feedback network transmits data through a communication protocol and interface, and each IGBT is integrated with a communication interface so that the IGBT can receive instructions from the distributed feedback network and send its own comprehensive status data, including temperature, current and voltage; the comprehensive status data is transmitted to the central node of the distributed feedback network in real time through the communication interface; the central node analyzes through a deep learning model and generates adjustment instructions to optimize the working state and current sharing effect of the IGBT; the deep learning model predicts future load changes by analyzing historical data and real-time status information, thereby adjusting the switching frequency, duty cycle and drive signal parameters of the IGBT in advance to adapt to the upcoming load changes; the deep learning model collects historical and real-time data, including the temperature, current and voltage of the IGBT; uses these data to train the deep learning model so that it can recognize the pattern of load changes; the deep learning model uses these patterns to predict future load changes; the predicted load changes are used to dynamically adjust the operating parameters of the IGBT, including the switching frequency, duty cycle and drive signal parameters, to optimize the response and performance of the IGBT.
[0010] Preferably, the distributed feedback network collects the temperature, current, voltage and other status information of each IGBT in real time through communication protocols and interfaces; the collected data is transmitted to the central processing unit, and the data is analyzed by a deep learning model to evaluate the working status and thermal distribution of the IGBT; the deep learning model identifies hot spots and electrical stress concentration areas, conducts heat through high thermal conductivity insulating materials, and exports heat from the IGBT chip; the packaging structure of the IGBT is optimized, and the DBC technology provides a heat conduction path by directly bonding the copper foil to the ceramic substrate while maintaining the electrical insulation properties, further improving the thermal conductivity and insulation performance; in terms of multi-layer circuit board design, each layer has a specific function to achieve electrical isolation and thermal management; the bottom layer is used for power distribution, using wide copper wires and a symmetrical layout to balance and minimize resistance and voltage drop; the middle layer focuses on signal transmission, using shielding and high-frequency materials to reduce signal interference and improve signal integrity; the top layer and additional layers are used for thermal management, and heat is effectively conducted and dissipated by laying out thermal vias, heat sinks and heat pipes.
[0011] (III) Beneficial effects The present invention provides a high-efficiency IGBT parallel topology optimization method, which has the following beneficial effects: 1. The present invention optimizes the switching time and conduction voltage of the IGBT by real-time monitoring of the working state of the IGBT and dynamically adjusting the driving signal, thereby reducing the switching loss and conduction loss and improving the overall power conversion efficiency.
[0012] 2. By integrating sensors and adopting adaptive control algorithms, the present invention realizes precise current control of IGBT, avoids local overheating caused by uneven current, optimizes heat distribution, and thus enhances the thermal stability and reliability of the power circuit.
[0013] 3. The present invention dynamically adjusts the switching frequency and duty cycle of the IGBT through the coordinated optimization of the distributed feedback network and the deep learning model, thereby reducing electromagnetic interference, improving the overall electromagnetic compatibility, and meeting the strict electromagnetic compatibility requirements in high power density applications. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION
[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0016] Embodiment 1: like Figure 1 As shown, an embodiment of the present invention provides a high-efficiency IGBT parallel topology optimization method.
[0017] First, temperature sensors, current sensors, and switch status sensors are deployed in parallel on the IGBTs to monitor the working status of each IGBT in real time; the sensor data of each IGBT is converted into a digital signal through an analog-to-digital converter and sent to the central processing unit for processing; the central processing unit uses data denoising to eliminate noise and interference in the sensor signal to ensure the accuracy and reliability of the data; after denoising, the central processing unit will generate comprehensive status data, including parameters such as the junction temperature, collector current, and gate voltage of the IGBT.
[0018] The central processing unit uses an adaptive control algorithm to dynamically generate drive signals based on comprehensive status data. The adaptive control algorithm is implemented by P drv = Qg * fin * ΔVg to calculate the drive power, where Qg is the gate charge, fin is the switching frequency, and ΔVg is the output voltage amplitude of the driver. The gate charge Qg is obtained by integrating the gate charge curve in the gate voltage off and gate voltage on regions. Using the real-time monitored current and temperature data, the central processing unit continuously adjusts the amplitude and duration of the drive signal to optimize the performance of each IGBT. For example, when the junction temperature of a certain IGBT is detected to rise to the set threshold, the central processing unit will automatically extend the IGBT's off time and reduce its on time to prevent overheating.
[0019] After generating the drive signal, the central processing unit transmits it to the gate drive circuit of each IGBT to adjust the on-time and on-voltage of the IGBT; at this time, each IGBT adjusts its current output according to the received drive signal, so that the IGBT can automatically share the load current when working in parallel; to ensure the current sharing effect, a current threshold is set for each IGBT. For example, assuming that an IGBT shares a total current load of 200 amperes in parallel, it contains four IGBT modules, and the preset threshold of each module is 50 amperes. If the current output of an IGBT exceeds this threshold, for example, reaching 55 amperes, the central processing unit will adjust its current output by extending its off-time and reducing its on-time to avoid damage due to overload.
[0020] Secondly, a distributed feedback network is established, which connects all IGBTs and transmits data through a communication interface. Each IGBT has an integrated communication interface that can send its comprehensive status data to the central node of the distributed feedback network in real time. The central node is responsible for collecting data from each IGBT, analyzing its working status, and generating adjustment instructions. This process is optimized through a deep learning model, which predicts future load changes based on historical data and real-time status information. When an increase in load is predicted, the IGBT parallel connection will adjust the switching frequency, duty cycle, and drive signal parameters of the IGBT in advance to adapt to the upcoming load changes, so that its IGBT is always in good working condition.
[0021] The deep learning model analyzes data to evaluate the operating status and thermal distribution of the IGBT; when the deep learning model identifies electrical stress concentration or excessive temperature in a certain area, the distributed feedback network automatically adjusts the operating parameters of the relevant IGBT to avoid performance degradation caused by local overheating.
[0022] For thermal management, the invention uses new insulating materials and heat dissipation design to construct a multi-layer topology structure to reduce thermal coupling and electromagnetic interference, thereby improving heat dissipation capabilities. The new insulating material is Taoxi silicone gel or ceramic substrate aluminum nitride; the bottom layer of the IGBT parallel connection adopts a wide copper wire layout to balance the current and minimize resistance and voltage drop; the middle layer focuses on signal transmission, using high-frequency materials and shielding technology to reduce signal interference; the top layer ensures effective heat conduction and heat dissipation by arranging thermal vias, heat sinks and heat pipes; the selection of materials and layout design need to comprehensively consider thermal conductivity and insulation properties to achieve better thermal management effects.
[0023] To verify the effectiveness of the IGBT parallel connection, different load conditions need to be simulated during the experimental phase and the working data of each IGBT needs to be collected. When the load increases, the central processing unit needs to adjust the drive signals of each IGBT in time to observe the current sharing effect. For example, when the load suddenly increases to 250 amperes, the IGBT parallel connection should be able to respond quickly and automatically distribute the load evenly to the four IGBTs to ensure that the current output of each IGBT is maintained within a reasonable threshold range. At the same time, the safe operation of the IGBT parallel connection is ensured by real-time monitoring of the junction temperature changes of each IGBT.
[0024] Through the intelligent adjustment of distributed feedback networks and deep learning models, the working efficiency and stability of IGBT parallel connection are improved; by optimizing driving signals, thermal management and layout design, this invention can not only achieve efficient IGBT parallel control, but also provide a new solution for the intelligent development of power electronic equipment.
[0025] Embodiment 2: For example, the present invention is applied to the inverter system of an electric vehicle to improve the current sharing capability of the parallel IGBTs and the overall efficiency of the system, thereby improving the dynamic response performance and operating stability of the vehicle; assuming that the inverter system of the electric vehicle needs to drive the motor to operate under different working conditions, including acceleration, deceleration, and constant speed cruising, the changes in these states will cause frequent fluctuations in the system load, resulting in uneven current distribution in the parallel IGBTs; the traditional static current sharing method is difficult to ensure consistent current distribution for each IGBT under these dynamic changes, which may cause some IGBTs to overheat or overload, ultimately reducing the reliability and service life of the system.
[0026] Sensors are deployed to monitor the current and temperature status of each IGBT in parallel in real time, and the collected data is transmitted to the central processing unit for processing; taking the acceleration state of an electric vehicle as an example, when the driver suddenly accelerates, the system load increases rapidly, causing the inverter output current to rise sharply; at this time, the real-time monitoring system will detect the current output deviation of each IGBT, and the current of some IGBTs may exceed the preset threshold, resulting in a momentary overload phenomenon; based on the adaptive control algorithm, the central processing unit immediately dynamically adjusts the drive signals of these IGBTs, such as shortening their on-time or increasing their off-time, thereby quickly reducing their current output and redistributing the current to other IGBTs that are not overloaded; this active current sharing control can be completed within milliseconds to ensure that the current distribution of each IGBT is balanced, thereby preventing some IGBTs from overheating due to overload.
[0027] When the vehicle decelerates or brakes to recover energy, the load on the IGBT suddenly decreases. At this time, the system needs to readjust the working state of each IGBT to avoid the situation where the current of some IGBTs is too low and the efficiency is reduced. The distributed feedback network of the present invention transmits the IGBT status information to the deep learning model through real-time communication with each IGBT. The deep learning model predicts the load change at the next moment based on historical operating data and current status, and adjusts the switching frequency and duty cycle of the IGBT in advance to adapt it to the new working condition. This process enables the IGBT parallel system to maintain a high efficiency and stable working state under the working condition of rapid load changes, thereby improving the power response speed and system safety of electric vehicles. In addition, under long-term high-load conditions, IGBTs are prone to generate a large amount of heat. If the heat is not dispersed and managed in time, it will cause local overheating, thereby damaging the performance of the IGBT. The multi-layer topology design and new insulating materials effectively reduce the thermal coupling phenomenon between IGBTs. The multi-layer layout of heat sinks, heat pipes and thermal vias can quickly export and dissipate the heat generated by the IGBT into the surrounding environment while maintaining the electrical insulation characteristics between the IGBTs. The thermal management capability of the IGBT is enhanced, allowing the system to operate continuously and stably in a high-temperature environment, avoiding performance degradation due to excessive temperature.
[0028] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A high-efficiency IGBT parallel topology optimization method, characterized in that: include: S1. Deploy sensors in each IGBT to monitor and generate comprehensive status data in real time; S2. Based on the comprehensive status data collected by the sensor, an adaptive control algorithm is used to generate a real-time dynamic drive signal, and the amplitude and duration of the drive signal are adjusted according to the status of each IGBT to achieve dynamic current sharing control of the IGBT under different working conditions; S3. Input the generated drive signal into the gate drive circuit of each IGBT, adjust the on-time and on-voltage of the IGBT, and then regulate its current output, so that the IGBT can automatically share the load current when working in parallel, ensuring the stability of dynamic current sharing; S4. Establish a distributed feedback network connecting all IGBTs, so that each IGBT can update its status information in real time after receiving its own regulation feedback signal, and transmit it to the distributed feedback network, so as to promote the coordinated optimization of the distributed feedback network and improve the current sharing effect; S5. Based on the status information in the distributed feedback network, the layout of the IGBT parallel connection is dynamically adjusted, and a multi-layer topology structure is constructed using new insulation materials and heat dissipation design to reduce thermal coupling and electromagnetic interference and enhance the overall stability of the IGBT parallel connection; S6. Combined with the data and status information collected by the distributed feedback network, the deep learning model is used to predict future load changes, and the switching frequency, duty cycle and drive signal parameters of the IGBT are dynamically adjusted according to the prediction results to improve the current sharing accuracy and the response speed of the distributed feedback network.
2. A high-efficiency IGBT parallel topology optimization method according to claim 1, characterized in that: Sensors are deployed in each IGBT, including temperature, current and switch status sensors; the sensors collect data about the working status of the IGBT in real time, including junction temperature, collector current and gate voltage; the collected data is converted into digital signals through the analog-to-digital converter ADC and transmitted to the central processing unit, and the collected data is denoised, calibrated and fused through data processing to generate a comprehensive status data.
3. A high-efficiency IGBT parallel topology optimization method according to claim 2, characterized in that: The central processing unit analyzes the comprehensive status data through an advanced data processing unit to determine the current working state of each IGBT; the central processing unit runs an adaptive control algorithm, which calculates the parameters of the driving signal required for each IGBT according to the analysis results of the data processing unit, including the amplitude and duration of the driving signal, so that the IGBT can be dynamically controlled under different working conditions; the central processing unit calculates the driving power P of each IGBT drv , where the driving power P drv It is determined by the gate charge Qg, the switching frequency fin and the actual output voltage swing ΔVg of the driver. The calculation formula is P drv = Qg * fin * ΔVg; The central processing unit is based on the driving power P drv The calculation results are used to optimize the amplitude and duration of the driving signal; The gate charge Qg is obtained by integrating the gate charge curve in the region where the gate voltage is turned off and the gate voltage is turned on.
4. The high-efficiency IGBT parallel topology optimization method according to claim 1, characterized in that: The generated drive signal is input into the gate drive circuit of each IGBT and transmitted to the gate of each IGBT to adjust the on-time and on-voltage of the IGBT. When the current output of an IGBT exceeds the threshold, the gate drive circuit will extend the off-time of the IGBT and reduce its on-time, thereby reducing its current output; On the contrary, the on-time is increased to improve its current output so that when it works in parallel, each IGBT bears an equal current load.
5. The high-efficiency IGBT parallel topology optimization method according to claim 1, characterized in that: The method establishes a distributed feedback network connecting all IGBTs, and the distributed feedback network enables real-time data transmission through communication protocols and interfaces; each IGBT is equipped with a communication interface for receiving instructions from the distributed feedback network and sending its own status information; the status information includes temperature, current and voltage, which are used to evaluate the working status of the IGBT and perform current sharing control; the comprehensive status data is transmitted to the distributed feedback network in real time through the communication interface, and the central node of the distributed feedback network analyzes the data, generates adjustment instructions based on the internal deep learning model, and optimizes the working status and current sharing effect of the IGBT; the collaborative optimization dynamically adjusts the operating parameters of the entire IGBT parallel connection according to the status information of all IGBTs through the distributed feedback network; the deep learning model improves the prediction and response capabilities of the distributed feedback network. By analyzing historical data and real-time status information, the deep learning model can predict future load changes, thereby adjusting the switching frequency, duty cycle and drive signal parameters of the IGBT in advance to adapt to the upcoming load changes.
6. The high-efficiency IGBT parallel topology optimization method according to claim 1, characterized in that: Based on the state information in the distributed feedback network, the working state and thermal distribution of the IGBT are evaluated; by analyzing these data, hot spots and areas of electrical stress concentration are identified; Use high thermal conductivity insulating materials to provide good electrical insulation performance and conduct heat to improve heat dissipation efficiency; optimize the packaging structure of the IGBT and use DBC technology to further improve thermal conductivity and insulation performance; manage electrical isolation and thermal management between IGBTs through a multi-layer topology structure, and adopt a multi-layer circuit board design, which includes power distribution, signal transmission or thermal management; reduce electromagnetic interference by optimizing the circuit board layout and component arrangement.
7. The high-efficiency IGBT parallel topology optimization method according to claim 1, characterized in that: The data and status information collected by the distributed feedback network, including current, voltage, and temperature, are combined, and the deep learning model is trained to identify the complex relationship between these data and load changes, so as to predict the dynamic changes of future loads; based on the prediction results, the IGBT parallel connection can be adjusted in advance, and the switching frequency, duty cycle, and drive signal parameters of the IGBT can be dynamically adjusted through an adaptive control algorithm; the deep learning model can identify factors that affect the IGBT parallel connection, including thermal coupling and electromagnetic interference. By optimizing these factors, thermal coupling and electromagnetic interference can be reduced, and the response speed and current sharing accuracy of the distributed feedback network can be improved.
Citation Information
Patent Citations
Current sharing method of IGBT parallel circuit and IGBT parallel circuit
CN112003460A
Dynamic current sharing control method and system for parallel power switch devices
CN115765405A
Flyback switching power supply frequency raising control method and device
CN117674544A
LED large screen monitoring method based on distributed sensor network
CN118574091A
A laminated busbar used for connecting a DC capacitor busbar with IGBT power modules connected in parallel
CN203482103U
Cited By
Power semiconductor device aging on-line diagnosis method
CN120009688A
