Study On Junction Temperature Smooth Control Of IGBT Module In Power Converter Under Power Fluctuation Condition

As a key component of high-power electronic systems, power converter plays an important role in safe and reliable operation of the converter system. Especially in renewable energy generation, converters often operate in harsh conditions with large-scale fluctuations of processing power, which leads to lower reliability and shorter sevice life compared with traditional power equipment. The reason is that there is great difference of thermal expansion coefficient among each layer of the power module, and there exists long-time random fluctuation of the processing power. These two factors result in dramatic variation of chip temperature and alternating thermal shock inside the module, which induces fatigue and aging problem as well as system fault. Meanwhile, with the development of technology, improving the reliability of power converter with new package material, new design and new manufacture processhas been restricted by cost and inner space of the module; what’s more, the faults can hardly be avoided just from the perspective of design and manufacture. Therefore, in order to improving the reliability of the power converter, the technology of delaying the fatigue and aging of the module has been studied in this thesis through suppressing and smoothing the thermal shock inside the power module from the systematic operation level.Firstly, the influence factors and the estimation methods of junction temperature have been studied. The research shows that the switching frequency, the amplitude and phase of the device voltage and current have an impact on the junction temperature; in addition, the switching trace of the device can also be used as a thermal control variable. Considering the accuracy and the calculation speed are hard to balance in the process of power loss calculation while using the traditional electro-thermal model, the average loss model over a line frequency period has been derived under the space vector modulation. Based on the model, the power loss distribution and thermal characteristic of the power module in a direct-drive wind power converter have been studied, under the condition of wind speed fluctuation. The junction temperature is the basis of thermal management. As the junction temperature of the module cannotbe measured directly,this dissertation proposes the method of artificial neural network to solve this problem. It is shown that this method has a relative high precision.Secondly, most of the existing junction temperatureadjustment through switching frequency regulation mainly focused on over-temperature protection and average temperature control. The thermal smooth control strategy based on switching frequency regulation is proposed. The adjustable range of the switching frequency has also been discussed. Then, the internal thermal control scheme has experimentally verified based on a self-designed inverter prototype. At last, a direct-drive wind power converteris taken as an example for thermal smooth control. By using the linear cumulative damage theory and rain flow counting method, the quantitative analysis of the control effect for the thermal management strategy is presented.Thirdly, as the regulation range of the switching frequency is limited in the process of thermal smooth control, a novel idea of thermal smooth control based on switching trace regulation has been proposed for the first time. In practice, the snubber circuitis often used as a supplementary for the power device, which has an effect for the switching loss of the device. Inspired by the idea, this dissertation proposes a new implementation method for junction temperature control based on switching trajectory regulation of power switch, by adding auxiliary switch in the traditional snubber circuit. Studies show that this method can smooth the junction temperature fluctuations effectively, reducing the impact of thermal stress. Compared with switching frequency adjustment based thermal control strategy, this method does not affect the quality of the output waveform of the converter.Lastly, because the control effect is limited relying on a single thermal management method,a multi-parameter combined thermal smooth control scheme has been put forward to remedy the defect. A thermal management power converter system through digital implementation has been designed. FPGA board is used for the whole system’s digital control. The embedded programming design idea is applied for both the power control and thermal management, so the system upgrade and maintenance is very convenient. The system not only achieves the goal of stable operation, meeting the requirements of power handling, but also reduces the thermal stress inside the power module via multi-parameter combined thermal smooth control. The junction temperature control effect has been verified by direct measurement of the chip temperature of the unpackaged modules with the help of infrared radiation thermometer. Considering that the thermal resistance from the module case to the surrounding environment can be dynamically adjusted, the author proposesan idea of active control strategy for heatsink thermal resistance adjustment through fan speed regulation. The hardware implementation shows that the external thermal control has slower response speed compared with internal management, but it can regulate the average temperature. The combination of both methods is an effective strategy for thermal smooth control.Overall, compared with other control strategies for smoothing thermal shock reported before, the proposed thermal management ideas and implementation methods have several advantages. Firstly, it does not affect the processing power of the converter system; secondly, it is more effective for smoothing junction temperature fluctuations as well as reducing the fatigue and aging rate; thirdly, it does not require extra energy storage system.