功率模块在线工作温热失效的预测模型研究

发布时间：2018-03-29 13:44

本文选题：IGBT功率模块　切入点：实时温度监测　出处：《河北工业大学》2015年硕士论文

【摘要】：解决清洁能源的开发与利用是我国现今面临的综合性战略问题,涉及到了经济、环境、能源等诸多领域,然而这一战略的落实和规模发展受到新能源发电系统可靠性低的制约。随着功率器件在新能源发电中的广泛应用,由于风、光、潮汐等的不稳定性使发电功率波动十分剧烈,从而致使功率模块受到持续不断的热力冲击,加速了功率模块的失效进程,成为影响新能源发电系统可靠性的关键因素之一,因此,寻找并构建能准确预测功率模块剩余寿命的模型是确保新能源发电系统稳定持续运行的基础。目前相关功率模块在由于热冲击所引起的衰退性及器件可靠性研究较少,对功率模块失效进程了解模糊,不能准确理解与把握相关模块的失效特征方式;更多的寿命预测的研究局限于可能损坏的预警,不能在实际工况下做出准确预测。因此,围绕风力、光伏等发电中常用功率模块IGBT的在线寿命预测,论文的主要工作如下:(1)研究并总结了引发IGBT失效的因素。主要可分为三种:由于内部结构缺陷和制造工艺引发的潜在失效;外部应力对其造成内部疲劳损伤积累失效;电路其他外部运行环境对其造成的失效。在此基础上,阐述了不断累积的热冲击使IGBT功率模块内部不同材料结构间逐渐发生变化并导致失效的过程。(2)提出并构建了基于实时监测IGBT功率模块温度的剩余寿命预测模型。深入分析现有的IGBT功率模块温度的模拟及探测方法,发现其只能向控制系统发出警告,并不能判断模块是否真正损坏,更不能由此推测模块的剩余寿命。针对上述问题,论文提出了一种新的预测方法,该方法基于实时监测IGBT功率模块温度,将模块总寿命和即时温度量化,通过累积运行过程中模块的损伤来实现实时监测IGBT功率模块寿命的变化,进而预测模块剩余寿命。(3)搭建模拟实际工况的测温闭环系统为寿命预测模型提供真实可靠数据,并运用云模型解决IGBT功率模块寿命的随机性与模糊性。从实际运行工况着手,采用CM100DY-24NF型IGBT功率模块,搭建测温闭环系统,控制IGBT功率模块的进行可靠性老化试验,实验模拟了模块几种工作状态下温度的动态波形,探究了两种温度变化下模块的不同工作条件下的可能原因,并提出将云模型引入可靠性分析中,与传统可靠性相比,云模型是定性定量相互转换的工具,把具有随机性和模糊性特征的性能参数序列的分布特征用参数可变的分布函数定量表达,从理论上提高了参数估计结果的合理性。论文在引起IGBT功率模块失效机理上进行了探究,并构建了基于实时监测温度的寿命预测模型,为新能源发电装置中关键性器件的在线寿命预测打下了一定基础。
[Abstract]:The solution to the development and utilization of clean energy is a comprehensive strategic problem facing our country nowadays, which involves many fields, such as economy, environment, energy and so on. However, the implementation and scale development of this strategy is restricted by the low reliability of the new energy generation system. With the wide application of power devices in the new energy generation, due to the instability of wind, light, tide and so on, the power fluctuation of power generation is very severe. As a result, the power module is subjected to continuous thermal shock, which accelerates the failure process of the power module and becomes one of the key factors affecting the reliability of the new energy generation system. To find and build a model that can accurately predict the residual life of power modules is the basis to ensure the stable and continuous operation of new energy generation systems. At present, there are few researches on the degradation and device reliability of power modules due to thermal shock. Understanding of the failure process of power module is fuzzy, and the failure characteristics of relevant modules can not be understood and grasped accurately. More research on life prediction is limited to the early warning of possible damage and can not be accurately predicted under actual working conditions. The online life prediction of IGBT, a power module commonly used in wind power generation and photovoltaic generation, The main work of this paper is as follows: (1) study and summarize the factors that cause IGBT failure. It can be divided into three types: internal structure defect and potential failure caused by manufacturing process, internal fatigue damage accumulation failure caused by external stress, and internal fatigue damage accumulation failure caused by external stress. Failure caused by other external operating environments of the circuit. On this basis, In this paper, the cumulative thermal shock causes the gradual change of different materials and structures in the IGBT power module and leads to the failure process. (2) the residual life prediction model based on real-time monitoring the temperature of the IGBT power module is proposed and constructed. Deep analysis of the existing IGBT power module temperature simulation and detection methods, It is found that it can only give warning to the control system and can not judge whether the module is really damaged or not, let alone predict the remaining life of the module. In view of the above problems, a new prediction method is proposed in this paper. This method is based on real-time monitoring the temperature of IGBT power module, quantifies the total life and instant temperature of the module, and realizes the real-time monitoring of the change of the life of the IGBT power module by accumulating the damage of the module in the running process. Furthermore, the residual life of the prediction module is predicted. (3) the closed-loop system of temperature measurement is set up to simulate the actual working conditions. It provides the true and reliable data for the life prediction model, and solves the randomness and fuzziness of the life of the IGBT power module by using the cloud model. The closed loop system of temperature measurement is built by using CM100DY-24NF type IGBT power module, and the reliability aging test of IGBT power module is controlled. The dynamic waveform of temperature in several working states of the module is simulated experimentally. This paper probes into the possible reasons for the different working conditions of the modules under two kinds of temperature changes, and proposes to introduce the cloud model into the reliability analysis. Compared with the traditional reliability, the cloud model is a tool for qualitative and quantitative interconversion. The distribution characteristics of performance parameter sequences with randomness and fuzziness are expressed quantitatively by the variable parameter distribution function. The rationality of the parameter estimation results is improved theoretically. The mechanism of IGBT power module failure is explored, and the life prediction model based on real-time temperature monitoring is constructed. It lays a foundation for on-line life prediction of key devices in new energy generation plants.
【学位授予单位】：河北工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN322.8

【参考文献】