多能源储能系统中三相逆变器故障诊断方法与参数辨识的研究

发布时间：2018-05-14 09:03

本文选题：多能源储能系统 + 三相逆变器　；参考：《浙江大学》2016年博士论文

【摘要】："十三五"规划首次将储能以专项规划形式纳入国家级规划,电力电子技术是利用电力电子器件对电能进行高效控制和转换的新兴学科,是储能环节和用户端的接口技术。面向长备用、安全可靠、高效率的防灾供电需求和风电、光电等新能源需求,多能源储能系统需要具备高的可靠性,并且具有故障诊断和故障预测的能力,持续不间断地工作以减少因故障停机而带来的经济损失。故障诊断技术和参数辨识技术作为提高多能源储能系统可靠性的有效途径,已经成为该研究领域的热点和重点。研究表明,在多能源储能系统中,功率变换器尤其是逆变器中的半导体器件和无源器件是容易发生故障的薄弱环节。因此,对多能源储能系统中三相逆变器的故障诊断技术和参数辨识技术方面的深入研究具有重要的理论意义和工程应用价值。本文具体研究内容包括以下四个方面:(1)提出了适用于故障诊断的三相逆变器电路的电压解析模型,并提出了基于桥臂中点间电压和基于桥臂中点电压的两种故障诊断方法,提高了诊断速度、诊断准确性,适用于各种负载情况。基于电压量的故障诊断方法能快速地诊断故障,不过需要额外传感器和硬件电路的支持,但是面向多能源储能系统的应用背景,故障诊断速度和冗余切换快速性是至关重要的因素。因此本文详细分析了三相逆变器电路每一种开关管开路故障情况下电路特征电压的变化,提出了适用于故障诊断的三相逆变器电路的电压解析模型。基于电压解析模型,提出了两种电压型的故障诊断方法,即基于桥臂中点间电压的故障诊断方法和基于桥臂中点电压的故障诊断方法。两种故障诊断方法都具有较强的抗干扰性和鲁棒性,基于桥臂中点电压的故障诊断方法的诊断时间比基于桥臂中点间电压的故障诊断方法的少,最大诊断时间为1/2个基波周期。所以基于桥臂中点电压的故障诊断方法是相对较优的一种故障诊断方法。(2)提出了基于灵活参考电压的电压残差故障诊断方法。从多能源储能系统成本角度考虑,基于电流型的故障诊断方法优于基于电压型的故障诊断方法。本文利用闭环控制系统已有的电压采样量作为故障诊断的电气状态量,结合残差理论提出了基于灵活参考电压的三相逆变器故障诊断方法,方便以子程序的形式植入到已有的闭环控制算法中。通过统计的方法给出了残差评价的决策函数和决策规则,在故障诊断速度和诊断准确性两个性能指标之间折中考虑选择比较阈值,在确保故障诊断准确性的前提下,提高了故障诊断的速度。(3)研究了多能源储能系统的基本结构和多功率模块并联系统故障诊断的需求,首次将蚁群算法思想应用到多能源储能系统故障诊断领域,提出了基于蚁群算法的多功率模块并联系统故障诊断方法。多能源储能系统中存在多个功率模块并联的情况,包括直流-直流变换多功率模块并联系统和直流-交流变换多功率模块并联系统。系统级的故障诊断技术相对于电路级的故障诊断技术,不需要诊断对象的精确模型,并且能够在不增加诊断对象额外成本的情况下,快速有效得进行故障诊断。对多功率模块并联系统进行故障诊断建模,提出了一种简单有效的环形比较图结构模型。然后将蚁群算法思想引申到环形比较图结构模型中,构建了适用于多功率模块并联系统故障诊断的蚂蚁系统。在此基础上,提出了基于蚁群算法的多功率模块并联系统故障诊断方法,实现了系统级的故障诊断。(4)研究了一类大功率应用场合下毫亨级的磁粉芯材料功率电感,建立了电感的模型并提出了电感的失效标准,通过最小二乘法实现了电感参数的在线辨识,实现了多能源储能系统的故障预测。实际电感可以等效为理想电感和等效串联电阻的串联,选择电感等效电路模型中的电感值作为电感的故障特征参数,将电感值偏小20%作为判定电感失效的标准。然后在建立多能源储能系统三相逆变器电路模型的基础上,探讨了一种能有效应用于参数性故障诊断的辨识方法。同时,针对三相逆变器电路中电感参数随负载电流偏置变化的情况,提出了一种改进的适用于变电感情况的参数辨识算法,实现了电感参数的在线辨识。
[Abstract]:In 13th Five-Year, the plan of "13th Five-Year" is integrated into national planning for the first time in the form of special planning. Power electronic technology is a new subject using power electronic devices to effectively control and convert electrical energy. It is the energy storage link and the interface technology of the user side. The demand for long reserve, safety and reliability, high efficiency, power supply and electricity, photoelectric and other new energy Source demand, the multi energy energy storage system needs high reliability, and has the ability of fault diagnosis and fault prediction, continuous and continuous work to reduce the economic loss caused by failure. As an effective way to improve the reliability of multi energy energy storage system, fault diagnosis technology and parameter identification technology have become the research. The research shows that in the multi energy energy storage system, the power converter, especially the semiconductor device and the passive device in the inverter, is the weak link of the fault easily. Therefore, it is important to study the fault diagnosis technology and the parameter identification technology of the three-phase inverter in the multi energy energy storage system. The theoretical significance and engineering application value. The specific research contents of this paper include the following four aspects: (1) the voltage analytic model of the three-phase inverter circuit suitable for fault diagnosis is proposed, and two fault diagnosis methods based on the middle point voltage of the bridge arm and the middle point voltage of the bridge arm are proposed, which improves the diagnosis speed, the diagnostic accuracy and the suitability. It is used for various loads. Fault diagnosis based on voltage quantity can quickly diagnose faults, but it needs additional sensors and hardware circuits, but the application background of multi energy energy storage system, the speed of fault diagnosis and the speediness of redundant handoff are the most important factors. So this paper analyzes the three-phase inverter power in detail. A voltage analytic model for a three-phase inverter circuit suitable for fault diagnosis is proposed in the case of open circuit fault of each switch tube. Based on the voltage analytic model, two voltage type fault diagnosis methods are proposed, that is, the fault diagnosis method based on the middle point voltage of the bridge arm and the middle point voltage based on the bridge arm. Fault diagnosis method. The two methods of fault diagnosis have strong anti-interference and robustness. The diagnosis method based on the middle point voltage of the bridge arm is less than the fault diagnosis method based on the midpoint voltage of the bridge arm, and the maximum diagnosis time is 1/2 basic wave period. So the fault diagnosis method based on the middle point voltage of the bridge arm is the fault diagnosis method A relatively superior fault diagnosis method. (2) a fault diagnosis method based on the flexible reference voltage is proposed. Considering the cost of the energy storage system, the current based fault diagnosis method is superior to the voltage based fault diagnosis method. This paper uses the voltage sampling amount of the closed loop control system as a fault. The fault diagnosis method of three-phase inverter based on flexible reference voltage is proposed based on the residual error theory, which is conveniently embedded in the existing closed loop control algorithm in the form of subroutine. By statistical method, the decision function and decision rules of the residual evaluation are given, and the fault diagnosis speed and the diagnostic accuracy are two. In order to ensure the accuracy of fault diagnosis, the speed of fault diagnosis is improved on the premise of ensuring the accuracy of fault diagnosis. (3) the basic structure of multi energy energy storage system and the requirement of fault diagnosis for multi power module parallel system are studied, and the idea of ant colony algorithm is first applied to the fault diagnosis field of multi energy energy storage system for the first time. The fault diagnosis method of multi power module parallel system based on ant colony algorithm is proposed. There are several power modules parallel in the multi energy energy storage system, including DC DC converter multi power module parallel system and DC AC transform multi power module parallel system. System fault diagnosis technology is relative to circuit level fault. The diagnosis technology does not need accurate model of the diagnosis object, and can diagnose the fault quickly and effectively without increasing the extra cost of the diagnosis object. A simple and effective structure model of the ring comparison graph is proposed for the multi power module parallel system. Then the idea of ant colony algorithm is extended to the ring. On the basis of the ant colony algorithm, a fault diagnosis method for multi power module parallel system is proposed, and the system level fault diagnosis is realized. (4) a kind of milli grade magnetic powder core work in a class of high-power applications is studied. The inductor model is established and the inductor failure standard is put forward. The on-line identification of the inductance parameters is realized by the least square method. The fault prediction of the multi energy energy storage system is realized. The actual inductance can be equivalent to the series of the ideal inductor and the equivalent series resistance, and the inductance value of the selective inductance equivalent circuit model is used as the inductance. The characteristic parameters of the fault are as small as 20% of the inductance value as the criterion for determining the failure of the inductor. Then, on the basis of the establishment of a three-phase inverter circuit model of the multi energy energy storage system, an identification method which can be used for the parameter fault diagnosis is discussed. At the same time, the inductance parameters in the three-phase inverter circuit vary with the load current bias. An improved parameter identification algorithm suitable for variable inductance is proposed to realize on-line identification of inductance parameters.

【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TM464

【参考文献】