车载大功率电力电子变压器鲁棒容错控制

Complexity and uncertainty grow dramatically as engineering systems getting more and more sophisticated. Model uncertainties, nonlinearities and strong coupling among multiple variables are significant characteristics of such systems. Fault diagnosis in closed-loop systems are facing challenges caused by broad propagation and complex evolution of malfunctions. A much trickier fact is that faults can be hidden in closed-loop systems. It is thus difficult to achieve effective fault diagnosis and fault-tolerant control using optimization of presentative performance of systems, while the utilized technologies are based on Euclidian distance. Power-electronic transformer is a promising approach to achieve light weight and high efficiency of high-speed trains. However, high-capacity vehicular power-electronic transformers are fragile with faults and abnormal working conditions. This can lead to severe casualty and huge loss. The proposed project will tackle these challenges in terms of fault diagnosis and robust fault-tolerant control. Methods that address intrinsic performance optimization of closed-loop systems will be established. v-Gap Metric shall be used as a primary tool for fault detection and classification as well as system behavior analysis, thanks its inherent capability of measuring closed-loop performance, identification and fault diagnosis. We shall propose innovative methods for modelling and identification, fault diagnosis as well as fault-tolerant control for high-capacity vehicular power-electronic transformers. Major scientific problems to be solved in this project include: mechanism and propagation channels of vehicle-network coupling electric oscillation, effective description of fault impact on closed-loop systems and design methods for fault-tolerant control addressing performance optimization. Through this project, we expect to bring forth innovations for the theory and methods of fault diagnosis and fault-tolerant control.

工程系统复杂性和不确定性快速增加，模型不确定性、非线性、多变量强耦合等显著呈现，闭环运行情况下控制系统故障诊断，不仅面临故障广泛传播与复杂演化以及故障模式快速增长等挑战性问题，更重要的是，现有基于欧氏距离的表象性能优化难以解决有效故障诊断和容错控制难题。为此，本项目拟以车载大功率电力电子变压器这个结构复杂、部件数量庞大、故障多发且后果严重的复杂系统为对象，基于v-间隙度量建立一套体现闭环系统本质性能优化的故障诊断与容错控制新方法，包括车载大功率电力电子变压器系统建模和辨识理论与方法，车载电力电子变压器系统故障诊断方法，车载电力电子变压器系统容错控制设计和面向对象的实验与应用研究等。重点解决的关键科学问题包括：车-网耦合系统电气振荡产生机理和传播机制、故障对闭环系统影响的有效定量描述、以及面向v-间隙度量性能优化的容错控制设计。取得一批原创性成果，建立一套有别于现有故障诊断与容错控制的新框架。

用车载大功率电力电子变压器取代传统牵引变压器，可以使牵引传动系统的重量和体积大大降低，是实现高速列车轻量化的必然技术选择。车载大功率电力电子变压器结构复杂，部件数量庞大，故障多发、后果严重，对其故障诊断和容错能力提出了极高的要求。列车运行往往要求在故障条件下，保持对强干扰的鲁棒性，并发挥冗余容量，保证一定的牵引力和性能，需要研究同时满足这些要求的新型控制理论和方法。.项目首次将整流器电压外环控制器与车-网耦合系统解耦，构建了更具一般性的车-网耦合动态模型，为稳定性分析和车-网耦合低频振荡抑制打下了基础；提出了车-网耦合稳定性判据，可以解析计算鲁棒稳定裕度，克服了既有稳定性判据只能给出数值解、且计算复杂的困难；提出了轻阻尼系统鲁棒稳定裕度优化方法，解决了当nu-间隙度量接近1时控制设计困难的问题；提出了利用H无穷回路成型设计方法和基于前馈电压补偿方法来抑制车-网耦合系统低频振荡。.面向车载整流器电感参数摄动、开关管和传感器故障、电机同步电机故障等，提出了系列故障诊断和鲁棒容错控制方法，显著增强系统的鲁棒性、改善系统的动态性能。提出了级联整流器电压均衡鲁棒控制方法，显著提升了级联整流器电压平衡的鲁棒性，改善了网侧频率波动、负载突变和网侧谐波条件的电压均衡动态性能，并消除了上下层系统之间的耦合效应。提出了DC-DC变换器、逆变器故障诊断和容错控制方法，在稳态性能优化的同时有效提升了变换器的动态性能和鲁棒性，能够有效地应对负载、频率和直流侧电压波动等变化，提升了控制系统的兼容性和可移植性。.基于以上研究，解决了车-网耦合系统电气振荡产生机理及稳定性判据、振荡抑制鲁棒稳定裕度的优化、故障对闭环系统影响的有效定量描述、以及面向nu-间隙度量性能优化的容错控制设计等关键科学问题。.部分成果在世界首台适用于单相25kV工频牵引供电制式的车载电力电子变压器原理样机上应用并装车试验成功。