首页> 美国政府科技报告 >Robust Model-Based Fault Diagnosis for DC Zonal Electrical Distribution System.
【24h】

Robust Model-Based Fault Diagnosis for DC Zonal Electrical Distribution System.

机译:基于鲁棒模型的直流分区配电系统故障诊断。

获取原文

摘要

A key element of the U.S. Navy's transition to an electric naval force is an Integrated Power System (IPS) that provides continuity of service to vital systems despite combat damage. In order to meet subsequent survivability standards under a reduced manning constraint, the IPS system must include a fault tolerant control scheme, capable of achieving automated graceful degradation despite major disruptions involving cascading failures. Toward this objective, online modelbased residual generation techniques are proposed, which identify explicitly defined faults within a stochastic DC Zonal Electrical Distribution System (DC ZEDS). Two novel polynomial approaches to the design of unknown input observers (UIO) are developed to estimate the partial state and, under certain conditions, the unknown input. These methods are shown to apply to a larger class of systems compared to standard projection based approaches where the UIO rank condition is not satisfied. It is shown that the partial-state estimate is sufficient to the computation of residuals for fault diagnosis, even in such cases where full-state estimation is not possible. In order to reduce the complexity of the system, a modular approach to Fault Detection and Isolation (FDI) is presented. Here, the innovations generated from a bank of Kalman filters (some of them UIOs) act as a structured residual set for the stochastic DC ZEDS subsystem modules and are shown to detect and isolate various classes of faults. Certain mathematical models are also shown to effectively identify input/output consistency of systems in explicitly defined fault conditions. Numerical simulation results are based on the well-documented Office of Naval Research Control Challenge benchmark system, which represents a prototypical U.S. Navy shipboard IPS power distribution system.

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号