Dr. MoYuen Chow, ECE Department
Dr. MoYuen Chow has been awarded $240,000 by the National Science Foundation to fund his research proposal titled "Small World Stratification For Power System Fault Diagnosis With Causality."
The award will run from September 1st, 2007 through August 31st, 2010.
Research Abstract - Small World Stratification for Power System Fault Diagnosis with Causality The objective of this project is to investigate and develop a fault diagnosis with causality methodology for power distribution systems using Small World (SW) strategy and Hierarchical Clustering (HC) technology to substantially improve fault diagnosis accuracy and reasoning. Causality is essential in fault diagnosis, yet it has not been well studied, especially with complex systems. Consider an anecdote of causality (or lack of causality) that ignorant people might think upon hearing a rooster crow, that the sun is rising, where the causal observation is established without sufficient knowledge of local dynamics (i.e., circadian rhythm) and/or global dynamics (i.e., the solar system). Historically, many pandemic, engineering, and socioeconomic disasters were due to erroneous causality. In contrast, identifying proper causality can trigger a major paradigm shift and technological development. For example, in medieval times, people believed that the sun revolved around the earth. After people correctly accepted the causality that the earth revolves around the sun due to gravity, the rebirth of classical learning brought about the renaissance of Europe, just as the development of Newtonian physics set off the industrial revolution. In general, diagnostic causality is easier to identify for a simple system (e.g., the analogy of the sun and the rooster) with humanity?s knowledge and heuristics; while the causality is much more difficult to identify for a complex system (e.g., the analogy of the sun and the earth) even with substantial domain expertise with the complex system.
Typical complex systems are large-scale, nonlinear, time-varying, and geographically dispersed with a wide range of dynamic operating conditions with both global and local network features. These systems (e.g., power distribution systems) often face harsh outdoor environments making them distinctly vulnerable to many different kinds of natural disturbances, which can be difficult to analyze and model. Faults are unavoidable and will happen in these systems. It is important to diagnose the faults with proper causality and restore the systems in a timely manner to maintain their vitality. Wrong fault diagnosis with wrong causality will affect subsequent fault management steps, including fault prognosis and fault mitigation, and the consequences can be enormous (e.g., the analogy of the sun and the earth). The project will use power distribution (a typical complex system) as a testbed to facilitate our discussion and to illustrate the effectiveness on the investigation of using Small World (SW) strategy and Hierarchical Clustering (HC) technology for fault diagnosis with causality on complex systems.
The intellectual merits of this project are: (1) The proposed work can provide effective and efficient use of available information to give accurate power distribution fault diagnosis and to discover proper fault causality knowledge, which can substantially enhance the subsequent power distribution fault prognosis and mitigation control. (2) The proposed work will provide a general framework from which to evaluate cases with similar symptoms, geographical settings and topological connectivity for the fault diagnosis and will give the context for causality. Not only can the framework be applied to power distribution networks, but it can also be applied to a wide variety of areas, such as power transmission systems, transportation systems, manufacturing systems, healthcare systems, energy systems, homeland security, etc., to encourage breakthrough opportunities and advancements in fault diagnosis technologies and performance improvements.
The broader impacts: There are extensive components of integrating research into education in this project, ranging from graduate and undergraduate, to high school students. Various means, including graduate and undergraduate research assistantships, senior design projects, and collaborating with the NCSU ECE K-12