Disaster Prevention Through Intelligent Monitoring

Abstract

Despite various tools and systems that can monitor complex engineering environments, bad things still happen regularly in all types of engineering industries. An intelligent system designed to monitor certain indicators, regardless of engineering industry, that might predict catastrophes would ultimately reduce the potential for loss of human life and property. In this article, 10 catastrophes were researched to identify their root causes and the various root cause combinations. These documented catastrophes covered a broad spectrum of engineering including oil, gas, nuclear, rail, air and space. The root causes identified in the investigation reports were grouped under 10 trait headings and their efficacy was tested using a qualitative fault tree of a credible catastrophic failure scenario. Each trait was adjusted to signify various levels of failure and fed into the prototype system representing the fault tree. While near real-time monitoring and trend analysis was investigated and shown to support an intelligent system that might predict catastrophe, one of the surprising additional results from the research was highlighting the need to standardize the approach to investigative reports and audits of existing systems. Reporting in the same “technical language” and looking for specific condition levels for each of the traits could provide a true picture of asset condition and the required funding prioritization, as well as assisting the dissemination of findings to all engineering industries.