Algorithms and methodologies of controlling unsteady flows

点击翻译按钮获取中文摘要

The objective of this project is to develop a computer-based virtual environment (VE) platform for supporting expert data-analysis and decision-making. The platform is illustrated for an important and pervasive industrial challenge: controlling unsteady flows (e.g., liquids/gas flow reactors). Embedded in the platform is a set of enabling algorithms that make robust and reliable recognition of key patterns from complex and massive flow data; and that render the recognized patterns effectively for three-dimensional visualization and multi-user interaction. Advancing the state of the art, these algorithms are unconventional solutions to current research challenges - such as incompleteness of data for flow pattern recognition, subjectivity of representations and inadequacy of simulation tools for real-time analysis. The significance of the VE platform is exemplified through applications to chemical flow reactors, a core business interest of our supporting organization NOVA Chemicals Corporation. The project fits well into the target area of Information and Communication Technologies, under the research topic of "(e) From Data to Knowledge to Action - unstructured pattern recognition." With strong industrial support, we envision such a VE platform to boost the productivity and competitiveness of many Canadian sectors; in which experts across disciplines must routinely decode complex and massive data together in their problem-solving and decision-making activities.