Multivariate time series pattern discovery in a complex continuous industrial process

复杂连续工业过程中的多元时间序列模式发现

• 批准号: 428632-2012
• 负责人: Wong, Andrew
• 金额: $ 3.47万
• 依托单位: University of Waterloo
• 依托单位国家: 加拿大
• 项目类别: Idea to Innovation
• 财政年份: 2013
• 资助国家: 加拿大
• 起止时间: 2013-01-01 至 2014-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=532593
• 关键词: Multivariate; time; series; pattern; discovery

With the advancement in sensor and information technology, organizations are now capable of collecting massive amounts of data continuously and storing them in large databases. The availability of these data makes it possible to solve previously unsolvable problems and to improve existing processes. In particular, the Ontario Government introduced smart metering sensors in all Ontario homes and small businesses back in 2010. Our industrial partner, Energent, believes that the data collected by these sensors contain valuable information that can improve energy consumption during peak hours and redistribute energy supplies to different regions across the province. The problem becomes how to extract this valuable information. While the availability of massive real-time data is an exciting technological advent, we are faced with the challenge of information overload - trying to make sense of the vast amount of data to determine the appropriate course of actions.

随着传感器和信息技术的进步，组织现在能够连续收集大量数据并将其存储在大型数据库中。这些数据的可用性使得解决以前无法解决的问题和改进现有流程成为可能。特别是，早在2010年，安大略政府就在所有安大略家庭和小型企业中引入了智能计量传感器。我们的工业合作伙伴Energent认为，这些传感器收集的数据包含有价值的信息，可以改善高峰时段的能源消耗，并将能源供应重新分配到全省不同地区。问题是如何提取这些有价值的信息。虽然大量实时数据的可用性是一个令人兴奋的技术进步，但我们面临着信息过载的挑战-试图理解大量数据以确定适当的行动方案。