Automated abnormality detection

自动异常检测

• 批准号: 131080
• 金额: $ 3.14万
• 依托单位: MILNE RESEARCH LIMITED
• 依托单位国家: 英国
• 项目类别: Feasibility Studies
• 财政年份: 2013
• 资助国家: 英国
• 起止时间: 2013 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=131080
• 关键词: Automated; abnormality; detection

Recent years have seen a huge growth in the complexity of manufacturing plant, and the measurement and control systems that it uses. Process control, condition indicating, performance measuring and QC parameters are routinely measured, howver the prupose to which this data is put are only two-fold: 1) keep the plant operating within a control regime and 2) assess the condition, performance and efficiency of the operating plant and hence of the process.The project will investigate the feasibility of developing a novel automated algorithmic approach to abnormality detection (i.e. a simple and automated means of assessing whether a plant is deteriorating, drifting in terms of efficiency/performance or operating in a sub-optimal manner.

近年来，制造工厂及其使用的测量和控制系统的复杂性大幅增长。常规测量过程控制、条件指示、性能测量和QC参数，但这些数据的目的只有两个：1）保持工厂在控制范围内运行，以及2）评估状况，的性能和效率的运作工厂，从而过程。该项目将研究开发一种新的自动算法方法来检测异常的可行性（即，评估设备是否正在恶化、在效率/性能方面漂移或以次优方式操作的简单且自动化的手段。