作者: Margret Bauer , Sethabile Madolo
DOI: 10.3182/20070606-3-MX-2915.00011
关键词:
摘要: Abstract Process data captured from the instrumentation of a plant is frequently sampled and quantised. Quantisation can result in poor quality if quantisation level large. This paper investigates effect on data-driven analysis methods, particular first order statistics, power spectrum an oscillation detection method. Analysis shows that large distorts results obtained process analysis. Investigation temperature measurements are prone to due low accuracy. An automated algorithm presented estimates historical data.