Premium
Weak signal detection in SPC
Author(s) -
Dohnal Gejza
Publication year - 2019
Publication title -
applied stochastic models in business and industry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.413
H-Index - 40
eISSN - 1526-4025
pISSN - 1524-1904
DOI - 10.1002/asmb.2480
Subject(s) - change detection , control chart , detection theory , statistical process control , computer science , signal (programming language) , detection threshold , outlier , step detection , process (computing) , threshold limit value , statistical power , pattern recognition (psychology) , statistics , artificial intelligence , mathematics , detector , real time computing , telecommunications , environmental health , global positioning system , programming language , operating system , medicine
Changes in the behavior of dynamic systems are detected based on changes in the monitored quantities or their characteristics. This detection usually takes place by monitoring the time evolution of a variable and detecting the change at the time when a predetermined threshold is exceeded. This threshold is determined on the basis of the detection scheme requirements, in particular, the probability of false alarms and the detection rate for the actual change. In some cases, however, a change does not come suddenly, but certain “hints” in the system behavior can be observed that may indicate a future change. For example, an increasing frequency of outliers can result in a sudden permanent change in the signal. The occurrence of some “unusual” frequencies often indicates an imminent change. For example, an increasing correlation value indicates an undesired process status. Detection of these “subliminal” hints can often improve the characteristics of the detection scheme, especially the detection rate for the actual change. In this paper, we will deal with the detection of weak signals in statistical process monitoring using a control chart with adaptive control limits.