Hands-On Industrial Internet of Things

EMAs

EMAs are the first layer of analytics, which are normally used for descriptive, diagnostic, or short-term predictions. Signals are acquired and sent to a rule engine directly using a data-stream mechanism. Simple rules analyze signal by signal to discover any potential deviations from standard behavior. A simple threshold rule is a common example. EMAs are not as simple as they might appear: the engine should avoid raising the same event multiple times within a time range due to oscillations of the signal. The following diagram shows a typical example:

Example of excursion-monitoring analytic raising the same event multiple time in a ...