What Is Anomaly Detection?
An “Anomaly” is a state or condition that is abnormal. Some people refer to this as “Asset Monitoring”.
Anomaly Detection is a mathematical process that continuously monitors data such as process conditions, materials properties and perhaps also quality results to detect whether something is abnormal or unusual. This monitoring can be on individual variables (“tags”), using limits, statistical distributions and frequencies or entire multivariate “systems” that describe the entire asset. Multivariate monitoring is like observing a spherical ball and when that ball distorts sufficiently then an alert is generated.
In order to create and calibrate an anomaly detection system you expose our system to “normal” operations which it remembers. Once calibrated, the anomaly detection system continually monitors the process conditions, materials properties and/or quality results, individually or as an entire system, to make its determinations. When something is not normal, an alert is issued in a variety of ways, such as by email or logged for review and acknowledgment via Intellect web services.
Why Do You Need Anomaly Detection?
As an asset manager, engineer, or operations manager you may not always be aware of what is going on. You certainly want to know when things are not operating normally.
Abnormal conditions most always precede a problem, so you can catch problems *before* they occur. Make corrections during office hours, on your schedule, instead of after a 3 AM phone call.
Abnormal conditions can also be good! Perhaps you are producing abnormally good product or experiencing high operational performance. You will be quite interested in what is going on and why
In Brief
- Anomaly detection informs you of abnormal operating conditions.
- Our detectors observe multiple measures, individually or as a system.
- Be aware of unusual conditions early on.
- Make corrections on your schedule, not 3 AM
Products Used
- Intellect Server
- Intellect Designer
- Intellect Anomaly
Benefits
Anomaly detection offers a variety of benefits, some of which are mentioned above:
- The engineers and asset managers can see unusual operating conditions for their processes.
- Avoid equipment damage, expensive repairs and downtime by being alert before problems occur.
- Capitalize on unusually good operations. See it, know it, understand how to make better product, yields, rates, etc.
- Observe processes using sophisticated multivariate, multidimensional analytics giving competitive advantage.
Many of these are cost avoidance or “soft” benefits, but valued by our customers.