For example, we applied Anomaly to monitor an oil and gas platform and found that frequently, more than once a day, there were serious upsets.  From automated "compressor trips" to operators turning hand valves, the persons managing the assets were unaware... until Anomaly.

We then applied Anomaly to a distillation operation... things were worse.  Whether it be significant feed fluctuations, temperature control loop instabilities, pressure fluctuations or simply out of specification quality, the process frequently had difficulties.

Keeping An Eye On What's Important
Anomaly monitors your key performance indicators and other critical variables “live” and on-line, alerting you to unusual conditions encountered.

Using a blend of statistics, expert rules and multivariate cluster modeling, Anomaly keeps a key eye on your operations, detects when systems start to go out of envelope and raises alerts to take action.  Anomaly is already in operation around the world for monitoring:

• Critical Large Capital Equipment
• Oil Platform Operations
• Oil Well Surveillance
• Distillation Column Operations
• Large Scale Pumps

Point-n-Watch
Anomaly is self-learning.  Deploying it can be as easy as pointing to the variables you want monitored and pressing the "GO" button.  Alternatively, you can pre-expose Anomaly to a history of machine operations and apply your own rules for "Good", "Bad" and "Outlier" operating conditions.

Integrates with Your Control Systems
Anomaly communicates with control systems either directly with OPC or via an OSIsoft PI Historian.  It uses an iManageData data handling template so you get the full benefit of data access, cleansing, filtering, preprocessing, conversion and transformation.

Properly Suited Technologies
Anomaly uses a blend of statistics, expert rules and multivariate clustering to understand what is "Normal", "Good", "Bad" and "Outlier" conditions.

Statistics are used to understand how variables are historically distributed and alerts you to infrequent values or N-sigma deviances.

Expert Rules are integrated so you may personally define what is "Good", "Bad" and "Outlier" conditions.

Multivariate Clustering detects when your system is distorted based on all variable values simultaneously.  There are cases where all variables may be "normal" but your process is stretched out of envelope considering inter-variable relations.

Additionally, the user can classify alerts as Good or Bad based on a single and a multivariate perspective.

Since Anomaly uses iManageData templates, you can monitor moving averages, rates of changes, logs, ratios, products, sums and a broad range of derived variables too.

For your convenience, NO programming is required.

Desktop and Server Versions
Anomaly is available in both a desktop version and a Process Intellect "process".  The desktop version enables you to construct monitoring solutions where the server process enables you to put multiple Anomaly solutions on-line in robust 24x7 operations.  The Process Intellect version enables the results of the alerts to issue email alerts and integrate with control and optimization logic.