ana::tool

ana::tool identifies unknown and unusual conditions and enables operators to react in a timely manner to faults in the monitored system, determines normality of this data and triggers an alarm when a significant deviation from the norm is detected

Start

• ana::tool is installed with application-specific default settings
• ana::tool requires training before it can generate alarms
• ana::tool automatically starts collecting data and indicates when it is ready for training.

Training

• ana::tool can train itself continuously or be trained manually
• Manual training can be easily performed via the graphical user interface of moni::tool
• For manual training a static time window is used
• For automatic training a moving time window can be used
• Events are automatically detected and can be removed, the remaining data represents the “normal state”
• ana::tool weights individual parameters and creates thresholds for alarms that optimize sensitivity and minimize false alarm rate Adaptation to slow changes
• ana::tool detects events as well as gradual changes (such as seasonal changes)
• Historical data is collected and patterns of different water compositions can be stored. ana::tool recognizes water whose composition matches these patterns and sensitivity is not affected by changes between known compositions.
• With automatic training with a moving time window, the system automatically adapts to slow changes in the water matrix. During periods when the water composition changes continuously, the sensitivity of the alarm system is reduced. The actual sensitivity is displayed.

Output

• ana::tool generates a single binary alarm value
• The value indicates how far the current water composition deviates from normality
• Additional single alarms can be defined, e.g. a static limit value for a parameter
• ana::tool requires user confirmation for communicated events
• The alarm value can be used to trigger various actions and reactions

Unique functions

• Open interface to all types of sensors and measuring devices also from other manufacturers. Other data sources can also be used.
• Online and real-time validation of data before event detection significantly reduces false alarm rate.
• Graphical user interface provides an overview of the entire station at a glance
• Web-based interface, allows access at any time from any location
• Designed to take advantage of the extensive capabilities of spectral data
• Includes an indicator of actual alarm sensitivity
• Can calculate the contribution of individual water sources A, B, C, … to the total water mix