TrendMiner provides interfaces to leading historians such as;
- OSIsoft PI
- Aspentech IP21
- Honeywell PhD
- GE Proficy
- Yokogawa Exaquantum
in order to assist process engineers to realise the value of their investment and quickly fault find and understand process relationships without the need for advanced modelling.
|Based on pattern recognition algorithms, patent-pending TrendMiner Search empowers the process expert user to find historical trends without the need for a special skill data scientist or big data infrastructure.||Billions of data points over thousands of sensors, but you are still flying blind without time and location specific context. Capturing and sharing important knowledge is key to the success of process behaviour analysis.||It's a given that things that occur once, will likely occur again somewhere in the future. The best way to increase the efficiency and safety of a production plant is by predicting abnormal behaviour based on live data.|
Why use TrendMiner rather than just doing things the same old way it's always been done?
- Searching for the relevant information on large historians is time-consuming
- The value of data is decreasing over time (rather a fix now than in months' time)
- Data is too often historised in case of (cf. Dark Data leading to large volumes of data to sift through)
- Knowledge on demand not available: Learning from each other and from historical events quite difficult.
- Knowledge is leaving the company: Baby-boom exodus “Within the next 10 years 43% of the knowledge workers in the industry will retire"
TrendMiner brings tools for the average user enabling them to achieve success rapidly in an environment that traditionally required power users and months of data analysis.
TrendMiner works out of the box and connects directly to industry leading historians to deliver value without the need for a resource-draining project.
TrendMiner (Joint Innovation) can be broken into four key capabilities.
|allows rapid access to historical data over long periods of time and for identified context without the wait associated with large datasets.|
|Search||brings data filtering and advanced searching such as pattern search, causal search, layer comparison search (tag compare of different time frames or layers) directly into a web-based trending tool. Automatic detection and labelling of periods of oscillation or drift over defined limits can be added to the historical data.
|Capture||allows operators to record end of shift notes and other input. Pattern searches can be labelled for rapid selection during future data analysis.
|Monitor||once an event has been defined TrendMiner can monitor for future occurrences and automatically create event records as they occur. Events can be predictive or simply a recognition of multi-variant data conditions.