Time to clean 20MB data? Around 5 minutes!
Measured data need to be quickly validated to prevent errors slipping into the system. But how can those one or two anomalous channels in 3000 be identified?
The automated ICE-flow data validation tools can be run immediately after each test run – they can even run while the next test begins. The validation procedures are defined ‘in the office’ and downloaded from the server to the test station to ensure that consistent and approved procedures are applied. More advanced users can access the system interactively, if authorized. Comprehensive procedures are provided, such as an automated anomaly detection sequence–which not only identifies the more obvious problems such as drift, dropout, spikes, overloads – but also abnormal variances between channels and several other subtle test errors that can only be detected statistically. Many of these can be corrected automatically.
No data is overwritten, and all changes are tracked for audit purposes.
Typical tests involve thousands, occasionally tens of thousands of long time histories, so ICE-flow has a graphical way of showing the current status of the data. As soon as channels of data are available each indicator changes from white (no data) to black. As validation and cleaning proceeds, the indicator light turns green when new data is “good,” yellow when a problem has been discovered (but the data can be used), or red when the data is “bad” and not recoverable or useable.
3000 channels - and any bad data can be seen at a glance!
Capabilities in this section are provided by GlyphWorks Anomaly Detection and/or ICE-flow TestManager.
Next stage in the process: Data classification and processing.
For more information on ICE-flow, click here
For more information on nCode, contact your sales office