The Benefits Of Good Data

The value of good, correlated, historical data can be immense:

Intermittent Failure Cause Tracing

The Half-Million Dollar Loose Screw!

As soon as Vsystem was installed and running with 100ms historical data being collected, it happened again at a paper mill. The web broke. This had cost over $500,000 in the past and now, at last, the mill engineers had good, correlated data to study from the motor drives and the PLCs. The data, shown in the picture below, shows the particular signals that indicate the fault (the analog traces) and the cause (the digital bars below the analog traces). The engineers quickly found the cause. Identifying the digitals as being from one PLC block led to the conclusion that the fault was caused by a loose screw causing intermittent power! The freshly installed monitoring system had paid for itself before its supplier had even had time to submit the invoice!

Process Signatures

A plant with a complex process found it difficult to characterize the process until Vsystem was installed. The Vsystem data enabled the staff to develop data signatures of key variables indicating good and bad operations. The screen shot below shows 174 pens or analog variables used by a plant using Vsystem to monitor the plant operation by looking for overall signatures and then drilling down to just the signals of concern by just clicking on the traces that they want removed from the display.

Predictive Maintenance

By daily monitoring data collected by Vsystem, engineers see problems early in their development and can then choose to either adjust production levels to maintain operation until the next scheduled maintenance shutdown or schedule an out-of-sequence maintenance period. This avoids very costly plant trips and allows operators to confidently operate closer to the edge of the envelope – increasing normal production as well. When trips to occur, data can determine if it is safe to restart immediately or if maintenance is required.

Process Problems and Optimization

By studying relevant Vsystem process data, engineers can better understand process issues and then implement modifications to the control of the process or address mechanical issues. The results are better product, higher and more reliable production, and lower costs.

Modern mathematical techniques can develop computer models from historical process data. The importance of data that fully represents all the plant and process measurements and influences is discussed in this white paper.

The Message:

The better the data, the more one can improve the process, reducing costs and improving quality!