Valid data is the basis for determining key figures and evaluating plant effectiveness in digitized production plants. Values generated in the production process from sensor data of machines are usually not validated or missing. Ultimately, no correct statements can be made about the efficiency of the machines and plants!
The causes for poor data quality are manifold. It can be non-calibrated data on the part of the OEM, or the sensors have outdated settings. Often updates of production programs cause changes in the value ranges of sensors. Another cause can be the change of the production to another product size, thereby values are outside of validities, if they are not adjusted.
With Pure Data, Quantis has developed an application to obtain data, check it and monitor its quality on a daily basis. The tool is indispensable for quality management, because only key figures from validated data allow a well-founded evaluation of plant effectiveness.
The basis is the raw data acquisition. Conditions and rules for checking the data streams are defined accordingly for each sensor of all machines, counters and states are defined per machine and sensor. Based on this individually created set of rules, the data streams are analyzed, evaluated, and validated during ongoing operation.
The set of rules that Pure Data creates for each plant assesses the complex relationships and dependencies that the machines in a plant have with each other.
The target is to calibrate the sensors of the production plants in such a way that the plant operator can generate and calculate relevant KPIs with the data obtained.
The application during operation makes sense since a permanent evaluation of the data streams guarantees a constant data quality. If limit values or validity values are exceeded or not reached, a notification is automatically sent via e-mail or dashboard. In case of malfunctions, it is possible to intervene quickly without disturbing the production process and, if necessary, to extend or change the rule catalog.
The calculated data is presented in an easy-to-understand dashboard. Quantis visually prepares the key figures for customers in its Pocket Factory platform and makes them available in self-generated dashboards with widgets such as heat maps, histograms and much more. The most important dashboards are already stored as templates in the Pocket Factory and can be easily adapted or extended as needed.
The application can be implemented in the production line within a short time.
Pure Data offers smart features to detect non-valid data. All values are checked and examined for deviations:
Defective sensors; values are at zero, although plant is running
Values that are outside of validity ranges; Voltage is over 240 volts, although the value must be between 220 and 240 volts
Dependencies; counter counts, but the machine stops
Wrong values; the speed of the machine is negative
Logical values; the operating hours counter may only count up by the value 1 in one hour
Plateau values; temperature of a machine remains constantly high for hours
Undefined values; values are delivered, whose condition is not present
Summary
Valid data from the production plants generate meaningful key figures on production. Permanent analysis ensures consistent data quality. In the event of changes in the data situation, the person responsible is notified by the system and can thus make quick corrections that do not interfere with the production process.
Pure Data can be used in all production plants across all industries. This is also possible for older plants.
Comments