Water plays an important role in many industrial processes. A smooth running of complex process water systems is the requirement for a functioning cooling process. Different sizes and measured values are decisive. These include, among others, the pH value, the redox value or the conductivity of the system water. In order to monitor these values, a wide variety of sensors capture and make data available.
Since 1932 the Dr. Hartmann Chemietechnik GmbH & Co. KG takes care of the process water systems of their customers. The goal of the free potential analysis of the experts from the Smart Data Solution Center Baden-Württemberg (SDSC-BW) was to support the engineers in their evaluations. The focus was not only on the qualitative aspects but also on the possibility of making predictions. It was examined whether and how the data obtained so far can be used to predict when a critical limit value is exceeded or fallen short of.
On the basis of Smart Data methods, an approach has been developed that makes use of both long-term and short-term features of the sensor data for the time series prediction. In addition, an extended Smart Data approach can predict the occurrence of relevant malfunctions very well. For a prediction time window of one hour, an accuracy of at least 95% is achieved. In order to apply and expand these findings in the future, the SDSC-BW experts will now turn to internal data and processes.
For further information, please visit: www.sdsc-bw.de
Data Innovation Community
Dr. Hartmann Chemietechnik GmbH & Co. KG, Smart Data Solution Center Baden-Württemberg
Andreas Meier, firstname.lastname@example.org
From May until June, 2017