Test bench for testing the aging process

The Fraunhofer UMSICHT test bench demonstrates a catalytic process.

Aging process

Making a precise statement about aging processes contributes to the efficient operation of technical catalysts, as this gives plants new opportunities to intervene in the catalyzed process with foresight. Plants can either counteract catalyst aging by taking appropriate measures (e.g. by changing the temperature) or by planning the replacement of deactivated catalyst material in such a way that plant shutdowns are minimized.

A typical aging process of technical catalysts is thermal deactivation (sintering). For the oxidation of methane emissions from ship engines, a technically relevant, heterogeneously catalyzed process, the aging process of the catalyst developed by Fraunhofer UMSICHT is to be recorded experimentally. In the catalyst considered here, the aging process means that it is less and less able to convert methane into CO2 over its lifetime. Process experts have the option of reactivating the catalyst from time to time by increasing the temperature, but this also shortens its lifetime. At a certain point, reactivation is no longer possible and the catalyst must be replaced.

Test bench

For the simulation of aging processes and thus for the generation of measurement data for the ML model, a trainer at Fraunhofer UMSICHT was used. Using the trainer, the conversion of methane (CH4) into CO2 can be measured with molecular precision and activation processes can be carried out. It is possible to record a large number of measured variables, whereby the variables methane, carbon dioxide, water and the furnace temperature have turned out to be the most relevant variables.

The system consists of three components:

  • Gas supply
  • Furnace unit
  • Analytics

More information

 

Project page
»PrädiKat«

You want to learn more about the project "PrädiKat"? Then visit the project page and find out more.

 

 

ML model

The task of the ML model is to predict the remaining lifetime of the catalyst as accurately as possible. You can find out how this is done on the page on the ML model.

 

Application

The PrädiKat tool includes a web application that acts as a communicator between the ML model and the user, displaying the result of the catalyst evaluation. More about the application can be found on the application page.