Walzenpresse - Entwicklung einer innovativen Walzenpresse zur kontinuierlichen Granulierung unter Einsatz von Process Analytical Technology (PAT) - AiF/ZIM


Project Title
Modelling the Process Behavior of a Dry Granulation Process
 
Funding Code
ZF4469701US7
 
 
Principal Investigator
 
 
Status
Abgeschlossen
 
Duration
01-11-2017
-
31-08-2020
 
 
 
Project Abstract
oll compaction is a process for the dry granulation of powdery bulk material. It finds application in pharmaceutical, chemical and food industry. In the first stage the powder is compressed between two counter-rotating rolls. Therefore the material is compacted due to the effective pressure the rolls exert. The resulting intermediate product is a high-density consolidated ribbon. Subsequently, the ribbon is comminuted into coarse pieces and drops into the sieving area of the machine. Within the latter the ribbon is granulated to a specific size.

The granulation process improves flow characteristics of the raw material and prevents dust formation by particle agglomeration. In contrast to wet granulation techniques, dry granulation can be used for processing moisture-sensitive and thermo-labile substances, as neither wetting nor high temperature drying is involved. Further advantages lie in a reduced production time, as no drying step is required. The main drawback of rolling compaction is the lack of quality assurance during processing. The state of the art is taking samples and analyzing them offline.

The goal of this project is the development of an inline measurement system in real-time to enable control and documentation of relevant quantities during processing. Based on these, a fully automated process control system will be installed to identify and adjust deviations during the continuous granulation process. This control system requires a computational material model to describe the influence of the recorded quantities on the product properties (e.g. strength, compaction, tendency of agglomeration). Concerning this matter, creating a comprehensive model that is suitable for real-time predictions will be a substantial challenge.