Catalysts, detergents, fertilizers, instant food products are examples of industries and private persons consumables, which are typically available as particles. The solid condition is in comparison to the liquid state advantageous due to a reduced transport weight, simplified dosage and higher keepability. One of the most prominent apparatus types for production of the mentioned products is the fluidized bed.

Especially in chemical industry, the continuous operation of fluidized bed processes has become the operation mode of choice as this comes along with higher yield and production rates due to the reduction of time consuming and not completely reproducible start-up and shut-down times and less downtimes. Continuously operated fluidized beds are often designed as elongated horizontal chambers.

The aim of this project is the autonomous optimization of a fluidized bed reactor regarding product quality by means of an optimized residence time behaviour. This will be possible with the enhanced experimental data collection by means of a Lagrangian sensor for the granulation state in conjunction with a detailed simulation approach simulating faster than real time. Finally, this can be combined in such a way that the rCFD simulations fed with live data enable an autonomous reactor control strategy to optimize the process conditions of the continuous fluidized bed spray granulation in such a way that the yield is increased and batch faults are markedly reduced. In comparison to typical research on fluidized beds in academia, the investigations will be done on pilot-scale instead of lab-scale, which allows the integration of 3D effects as well as the investigation of scalability potential. The aim is the development of a process optimization tool, which can be easily transferred to other processes and also larger process scales. By achieving these goals, the design of continuous fluidized bed processes in academia and industry will be simplified, the processes become more predictable and material and energy will be saved.