Dynamic real-time optimization under uncertainty of a hydroformylation mini-plant

- Dynamic real-time optimization under uncertainty using chance constraints.
- Sparse grid sampling and back-projection for quantification of constraint adherence.
- Application on continuously operated mini-plant.
- Hydroformylation of long-chained olefins in microemulsions.

In this contribution, the chance constrained optimization approach is applied for the dynamic real-time optimization under uncertainty of a hydroformylation mini-plant. For this purpose, a dynamic model of the plant is designed in MOSAIC consisting of 23 units, 12 components, and 25 streams. Three uncertain parameters were identified: the activation energy of the main reaction, a reaction inhibition factor, and the feed stream flows. The goal of the optimization is the maximization of the product content in the outlet stream, while minimizing the catalyst loss for economic reasons. The problem is solved with IPOPT every 5 h whereby set-points for seven controls are calculated. A fully automated mini-plant is operated for 200 h, during which offline gas chromatographs analyse samples on an hourly basis to provide information on the simulation and optimization runs. Overall, a stable mini-plant operation with a constant yield and efficient phase separation was achieved.