Ontology-based computational engineering
Generate an ontology for standard chemical engineering applications : The observation is that one needs a very few basic equations to simulate a very wide range of chemical engineering applications. We would like to establish such a base library which we then can use to generate process models and simulate them for given conditions.
Implement model for some of our experiments and simulate process : We have a number of processes in the teaching laboratory, which we aim at simulating their behaviour given a set of operating conditions. One of the more complex problems are to establish a module for the software environment that provides the link to the computational fluid dynamic code, specifically OpenFoam.
Implement the models for the biofuel process EcoLodge: We want to model the behaviour of the plant used in this project with the aim to improve performance.
Design of experiments: The MoDeNa project uses meta models to increase performance of the computations. The idea is to replace complex models by approximations that are much quicker to compute whilst reducing accuracy on an acceptable level. This is done by fitting meta / surrogate models to the simulated complex model.
Bio-fuel project EcoLodge: The project seeks to produce bio-fuel using two coupled bio reactors linked to different separation units. The process is also modelled and simulated with the objective to explore alternative units, in particular separation technologies, and configurations, as well as optimise the operating conditions.
Reactor technology – simulation technology
Static mixer reactor: This is an advanced reactor type that is used in specialty chemical production. We have modelled such a device for BASF’s specialty polymer research division and are planning to use our installation for the teaching laboratory. The reactor is equipped with a fibre-brag temperature measurement device, which measured 8 temperature in the center-axis of the reactor. The technology is using modified glass fibres and lasers.
Residence distribution bench: We are developing measurement devices for a flexible bench arrangement where the apparatus of interest is preceded with a pump, an tracer injection and a tracer measurement unit. At the outflow, a second, identical measurement device is added. The signals from the two measurement devices are used to identify the residence distribution of the apparatus. Fluid (CFD) simulations of the injection and measurement devices are used for design of the devices and also for validation of the established models for the bench and the apparatus.
Model-process interface / meta models
Process identification: Multi-wavelets shall be used to identify process models for multi-scale processes.
Discrete-event dynamic systems
Design for safe processes: In the past we have developed a method to generate discrete-event dynamic models from continuous models. These can be utilised for control, but also for hazop analysis and design of alarm and warning systems.