How computational fluid dynamics gives you competitive advantage
The water industry but also other process industries use piloting as a crucial step before full-scale implementation of a treatment or production process. Everyone with piloting experience knows that it costs significant amounts of money and effort. But apart from that, time to market is more important than ever. Hence, piloting should be done carefully, smartly and efficiently, and the following question is a crucial one to answer: 'what exactly do we want the piloting to teach us?'. A second question that follows is: 'can we reach (part of) these objectives + learn more using 'virtual' (i.e. computer-based) piloting?'. Computational Fluid Dynamics (CFD) has huge potential to complement, strengthen, reduce or replace expensive piloting efforts. Taking that route, the process is built, operated and modified using a 3D computer model, which offers unprecedented testing freedom and high problem solving speed. We will talk about this topic at numerous events in 2020.
It's time to realise that CFD cannot only be used for fixing performance issues (reactive usage) or checking a design decision (that in fact already has been made). Another level of CFD usage is the proactive and innovative use for scale-up and disruptive design optimisation. And CFD and real-life experimenting can go hand in hand here.
Why the time for computer-based process design is right
I should not have written this article in 2000. 2020 is the perfect timing. We have succesfully completed multiple cases and are currently working on many more. Some marjor drivers are shown in the following figure. Societal and market pressures include:
- Process intensification (doing more with less recources, space, ...)
- Resource recovery and efficiency
- More stringent and new regulation
- Speed of society and the market
- Smarter end users (which means they become more demanding)
Why do technologists invest in piloting trials?
Piloting is typically applied to:
Scale-up a process (unfortunately, water or other molecules do not scale along with a reactor)
Test process improvements (eg design modifications)
Assess how a certain process would perform at a specific site or in a specific situation
Regardless of the objective, we have observed 3 categories of learnings:
And this brings us to the core of this discussion. Realising that on-site staff, analytical measurements and expensive equipment can easily consume hundreds of thousands of EUR in a relatively short period of time, we should ask ourselves the crucial questions mentioned earlier.
The value of virtual piloting
Category 1 mainly relates to on-side experience building and accounting for the unexpected. For example: will these sensors function well (eg will fouling be an issue?),will unexpected dynamics or disturbances occur, ... ? And some processes rely on complex (bio)chemical reactions for which no suitable kinetic models exist yet. Also, it serves as a demonstration case with strong convincing power. Real-life testing is very important here.
Category 2 relates to learnings that can actually be assessed based on CFD. We have seen cases in which new equipment was built to assess hydrodynamic aspects and mixing. If performance was bad, the equipment (eg reactor) was manually opened and modified (eg inlet structures or mixers) after which new trials followed. In many of these cases, we could have completely replaced these tests by virtual CFD simulations. Using CFD, much more designs can be tested and on-site costs and efforts are drastically reduced. Any operational or design variable can be changed rather easily.
Category 3 relates to learnings that can not be measured or tested in real life. In some cases, certain experiments would be too costly (e.g. special pumps or other equipment are needed) or infeasible (e.g. 'what if particle or granule size would vary'). In other cases, the data that really matters cannot be measured accurately and in high resolution. Examples are shear (relevant for membrane applications, or mixing of specific biomolecules),shear distributions (relevant for eg flocculation),gas holdup (many treatment processes),... . A revolutionary application is the coupling of CFD with process kinetics. Especially if the kinetic mechanisms are complex, the simulations can lead to disruptive insights. In this category, we talk about 2nd or higher order impacts of mixing on process performance that cannot be assessed with human brain thinking.
It's not science fiction. The number of technology developers and vendors in our client portfolio has grown tremendously, exactly because of the points addressed here. Sure, some of you reading this will try to think about the weaknesses and threats. I'm doing the same if I look at the non-CFD route. I'm saying here that if you look at the opportunities in 2020, they are significant...
CFD can be used:
- To fix (urgent) performance issues (troubleshooting) (reactive),or
- To design processes properly before construction or implementation in real life (proactive),but also
- To explore novel designs based on the in depth insights and testing speed and freedom (innovation and disruption). We are already working towards this application.