Can't Scale it Down?

My last column discussed scaledown as a method for resolving process problems. However, scaledown, like scaleup, is not easy to accomplish successfully. You might find numerous hidden traps along the way.

Perhaps the problem cannot be duplicated in the laboratory or pilot plant. In other words, the process cannot be scaled down. Some processes cannot be scaled up, and some processes cannot be scaled down.

Solids systems tend to be nonscaleable. Filtration is a prime example. Crystals that are to be filtered are made the very first time in the plant. All prior laboratory studies are suspect.

Much in filtration depends on cake porosity and particle shape. Small changes in cake porosity and particle shape can bring about big changes in filtration rate and production. However, cake porosity and particle shape are not measured and certainly not controlled directly. In these cases, it is fortunate that nature is reproducible.

However, if the plant has an upset, a scaledown laboratory study probably would not contribute to the understanding of the situation. Time might be better spent in cause-and-effect plant studies and experiment design.

Problems often are not what they seem to be. A problem at a particular process step might have originated upstream. If a filter press is not filtering, the press might not be to blame. The filter press might be doing exactly what it is supposed to do.

Problems in filtration can be caused by a whole host of items, including an impurity increase, a feed location change, an increase in turbulence and a change in the heat transfer coefficient. A scaledown of a filtration process is not likely to identify these changes.

Another example of a potentially nonscaleable process is heat transfer with solids. In the plant, the solids in the outer perimeter of the tank might fuse together because high local heat transfer in their surface layer combines with poor mixing and solid suspension in that region. In the laboratory, where mixing, solid suspension and heat transfer are excellent under the poorest of conditions, replication of plant conditions is difficult. Engineers can waste a lot of time and effort trying to accomplish scaledown on a nonscaleable process.

Suppose you could scale down the phenomenon and study it in the laboratory. You know everything about the mechanism causing the plant to have problems.

What are you going to do about it? Even with this information, you might not be able to make changes in the plant to solve the problem.

Knowledge and information generally help lead to solutions, but not always. Suppose you were able to replicate the solids fusing phenomenon above. However, you did the replication in a confined space with no mixing and solids suspension.

How are you going to use this information? You just proved you could fuse the solids. However, you still have no idea how to solve the plant problem.

So what do you do? This situation recalls the old adage from graduate school: It does not matter what you do ," just do something. If the plant problems are significant, then it is reasonable to take educated guesses related to their solution. A surface coating that disappears upon processing might inhibit solids fusing.

Do not be meticulous and fastidious.

Easy to say, but studying and solving problems in the plant are more difficult. Plant changes are expensive, and success might be difficult to assess. Political risks also are associated with plant changes. Plant acceptance issues arise. As one problem is solved, another arises.

Be prepared for political problems. An important initial need is to put plant personnel on your side. Make only small changes to explore possibilities.

It is important that you let the process speak to you. You must be prepared to listen to everyone and be prepared to learn. You do not ," and cannot ," know everything.