It is the fond hope of engineering academics everywhere that design can be reduced to a problem that a computer can solve. Unfortunately, design is a perfect example of the thing computers cannot do at all, creating innovative solutions to ill-defined problems.

What computers are great at is optimizing very well-defined multivariable problems. Mathematical and graphical techniques and related computer programs to do this were developed back in the 70’s, and the use of these have come to replace proper design in many universities. However:

Premature optimization is the root of all evil. – Donald Knuth

Contrary to popular academic opinion, availability of accurate information means that design optimization is most usually and always best applied after a plant has been built, commissioned and operated for a while.

Design optimization tools such as modeling, simulation, and pinch technology are therefore poorly suited to use during the plant design process, with a few notable exceptions. Their use in academia is almost always misuse grounded in a lack of understanding of the constraints of professional practice.

The main thing that those who apply academic process optimization techniques to process plant design fail to understand is that the iterative nature of real design processes means that there is already a complex design optimization process going on.

Professional engineers, however, understand that each stage of design has its own natural resolution. There is no practical point in applying a technique with a resolution finer than the model it is being applied to.

The second thing which the academic approach fails to address is that you cannot meaningfully optimize a model which has not been verified by input of real world data.

Like microscopes, all design techniques have what we might call a limit of resolution. Microscopic resolution allows us to distinguish accurately between two lines. A microscope with insufficient resolution for the task to which we put it may give the appearance of two lines where there is really only one, or one line where there are actually two.

Similarly, a design technique with insufficient resolution may make two options seem equal where one is actually better, equal options significantly different, or even the better one worse. Misuse of process optimization tools for design is to me akin to what is called in microscopy “empty magnification”, where you make an image look bigger, but it actually holds no additional information.

What is known in academic circles as “Process Integration” is not design integration. It probably isn’t even really process integration. The process designer’s “process integration” balances a number of mutually dependent considerations. The design needs to be safe, robust and cost-effective, but safety and robustness do not come for free. A balance has to be struck.

As P&ID, GA, PFD, process and hydraulic calculations are developed, many choices have to be made about the broad outlines of plant layout, degree of redundancy of equipment, and basic approaches to safety.

Potential hazards have to be identified, quantified, eliminated or controlled. Materials and equipment have to be specified. Whilst doing this, installation, commissioning, maintenance, and non-steady state operation of the plant have to be considered. Past experience with other similar plants needs to be incorporated.

The process designer does not do this in a vacuum – they need to integrate the requirements of and insights from other disciplines. Optimizing a few aspects of the process, or even the whole process chemistry is not optimizing the overall plant design. It may actually be making it less optimal.

What is often meant by Process Integration in academia is use of a mathematical analysis of a system using one of what is now a wide range of mathematical, graphical or computer based tools, originally developed for beginners.

The problem these tools solve is one of handling a multiplicity of possible solutions. It isn’t so much that there are an infinite number of possible solutions to the question each of which has a number of subtly different implications, as that there are a great number of permutations to winnow for the best value of a single numerical selection criterion.

The tools can perform this winnowing process for us, but the fact that there is essentially one right answer, and a computer can find it better than a person tells us that this isn’t really engineering, and the problem is essentially trivial.

These tools may have some limited use in the final stages of designs which use a lot of energy, and have clear possibilities for substantial recovery of that energy. They may also be of use in identifying possible improvements to existing processes.

Starting a design from heat integration of a process at steady state without consideration of cost or other implications is trying to fit a job to a tool rather than the reverse.

Another buzzword in academia is “process intensification”. Professional engineers make processes as intense as they practically can, but no more so, to paraphrase Einstein.”

In summary, there is a radical disconnect between academic “process design” and real world process plant design which my book can help you understand.

Read more from Sean on SciTech Connect

An Applied Guide to Process and Plant Design is available for purchase on the Elsevier Store.

Use discount code “STC215” at checkout and save up to 30% on your very own copy!

About the Author

Professor Moran is a Chartered Chemical Engineer with over twenty years’ experience in process design, commissioning and troubleshooting. He started his career with international process engineering contractors and worked worldwide on water treatment projects before setting up his own consultancy in 1996, specializing in process and hydraulic design, commissioning and troubleshooting of industrial effluent and water treatment plants.

In his role as Associate Professor at the University of Nottingham, he co-ordinates the design teaching program for chemical engineering students. Professor Moran’s university work focuses on increasing industrial relevance in teaching, with a particular emphasis on process design, safety and employability.

Connect with Sean on LinkedIn here, check out his Facebook page here and stay up-to-date on his thoughts, research and practice at his personal blog here.