Why I Needed to Write "An Applied Guide To Process and Plant Design" #3

Because there was no realistic textbook...

Upon taking up my academic post, I searched for publications which would support my teaching of a professional design philosophy. I found it extremely hard to identify a textbook whose methodologies were consistent with consensus professional practice. Furthermore, I found that there was little research to validate my professional experience, whether on engineering practice generally, the design process specifically, or on engineering education. My literature search was thus significantly more challenging than I had envisaged.

I reviewed a wide range of potentially relevant engineering textbooks published between 1970-date. Those which appeared to contain a design methodology closest in intention to the professional consensus design were as follows:

·        Sinnott & Towler’s ‘Chemical Engineering Design’ (2009)

·        Baasel’s ‘Preliminary Chemical Engineering Plant Design’ (1989)

·        Backhurst and Harker’s ‘Process Plant Design’ (1973)

·        Luyben’s ‘Principles and Case Studies of Simultaneous Design’ (2011)

·        Sandler & Luckiewicz’s ‘Practical Process Engineering’ (1986)

·        Peters, Timmerhaus & West’s ‘Plant Design and Economics for Chemical Engineers’ (2003)

·        Wells & Rose’s ‘The Art of Chemical Process Design’ (1986)

·        Koolen’s ‘Design of Simple and Robust Process Plants’ (2001)

·        Pahl et al’s ‘Engineering Design: A Systematic Approach’ (2006)

·        Erwin’s ‘Industrial Chemical Process Design’ (2013)

·        Douglas’s ‘Conceptual Design of Chemical Processes’ (1988)

·        Mecklenburgh’s ‘Process Plant Layout’ (1985).

I explored the context and content of each of these, and made the following observations.

The classic Sinnott & Towler contains a solid attempt by non-plant designers to provide such a methodology. However, its explanation of the professional design is not explicit. In my experience this has a tendency to lead to highly unrealistic student plant design work, apparent to anyone with substantial plant design experience who witnesses the outputs of a standard student “capstone” design project.

Baasel’s ‘Preliminary Chemical Engineering Plant Design’ was authored by an academic who undertook a placement in industry. In doing so Baasel comes to the view that process design is really plant design, and acknowledges the importance of what is so often glossed over in academia. Whilst the text has enduring value, there is no realistic modern design methodology to be found in this book.

Despite its authors being academics, Backhurst and Harker’s ‘Process Plant Design’ demonstrates a clear understanding that what engineers do differs from what universities teach them to do. The authors perhaps overemphasise certain areas (which they themselves describe as arbitrarily chosen) rather than offering a system level design methodology. Overall, however, the book is of limited value as a realistic design methodology due to its age and partiality.

Luyben’s ‘Principles and Case Studies of Simultaneous Design’ considers only steady state economics and dynamic controllability aspects of a process. Design for controllability is strongly emphasised in this approach, supported by computer modelling. I would consider this too partial an approach to represent a realistic design methodology, or even part of one. The problem is that in achieving his stated aim of producing something smaller and less encyclopaedic than ‘Perry’s Handbook’ (Green & Perry, 2007), Luyben has chosen only two elements. These are without a doubt important partial design elements, but optimising only two of the many variables in a design will not produce an optimal solution. The text thus offers a methodology for process design, not process plant design; partial rather than total design, and the substitution of modelling for design. Moreover, ‘Perry’s Handbook’ in any case offers no whole-plant design methodology. Its focus is on the areas of unit operation design traditionally favoured in university curricula, namely, engineering science and mathematics.

Sandler and Luckiewicz’s ‘Practical Process Engineering’ is potentially far more useful to a new designer, although it has regrettably become obsolete in some respects. The book represents the joint endeavor by an academic and a practitioner to remedy the shortcomings of academic engineering courses. It has a great deal to say about the importance of drawing, and practical hydraulics, and contains useful practical knowledge on trace heating, lagging, electrical power and motors. By contrast, it is less substantial on the matter of unit operation design - the term ‘vessels’, for example, covers everything from reactors to storage tanks. Despite its shortcomings, it still contains useful material that none of the other books cover, and I decided to use, expand upon and update some of this unique content in my book.

As the title suggests, Peters, Timmerhaus and West’s ‘Plant Design and Economics for Chemical Engineers’ has a lot to say about economics, but it also contains valuable content on technical report writing and delivery. Whilst the design content is not complete, the book has particular relevance for chemical engineering educators as a source of cost curves for academic costing exercises.

Wells and Rose’s ‘The Art of Chemical Process Design’ was authored by an academic and a software company representative. This book works on the premise, which I do not share, that design begins in the laboratory and proceeds via simulation and modelling. Whilst it recognizes the iterative nature of design, there is little art to be found in this book; rather it reads more as an attempt to fashion a science out of the art of design.

Koolen’s ‘Design of Simple and Robust Process Plants’ contains useful summaries of the implications of ten approaches to process or process plant design. As for Koolen’s own approach, simplicity and robustness are indeed heuristics in plant design but, of these, the professional consensus is that simplicity is not a key metric; the three most commonly emphasized parameters are cost, safety and robustness. Simplicity can often be an indicator of expert design (as expressed by Kelly Johnson in the KISS principle ‘Keep it Simple, Stupid’- (Rich, 1995)) but, like any other single metric, it can be overemphasized. Koolen quantifies complexity, which is an interesting and potentially fruitful approach, but his discussion of partial approaches, combined with a theorist’s attempt to substitute modelling and simulation for design, are indicative of a normative rather than positive design methodology.

Pahl et al’s ‘Engineering Design - A Systematic Approach’ is a general engineering text which captures the essence of engineering design practice (Eckert 2004, Anderson et al 2010) in a way that so many chemical engineering texts do not. It gives a number of realistic design methodologies, and clearly analyses and explains key aspects of the engineering design process, including those often glossed over in an academic setting. The book suggests both explicitly and implicitly that German design teaching is far more closely aligned with professional practice than in the English-speaking world. I found this book to be most closely aligned with a realistic chemical engineering design methodology and used it in my design teaching prior to the publication of my own work.

Erwin’s ‘Industrial Chemical Process Design’ explains how to use MS Visual Basic to carry out tasks to support engineering design (rather than solve engineering problems - computers can't solve problems). The book has a narrow focus, as evidenced by statements such as: “The great majority of the process engineer's work is strictly with organic chemicals”. Whilst this is a clear reflection of Erwin's own professional experience, as opposed to that of the majority of process engineers, Erwin is a practitioner, and his approach is practical. He recognizes upfront that even the programs he supplies with the book on disc “have not been put through an exhaustive beta test”, and goes on to suggest that the “quest of this book, to correct the program through your good ability”. Erwin is not implying that we should forget the IChemE's Guidance on Use of Computers (1999). This book will be valuable to those who wish to use MS Visual Basic to carry out design tasks (especially those in the petrochemical industry), with the caveat that the time taken to write and test a programme might compare poorly with that of other methodologies, unless the programme is to be reused many times.

Pahl et al provide a useful counter to the oversimplification of problems, by differentiating between a task and a problem. When there is a well-established methodology which produces unambiguous guidance on how to choose between design options, following such an approach is not a problem-solving activity, but is merely a task. Where data is gathered, entered into a computer program or design methodology, and a reasonably straightforward answer is produced, this is not engineering.

Much of what is done in academic engineering courses nowadays however focusses on such scenarios. Sinnott & Towler and other textbooks provide ‘turn-the-handle’ methods to ‘design’ various unit operations. So-called ‘design projects’ consist in many educational institutions worldwide of getting students to knit together a group of tasks of this nature, operating a simulation program like Hysys, or grinding through pinch analyses.

These books did not embody the concepts of professional practice articulated by observers such as Trevelyan (2014); rather they appeared to reflect the misconceptions which Trevelyan has also described.

In attempting to teach what I had learned in practice, it thus became clear that there were no realistic textbooks in the area of process plant design. Instead, the approach which had come to dominate teaching in the area had much in common with the work of Douglas (1988) which, it is explicitly stated, was intended to address the difficulty in teaching process design in an academic environment free of experienced process designers.

The approach to researcher-led ‘process design’ elaborated in academia can be traced back to Douglas’s seminal book ‘Conceptual Design of Chemical Processes’ (1988). As the title suggests, it attempts to design chemical processes, rather than process plants. Its author understands that the expert designer proceeds by intuition and analogy, aided by ‘back of the envelope’ calculations, but identifies the need for a method which helps students and academics to cope with all of the extra calculations they have to do whilst they are waiting to become experts.

The arguments underlying the academic approach since built on it are helpfully set out explicitly. It assumes that the purpose of conceptual design is to determine process chemistry and parameters such as reaction yield. Choices between technologies are not considered. Pumps are explicitly assumed to be a negligible proportion of the capital (capex) and running cost (opex) of a plant, whilst heat exchangers are assumed to be a major proportion of capex and opex. It is implicit in the chain of assumptions used to create the simplified design methodology that a particular sort of process is being designed. Like all design heuristics, it has a limited range of applicability. Whilst the author mentions other industries, the text is based throughout upon an example taken from the petrochemical industry, and it is clear that the approach is also most suited to that industry. While omitting many items which are of great importance in other industries, it finds time for pinch analysis, which was relatively new when the book was written.

Whilst all this may have been a worthwhile approach for the budding process designer in the petrochemical industries of the 1980s, there are many process plant designs in 2018 which do not contain a single heat exchanger. There are many industries where process chemistry is given to chemical engineers by chemists; and consideration of process chemistry is not an engineering task.

Douglas attempts to offer a beginner a way to choose between potential process chemistries and specify the performance of certain unit operations in a rather old-fashioned area of chemical engineering. However, the problem for which a methodology is offered is not one I have ever been asked to find a solution for. When I am asked to offer a conceptual design, I am being asked to address different questions, on plants with a different balance of cost of plant components. Petrochemical plants of the sort used as the example in this book are no longer built in the developed world nowadays. That said, the approach in the book does have a coherence that more recent developments based on it do not. For example, a good amount of effort goes into as rigorous a costing as is possible at the early design stage (ignoring the issue of the items which are left out).

In summary, whilst Douglas provides a plausible and pragmatic approach to the limited problem it sets out, the book seems to initiate the early subtle wrong turns which led engineering education, via successive oversimplifications and misunderstandings, away from a professional design philosophy and towards the researcher-led approaches which prevail in modern engineering teaching and textbooks.

#chemical #engineering #reallychemeng #processplantdesign #design #processdesign #research