Stages of Process Plant Design

Staged design is the norm in professional practice, each stage providing the information needed to pass on to the next stage for approval. Conceptual design decides plant size, location, and broad technology. Detailed design works out all but the very finest details of design. The final stage of design produces detailed deliverables allowing the plant to be built in accordance with the designer's intentions. These stages may be deliberately combined in order to fast-track the project, though this will have negative cost and quality implications.

Keywords

Design stage; conceptual; detailed; construction

General

Process plant design (and indeed almost all design) proceeds by stages which seem not so much to be conventional as having evolved to fit a niche. The commercial nature of the process means that minimum resources are expended to get a project to the next approval point. This results in design being broken into stages leading to three approval points, namely, feasibility, purchase, and construction.

This is why Pahl and Beitz's systematized version of the engineering design process resembles that which applies to all engineering disciplines (including chemical engineering's process plant design) as practiced by professionals. It may very well also apply to fashion design. Design is design. Is design.

Note that in recommending Pahl and Beitz's approach I am not seeking to enter the academic debate on how the design process ought to be done. Having read many books on engineering design across many disciplines, I found Pahl and Beitz's description to be one of the closest to how design is done. That is the subject of this book.

The basically invariant demands of the process are the reason why everyone who designs something professionally does it basically the same way, even though chemical engineers are often nowadays explicitly taught a radically different approach in university (if they are taught any approach at all).

Conceptual design

Conceptual design of process plants is sometimes carried out in an ultimate client company, more frequently in a contracting organization, and most commonly in an engineering consultancy.

In this first stage of design, we need to understand and ideally quantify the constraints under which we will be operating, the sufficiency and quality of design data available, and produce a number of rough designs based on the most plausibly successful approaches.

I am told that, in the oil and gas industry, the conceptual stage starts from a package of information known as Basic Engineering Design Data (BEDD), which is often confused with (Process) Basis of Design. BEDD includes, typically, information to start the concept design such as:

 General plant description

 Codes and standards

 Location

 Geotechnical data

 Meteorological data

 Seismic design conditions

 Oceanographic design conditions

 Environmental specifications

 Raw material and products specifications

 Utilities

 Flares

 Health Safety Environment (HSE) requirements

Other industries have alternative formats, but initial information packages ideally cover many of the same areas.

Practicing engineers tend to be conservative, and will only consider a novel process if it offers great advantages over well-proven approaches, or if there are no proven approaches. Reviews of the scientific literature are very rarely part of the design process. Practicing engineers very rarely have the free access to scientific papers which academics enjoy, and are highly unlikely in any case to be able to convince their colleagues to accept a proposal based on a design which has not been tried at full scale several times, preferably in a very similar application to the one under consideration.

The conceptual stage will identify a number of design cases, describing the outer limits of the plant's foreseeable operating conditions. Even at this initial stage, designs will consider the full expected operating range, or design envelope.

The documents identified in Chapter 3 are produced for the two or three options most likely to meet the client's requirements (usually economy and robustness). This will almost always be done using rules of thumb, since detailed design of a range of options (the majority of which will be discarded) is uneconomic.

This outline design can be used to generate electrical and civil engineering designs and prices. These are important, since designs may be optimal in terms of pure "process design" issues like yield or energy recovery, but too expensive when the demands of other disciplines are considered.

At the end of the process, it should be possible to decide rationally which of the design options is the best candidate to take forward to the next stage. Very rarely, it will be decided that pilot plant work is required, and economically justifiable, but this is very much the exception; design normally proceeds to the next stage without any trial work.

There are academic arguments for including formal process integration studies at this stage, though this is incredibly rare in practice. The key factor in conceptual studies is usually to get an understanding of the economic and technical feasibility of a number of options as quickly and cheaply as possible. As many as 98% of conceptual designs do not get built, so you don't want to spend a fortune investigating them.

Client companies have advantages over contractors in carrying out conceptual designs, as they may have a lot of operating data unavailable to contractors, however, they do usually lack real design experience.

Contractors are in the opposite situation, while the majority of staff employed by many consultancies tend to have neither hands-on design experience nor operational knowledge.

In an ideal world, therefore, client companies would collaborate with contractors to carry out conceptual design. In the real world, this cooperation/information sharing is less than optimal.

"Conceptual design of chemical processes"

Douglas wrote a book of this name which essentially attempts to design chemical processes (whatever they are), rather than process plants.

He understands that the expert designer proceeds by intuition and analogy, aided by "back of the envelope" calculations, but sees the need for a method which helps academics and beginners to cope with all the extra calculations they have to do while they are waiting to become experts (who know which calculations to do).

The arguments underlying the academic approach which has since been built on Douglas's approach are helpfully set out in explicit detail. There is an assumption that the purpose of conceptual design is to decide on process chemistry and parameters such as reaction yield. Choices between technologies (the usual aim of conceptual design exercises) are not considered. Pumps are assumed to be a negligible proportion of the capital (capex) and running (opex) cost, and 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, the methodology has a limited range of applicability. While it mentions other industries, it is based throughout upon examples taken from the petrochemical industry, and it is clear that the assumptions it makes are most suited to that industry.

Having declined to consider many items which are of great importance in other industries, Douglas finds time for pinch analysis, which was quite new when the book was written. Perhaps this really was a worthwhile exercise for the novice process designer in the petrochemical industries of the 1980s, but there are many process plant designs in 2014 which do not have a single heat exchanger.

In the majority of industries, process chemistry is a job for chemists, and from the plant designer's point of view is in any case usually limited to choosing between a number of existing commercially available process technologies.

Douglas offers a plausible approach to the limited problem he sets out to solve, few of whose assumptions I can argue with in the context of his chosen example. He attempts to offer the beginner a way to choose between potential process chemistries and to specify the performance of certain unit operations in a rather old-fashioned area of chemical engineering.

However, the problem which this methodology seeks to address 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 do not really get built in the developed world any more.

The approach does, however, hang together coherently, in a way more recent developments based on it do not. A good amount of effort goes into constructing as rigorous a costing as is possible at the early design stage (ignoring the issue of the items which are left out).

In essence this book seems, in my view, to reflect the slight wrong turns and oversimplifications which, followed by successive oversimplifications...