Inadequate Preliminary Engineering Design: A Common Reason for Major Problems

I suspect most people have heard the truism “Take the time to do it right or take the time to do it over”.? Nowhere is this truer than in the design of a pilot plant. I am all too often asked to do a design review, a cost or schedule review, or called in to consult on a startup or operation encountering major problems only to find the main issue was a badly inadequate preliminary design. I am often consulted as to the best approach to getting the lowest cost new unit in the fastest manner. The client is almost invariably aghast when I suggest that taking their time and spending enough money on the design in the beginning is the best approach to both. To many organizations this is counter intuitive.

?

Contractors must bid on numerous potential projects. A relatively few are usually approved. (Figures of 25-50% are often quoted. The higher number is usually for a firm that has a somewhat dedicated “stable” of clients who will usually go to them for almost any project.) As a result, there is a strong incentive to develop a design with minimal engineering and effort. That makes the contractor look good to the client (they were fast and responsive) and minimizes their out of pocket costs (which won’t be recovered if the project is not approved). It also helps them crank out enough proposals to keep the pipeline of potential work full enough to stay busy. These designs are rarely thought through or analyzed in detail. They are full of inconsistencies. This flow loop has blocks, bleeds, and bypasses and this one does not; do both need them? Does neither need them? Is there a reason one does and the other does not? They are usually based on the contractor’s “experience’ or “similar units”. Is the distillation tower even close to properly sized? Is the separator large enough? Little if any calculations have been performed. Line sizes are too small, probably leading to excessive pressure drops, or too large, leading to the potential for settling or at least higher residence times. Control valves will be very small leading to significant operating and maintenance problems. Invariably many of these problems will emerge and lead to costly change orders later. Others will go unnoticed until operation when problems arise that adversely influence safe and efficient operation or make data scatter and non-repeatability a major issue. In many cases they can be fixed with time and money; sadly, much more of both will be required than if the design had been more complete and well analyzed properly from the beginning. Often the best possible results are still very poor as rebuilding a large part of the pilot plant or taking a different approach is just not economically feasible. Many, fortunately not all, contractors will not be bothered with the issue as they can address it with change orders or revised estimates and recoup their costs (and more) later. The client is not so satisfied.

?

Many organizations are not much better. They similarly rush into the design process so they can develop a cost and schedule estimate to se if the project is economical enough to be approved. They often lack the knowledge base of a contractor, so their preliminary designs are frequently based on guesswork or a fast discussion with a contractor. Numerous areas are overlooked or not resolved. Does the line need to run at 35 C requiring heat tracing or can it tolerate cooling to room temperature? Is condensation in the line an issue or not? Questions such as these can have significant cost and schedule impact and hence need to be resolved early in the design process. Similarly, many components are sized with only cursory – if any – analysis. I understand if you think a 5 gal can will be large enough for the entire run but at your maximum flow rate, but the can will be emptied in three hours which is unlikely to be long enough to complete a run. Have you considered the vapor flow rate to the separator as I think it is undersized by a factor of 3 or 4.

?

When these types of issues surface early enough, they can be addressed. Unfortunately, often by the time I get involved a preliminary cost and schedule have been developed based on erroneous assumptions and it is my unenviable task to try and explain why neither is remotely credible as the basic design is too flawed. Most times the organization listens and works with me to revise the design appropriately leading to a revised, and more realistic, estimate and schedule. This always results in different expectations, wasted time and effort, and often needless delays. Other times, the organization is unwilling to realistically address the issues, usually because of the political fallout, and begins an increasingly desperate attempt to salvage the design. They cut the desired flow rate but do not think through how much longer it will take to accumulate the desired amount of product. They add a diluent to avoid having to heat trace the system but fail to acknowledge how much more challenging it will be to do the final separation. They ruthlessly eliminate all desired features without a careful evaluation of which of them are not just desirable but critical. This often leads to a death spiral where the proposed unit becomes less and less likely to be able to achieve the desired end results.

?

The worst case is where this type of organizational design is presented to a less experienced contractor. In this case the two issues feed themselves and produce the worst possible result as both sides increasingly support an unrealistic design and attendant incorrectly low cost and schedule.

?

So, what can an organization do to avoid these issues?

?

The answer is to clearly get a reasonably detailed and carefully developed design before any estimating or scheduling. The design needs to be realistic. It should start with a carefully thought through block flow diagram and mass balance on all streams. This often involves some assumptions, estimates, or suppositions which should be clearly documented and presented for assessment by others. Then the bock flow diagram should be expanded into a process flow diagram (PFD) showing all the major components, lines, and key equipment. Developing this into at least a preliminary piping and instrumentation drawing (P&ID) is always prudent as this often highlights issues that have not been apparent until that point. (See How Big Things Get Done: The Surprising Factors That Determine the Fate of Every Project, from Home Renovations to Space Exploration and Everything In Between by Flyvbjerg and Gardner for a great discussion of how the best projects “Think slow and act fast”.)

?

At the PFD and P&ID stage, the organization needs to constantly ask questions such as:

?

• How will I fill the feed vessel?
• How will I empty the product vessel?
• What happens if I need a higher or lower flow?
• How will I control the flow, level, and pressure?
• Do I need a sample here and how is the best way to take it?
• How will I clean this out? Purge it out? Vent it down? Drain it out?

?

These and dozens of other questions are critical to getting the proper design.

?

Similarly, the organization should reassess all the assumptions that went into the design. Are we sure that precipitation and attendant plugging is not an issue? Should we test to make sure that condensation is not an issue at the proposed room temperature? How good is our assumption that no recycle will be required? Concerns that the basis may not be correct are always present. It is not realistic to try to resolve 100% of them 100% of the time. However, it is critical to resolve the most important ones all the time. It is just as important to make sure that you have fallback positions in case you are wrong. Having to add a recycle system is not desirable and will be more costly and time consuming after initial construction. It can be done if room was provided, if some necessary taps or tees were provided and if the consequences of the need for a recycle were included in the original design. Will you need a compressor? How will you determine the recycle rate? Do you need sampling perhaps online to evaluate the outlet product and or recycle stream? As long as you have thought through the consequences and done some advance planning, the extra costs and effort can be minimized (although never eliminated).

?

Design reviews are critical elements of getting the pilot plant design right. (See Pilot Plant Design Reviews: An Important Tool for Minimizing Start Up Costs and Improving Operability,

https://www.dhirubhai.net/pulse/pilot-plant-design-reviews-important-tool-minimizing-start-palluzi/ for a more detailed discussion.) These take time and cost money. They require experienced pilot plant personnel. The reviewer should be a "cold eye”, someone who has no investment in the outcome. The review should include at least a preliminary safety review to identify safety issues and ensure they were addressed. (See Preliminary Safety Reviews: A Tool To Save Money, Avoid Problems, And Ensure Safer Operations,

https://www.dhirubhai.net/pulse/preliminary-safety-reviews-tool-save-money-avoid-problems-palluzi/ for a further discussion.)

?

All designs should include a resident cynic, someone charged with ensuring that desirability does not overtake reality thus avoiding excessively optimistic, and usually unfulfilled, outcomes. Too often I have been told something like “if it plugs the process would not work” and so this potential is ignored. When it does plug, the organization is thus presented with a choice of stopping all work or trying desperately to address something better delt with during the design. Having the tenth person, a person tasked with always taking an opposing view on everything to force a more careful evaluation, is often prudent.

?

Another factor leading to inadequate designs is incomplete drawings. Drawings are often rushed, riddled with errors, omissions, mistakes, and typos. The assumption is they will be fixed “later” before construction. However, when large numbers of errors exist on a drawing it becomes progressively harder to ensure all of them are caught. Reviews become more cursory and incomplete. Problems are overlooked. Simple errors continue to grow and multiply. Cost estimates and schedules are badly underestimated, leading to further problems. In 49 years of doing pilot plant designs I cannot honestly say I have every produced a P&ID set that is 100% perfectly correct; some sort of mistake or typo always creeps in. But I do ensure they are 99.9% correct so that hopefully the error is trivial and easily caught or corrected. If you do not take the time to get it as correct as you can then you will suffer later. Contractor drawing that still have numerous “small”, or “trivial” errors and typos invariably are not constructed properly and come back to create problems later. A good as built set of drawings does not mean the unit will ever work properly. I know someone will argue, validly, that perfect is the enemy of good but I will strenuously argue that good is often- sadly – not good enough for success.

?

Finally, pilot plant work is inordinately equipment dependent. One can recognize the need for a smaller pump, a higher pressure device, or a higher temperature component but – it they don’t exist – the design cannot work. All too often when I mention that a component may be an issue to locate the concern is brushed off with a casual comment that the organization will just look harder. Many times this is true and the component or equipment is found but not always. Often, even if the component is located it will have problems with the proposed service. It may meet the temperature and the pressure requirements but not be available with the right materials or elastomers. Custom or customized components are costly and long delivery. They also have an increased risk that they will not work as specified. New or novel devices may be the perfect answer, but some will, sadly, not work as proposed. A design based on these is almost certain to encounter insurmountable issues. A good design always ensures that the equipment it needs is available. I may not have the nitrogen regulators selected but I make sure all the pumps and instruments are available.

?

I understand the compelling need to get many pilot plants in operation as soon as possible. However, without careful design organizations often find themselves with pilot plants that do not work well enough to meet their needs. Or, that required a lot more time and money to get to the point where they provide the necessary data or product. Spending enough time and effort of the design upfront can avoid needless lost time and money later and is almost always faster and less expensive in the end. See my discussion on “We Need It Faster”: Issues with Accelerated Pilot Plant Schedules, https://www.dhirubhai.net/pulse/we-needfaster-issues-accelerated-pilot-plant-richard-palluzi/ ?for more information and other concerns,