Pilot Plant Cost Estimates: How Do I Get An Accurate (Enough) One?

We need an estimate of the cost for our proposed new pilot plant. We do not have any real funding or time for the estimate. We have not done any significant design work. We have many areas of uncertainty. Despite this, we need a realistic cost and schedule to evaluate our options. What can we do?

?

This conversation, or a similar one, has occurred several times a year for every year of my career. Organizations require an estimate of what a pilot plant will cost early in the selection, decision, and/or budgeting process but have no real funding or time to make one since they do not have a completed design. It is then my sad duty to explain their, often limited, options.

?

Option 1: Hire an experienced pilot plant designer to make a conceptual estimate based on what the designer thinks the final design requires. This is a high risk option as you need three things in that designer. First, they must be very experienced in pilot plant design and construction. Second, they must be good pilot plant estimators, which are not common at a preliminary or conceptual basis. Third, they need to be able to accurately assess and visualize what the final design of your unit will be from very cursory information. Many experienced pilot plant designers fail badly on this last point. Locating someone with all three required skills is very difficult. (See "But What Will It Cost?" The Trials and Tribulations of Pilot Plant Cost Estimating, https://www.dhirubhai.net/pulse/what-cost-trials-tribulations-pilot-plant-estimating-richard-palluzi/ for the problems of estimating this type of work.) There are a few proprietary systems that can help, but finding and evaluating them is difficult. (See Developing Screening Estimates, R P Palluzi, Chemical Engineering Progress, July 2011 for a discussion of one system.)

?

An alternative is to go to a firm that specializes in designing and constructing pilot plants and ask them for an estimate. This has several potential issues. These firms are not common so there are limited alternatives. Many have areas of specialization but struggle in others. Finding one that can estimate your pilot plant from a conceptual design is challenging. No pilot plant design form can afford to spend significant effort to develop a useful design and credible estimate for free and stay in business. Hence, either you need to expect to pay for their work to develop the design and estimate or you will get a quick guess. These are often spectacularly wrong. Such firms, particularly larger ones, are often more expensive as they need to involve more of their groups and people in developing a realistic design and estimate. Instead of one person determining the probable control system, piping, and wiring, they need to involve three. Too often, they also oversell the accuracy of their guess or even their estimate basing them on their limited design work. Few have good information regarding how good these initial preliminary or scoping estates were versus cost of the final design. ??Being told that the unit may cost $X is a major problem if the final design costs $3X. Understanding why the initial estimate was so wrong is usually possible but rarely helpful. Often, if one cannot find an experienced pilot plant estimator then they remain the only viable choice. (See Contracting Pilot Plants and Research Support: The Good, the Bad, And the Ugly, at ?https://www.dhirubhai.net/pulse/contracting-pilot-plants-research-support-good-bad-ugly-palluzi/ for a more detailed discussion of common issues with contracting a pilot plant.)

?

Option 2: Bite the bullet and do enough design work to get a credible estimate whose accuracy meets your needs. The table below shows my opinion of the design work required for the class of estimate desired. Please note that the design areas are the minimum design work that needs to be completed before the estimate is started. More is always better.

?

Often a good preliminary design that has, at least, identified and addressed any significant issues or requirements may be good enough with an experienced pilot plant estimator. (See Pilot Plant and Laboratory Unit Cost Estimates: Should We Progress the Idea or Not? at https://www.dhirubhai.net/pulse/pilot-plant-laboratory-unit-cost-estimates-should-we-progress/ for a further discussion of how accurate and estimate you may need to make a decision.)

Key:

?A- Minimum required or higher risk of significant error

B – Preferred for better accuracy (target)

R – Virtually required or? higher risk of significant error

?The estimate classes referred to are indicated below. Other classes, while less common, do exist.

An experienced pilot plant estimator can often do better than the accuracies shown. An inexperienced pilot plant estimator will usually do worse, often much worse. Recognize that even an experienced process estimator will usually struggle to achieve these accuracies if the final design is too different from what they assumed it would be.

?

Many groups believe they address this problem by adding a large enough contingency to their estimates. Contingency is supposed to be an estimate for the uncertainty in the estimate, the knowledge that historically predicable extra costs will arise that – at the moment- cannot be identified. You don’t know how much bad weather will affect construction but you do know there will be some. You cannot be sure how much equipment delivery delays will cost but you are sure there will be some. Those are rather classic examples of contingency as reducing their probable impact is essentially outside one’s control. You can do some better work at estimating the impact by checking local historic weather or the problems on past projects with late deliveries but you cannot do much to control them. And you know some will inevitably occur. There are, however, numerous other costs that are included in contingency for a conceptual estimate that are more uncertain and yet can have a major cost and schedule impact based on what the estimator includes Iin the costs or assumes as the basis How many compression fittings and valves will be required versus what was assumed? How much support steel will the final design require? How many bolts and nuts, how much wire, how many terminal blocks, and a host of other things will eventually be needed? Many estimators have few tools for evaluating these costs and so often count on a larger contingency to cover them.

?

Contingencies range from 5-100% for unclassified conceptual or screening estimates to 5-10% for Class II control estimates. While a valuable and critical tool for estimating, no amount of realistic contingency can address an inadequately designed project. A poor or badly incomplete design ?leads to the estimator ?“guessing” what contingency may be necessary. The author, having been placed in such a position numerous times, can assure you that no one will – or should - accept a 200% or higher contingency. Most will baulk at 50% or more. Worse, these type of large contingencies hide that the estimate is, at best, a SWAG (scientific wild ass guess) and probably not worth the paper it is written on.

?

Studies on commercial capital projects, small and large pilot plants, lab units, and almost any other work continually show that the more design work provided up front, the better the estimate. This front end or front end loaded engineering and design work is critical to try and identify what really will be necessary for the project to succeed. So, just in case you can’t tell, I strongly recommend Option 2 as the best approach.

?

To try to minimize the potential for a truly terrible and inaccurate estimate, I can suggest the following guidelines.

?

• Engage an experienced pilot plant designer as soon as a need is recognized. This allows the designer to help guide you towards the key (i.e., costly and/or time consuming) elements of the design within the limited resources (time, money, information) available. This avoids wasting substantive time and effort pursuing “pipe dreams”, approaches that even a cursory review will indicate have little or no chance of meeting the basic project requirements. It allows the organization to spend more time and effort on more realistic designs that are more likely to meet their long term needs at a cost and schedule they can afford.

?

• Evaluate potential project costs with probable budgets as early as possible in the process. When the means (funding and schedule) ?are clearly unlikely to meet the ends (the required pilot plant) then the organization really needs to try and find a different path forward or a way to secure more funding. Waiting until further down the road never makes this easier. Hoping that more design work will eliminate costly problems or find faster and cheaper approaches is commendable but all too often unavailing. Gross mismatches of ?estimates versus funding only waste time and money with no good result. Hoping a clever design can cut costs or schedules by 10-20% is challenging but not unrealistic. Hoping it can cut them by 20-50% is progressively more unlikely. To achieve these sorts of reductions, a significant basis change - such as a smaller scale, less process operations, or eliminating areas of work - is always required. Walking away from a proposed unit is never easy but walking away before spending double the time and effort to get the same, non-viable answer is a lot more painful.

?

• Recognize that getting this assistance and a good estimate will take time and money so allocate enough for both. It will always cost more and take longer than you would like, but this expenditure early in the process pays rich dividends by avoiding problems later.

?

• Make sure you engage an experienced pilot plant specialist of pilot plant organization to make the design and estimate. Many commercial design organizations claim to have adequate pilot plant experience, much fewer few do.

?

• Be willing to compromise. Talk to the designer about lower cost/faster options that involve doing less, taking longer, or utilizing a different approach. Identify areas of lower risk that perhaps can be eliminated from the unit. Paying more to get higher purity feeds or slop more waste always hurts long term but may be a viable strategy when total unit costs are simply too high. Adding these capabilities back later can often be reasonably economical, although always more expensive and time consuming when done later, with good design work and initial planning. Some suggestions will simply not be viable, but a willingness to keep an open mind can result in some that warrant consideration.

?

• Be willing to consider hitherto undesirable alternatives. Acknowledge that you chose the wrong location and that a different one may allow a much lower cost and/or faster schedule. Consider that the pet requirement of a vocal team member which adds substantial cost may not be absolutely required upon a more careful review. Evaluate if taking 3 months rather than one month to produce the desired amount of product may well be a more cost effective approach for the initial stages. Decide if the time and expense of some additional larger scale laboratory work on a small laboratory scale pilot unit might not provide enough information to progress to the next stage (or decide progressing further is not viable). Again, some of these alternatives may not be viable. However, all too often they are not evaluated carefully and too casually discarded.

?

• Recognize when adequate funding is just not available. While stories of startups that staggered from one funding crisis to another yet were ultimately successful abound, the reality is that many more simply go under when their reach exceeds their grasp. If you cannot afford the path forward, trying to progress a plan without enough funding rarely ends well. Sometimes it is better to not try until you have secured more funding or developed a less expensive alternative approach.

?

• Consider cold eyes reviews of the proposed design estimate and schedule. These can often be helpful in evaluating alternatives not previously considered or identifying estimates or plans that are not realistic. The costs of such cold eyes reviews are trivial when compared with what they can identify or prevent.

?

I recognize some of these suggestions are rather unpalatable to a struggling organization. The results of progressing research without a good estimate of the costs are, however, usually much worse. So, consider the quality of your estimate(s) carefully.