Sound purchasing decisions depend upon properly soliciting and evaluating proposals from various vendors. This article will focus on the first step: assembling a request for quotation (RFQ) that specifies your requirements in a manner that assures you will receive offerings that meet your needs and that you can compare. A follow-up article will look at evaluating the quotations you receive. While we will focus on distillation/absorption/stripping column internals, many of the general principles apply to all types of process equipment. Some, as we will point out, relate particularly to new units and others to retrofits/revamps.
If you want to have any hope of comparing quotations, the RFQ must be specific enough so that all offerings are on a common basis. In addition, management must clearly define what you absolutely must achieve as opposed to what is desired. Penalties for missing both throughput and product quality specifications must be identified. In the throughput case, does the amount reflect what marketing believes it can sell, or what a signed contract, with penalties, calls for? In the quality case, is the specification what is desired, or is it a threshold value that must be achieved to meet the definition of product? In some instances, an upper limit on total cost or downtime can be a valuable aid in the screening process.
Are you looking at packing, trays or both? A detailed treatment of this question is far beyond the scope of this article. In general, trays tend to be more predictable, whereas structured packing wins at vacuum; it can be a toss-up between the two at atmospheric pressure. High pressure and/or high liquid loads favor trays and random packing. Some special considerations that can influence the choice of device are the corrosiveness of the system and whether it contains suspended solids or is otherwise fouling. Vendors need to know if the material is non-Newtonian because design models are normally based on Newtonian fluids.
For either a new column or a retrofit/revamp, you need a set of design loadings. The most common way of obtaining them is from a simulation. However, make sure you can really believe the simulation [1, 2]. (Also, when running the simulations, determine the sensitivity of the separation to reflux versus the number of stages so high-efficiency options can be evaluated.)
Don’t base actual design loads on the simulation alone, though. Consider additional factors. For example, how certain are you of the physical properties of the system? Are they well-known hydrocarbons, or an exotic collection of components whose properties must be estimated by some esoteric means? Is the feed composition or quantity likely to change as the upstream catalyst ages? Is there a startup case? Is there an alternative feed case? Is there a future case? The minimum rate that realistically might occur can be important. Trays not only might weep, but can lose their downcomer seal. Excessive weeping has been blamed for vibration leading to tray damage. A minimum rate to a packed tower might affect distributor performance.
Size up the column
For a new column, it is always a good idea to do your own preliminary (or final) column sizing. Excellent sources for sizing methods for nonproprietary hardware are available [3, 4]. Members of Fractionation Research Inc. (FRI) have access to its rating methods. The major vendors, such as Koch-Glitsch and Sulzer, offer user-oriented rating programs via their Web sites. An important point to bear in mind on new installations is that the cost of a shop-fabricated column with its trays or packing (in ordinary metallurgy) is usually less than half the total erected cost of the column, reboiler, condenser, instruments, pumps, piping, etc. Thus, modern high-capacity devices that save six inches or so on the diameter of a new column might not be justified.
For a retrofit/revamp, find out the original design and rating, if possible. When you hear: “My device will give you 25% more capacity,” you should ask: “25% more than what?” Most likely the column was designed for about 75% of flood; this has been standard practice for more than 40 years. Therefore, the tower shell with redesigned, modern internals might be good for a significant throughput increase without going to the more exotic proprietary internals. Hopefully, plant records are good enough to check for observed bottlenecks. A talk with operators who have been around for a while invariably is a good idea, and a test run, if possible, is always useful.
Also, make sure to check the auxiliaries. Simulations will give you the required condenser and reboiler duties at the desired new conditions. It is pointless to spend a lot of money on new internals if the reboiler cannot handle the required duty. This is when it is particularly useful to have determined the reflux/theoretical-stage relationship necessary to achieve the desired product rate and purity. If the desired new conditions make the existing column shell and auxiliaries very tight, consider all options. High-efficiency internals might allow reflux reductions. Conversely, high-capacity devices at slightly lower efficiency might save a column shell if the condenser and reboiler are adequate. Low-pressure-drop packings, where trays might otherwise do the job, might ease reboiler approach temperatures.
Check the system limit
In many cases for a retrofit/revamp, assessing the system limit or ultimate capacity of the column shell can be very useful. The concept is quite simple: In a turbulent vapor/liquid flow field, there is a maximum, stable drop size that is a function of system properties only. If that drop size is less than or equal to the size that will be entrained due to the velocity of the rising vapor, all the liquid will be blown out of the column. This is a hardware-independent value; it can govern for devices that depend upon gravity separation, particularly in low-surface-tension applications, such as hydrocarbon services.
In refining applications, system limit tends to govern at pressures above 150 psig or below atmospheric, and in atmospheric-pressure services involving high liquid rates. Internals vendors almost never report or even check this value. Published models, their derivations, comparisons with test data and recommendations are summarized in references 5 and 6. The original versions developed by FRI have been released and are available through the Oklahoma State University Library. For packing and dualflow trays, the correlation given in its Topical Report No. 34  applies. For trays with downcomers, a version of this correlation is described in the FRI Annual Report for 1961 . Some proprietary devices on the market assist gravity and claim to get past this barrier. They tend to be very expensive, have relatively high pressure drop and normally are only justified when the column loading approaches the system limit.
Several other important constraints must be determined before you actually go out for quotations. Vendors need to know how much downtime is acceptable. Another consideration that affects both time and cost is whether welding to the column shell, such as for moving tray rings, will require the vessel to be recertified. The best source of advice on this point is your insurance company or state inspector. They also can advise if the vessel was fabricated to a superceded edition of the code and whether this presents any complications. If the vessel will have to be recertified and it was shop fabricated and tested, you need to verify that the foundation can sustain the hydrostatic test load. Up-to-date shell drawings showing manway sizes and locations must be assembled for inclusion in the inquiry document package. Other information that the suppliers might require before a firm quote can be given are discussed below.
Preparing the RFQ documents
If you want to be able to compare quotations on a reasonable basis, it is essential that you do not hand the vendors a duty specification followed by a blank sheet of paper. Specify as many things as you are certain of, including your company standards. To take advantage of the vendors’ experience and know-how, tell them you will accept alternatives to the base bid, not in the base bid. Also, you need the vendors to supply their definition of percent of usable capacity or percent of flood for all devices offered. These commonly used terms do not mean the same thing to all vendors. Some use a capacity definition based upon loss of separation efficiency, whereas others use hydraulic operability as an upper limit. There can be a considerable difference between these two loadings depending upon the device and system. In addition, particularly for proprietary devices, ask the vendor to state the number of installations of this device, and how many are in this service and the size; also, request operating or test data. If a vendor is going to report on throughput of an existing column before and after installation of this device, make sure the vendor tells you all that was altered (the existing unit might have been limited by a reboiler and/or condenser that also were changed).
For the technical specification part of the RFQ, the first thing obviously is the design column loadings. Modern simulation programs give the loading for each stage. For practical purposes, columns are divided into sections in which one design loading is used to set a constant tray design or packing size (although loads from top and bottom of a section might be checked if large changes are occurring). Feed and product withdrawal points, etc., usually define such sections. For each section, you need to specify the number of trays or height of packing. (Note that the prediction of stage efficiency falls to the purchaser.) Within each section, the stage with the maximum load is chosen. Common practice is to use vapor to and liquid from a stage. For distillation columns with a feed near its boiling point, the bottom section normally governs. There are some common cases, such as strippers or packed fractionating beds, where the maximum load occurs at the top and the limiting section loads consist of liquid to and vapor from the section. You must specify the vapor and liquid mass rate and density. Liquid viscosity and surface tension are required by most rating methods. Vapor viscosity and molecular weights are also useful. Diffusivities and slope of the equilibrium curve are needed if efficiency estimation is desired.
You might wish (or need) to specify the maximum-allowable-pressure-drop per tray or foot of packing. Based on the considerations discussed above, you might wish to set the maximum allowable approach to flood (or entrainment effect point for low-pressure columns). Also, if you have determined a derating factor for this system, you must specify it and explain its basis — e.g., foaming.
Under trayed column requirements, the first item is the diameter and number of trays in each section followed by tray spacing and tray type (moving valve, fixed valve, dualflow, sieve, bubble cap, etc.). You might wish to fix the number of passes or hole size for a sieve or dualflow tray. If you are asking for a valve tray, it might make sense to specify fixed or moving valves (and whether you want one or two weights of valve if moving) and set valve size (standard or mini). This is where you need to get all offerings on a common basis, with vendor alternatives as options. Also, you should specify the material, particularly if there is a process requirement. Based on your experience or company standards, you might wish to set the maximum downcomer velocity or calculated backup.
For a packed column, the first item is the diameter and height of each bed. A single section might have more than one bed if redistributors are being used. The packing type can be structured or random, the choice of which is outside the scope of this article. For structured packing, the type and size is required. For random packing, the type and size must be specified, as well as the volume to be shipped. (The packing is shipped in boxes or bags and then dumped into the column where it settles differently – usually 10% to 15% extra is required.) Also, for random packing, specify if you desire bed limiters or hold-downs.
For any type of packing, define the material of construction and its thickness (at least to get all quotes on a common basis). You also might wish to specify the type and material of the supports. In addition, the supplier needs to know the diameter and location of loading manways.
Detailed discussion of liquid distributors is beyond the scope of this article. They are expensive and take up space that could be occupied by packing. However, inadequate liquid distribution has long been cited as the reason for packed column failures . Lockett and Billingham provide a good guide for checking the sensitivity of the separation in question . At the RFQ-preparation stage, the key requirement is to specify what is needed in such a way as to obtain comparable quotations. For gravity distributors, the key parameters are the number of distribution points per unit area, the percent open area for vapor flow, possibly a minimum orifice size, a minimum liquid level at reduced flow conditions, the installed level tolerance and whether a shop water test is included in the quotation. For pressurized types, key parameters are the number of nozzles and the pressure drop.
Other important details
Whether the column is trayed or packed, the supplier needs to know the feed condition, particularly if the feed is partially or totally vaporized. You must supply the quantity and density of feed vapor and liquid at the column, both after any control valve and at the feed-point elevation. Many failures have been traced to not properly taking flashing into account.
All vendors should be asked to quote the installation services included or, if not included, the services offered and the rates. They also should detail the installation time required for the base case and for all options.
Many forms are available for specifying distillation-column internals requirements. FRI has compiled and released a set of forms as a service to the industry.
By following the principles we have outlined, you should receive adequate and comparable proposals. In the next article, we will discuss how to evaluate these quotes.
John G. Kunesh is a part-time consultant on distillation based in Red River, N.M. He served as technical director of Fractionation Research Inc., Stillwater, Okla., until his recent retirement. E-mail him at JGKunesh@aol.com. Raymond M. Sowiak is a senior process engineering specialist at Sunoco, Philadelphia. He is Sunoco’s technical representative to FRI and is a member of FRI’s Design
|1. Kister, H.Z., “Troubleshoot Distillation Simulations,” CEP, p. 63 (June 1995).
2. Kister, H.Z, “Can We Believe the Simulation Results?” CEP, p. 52 (Oct. 2002).
3. Kister, H.Z., “Distillation Design,” McGraw-Hill, New York (1992).
4. Stichlmair, J.G. and J.R Fair, “Distillation: Principles and Practices,” Wiley-
VCH, Hoboken, N.J. (1998).
5. Fitz, C.W. and J.G. Kunesh, “Column Hydraulics: System Limit/Ultimate Capacity,” Chem. Eng. J., 88, p. 11 (2002).
6. Stupin, W.J. and H.Z. Kister, “System Limit: The Ultimate Capacity of Fractionators,” presented at Intl. Conf. on Distillation and Absorption, Baden-Baden, Germany (Sept. 2002).
7. Stupin, W.J., “Ultimate Capacity of Fractionators,” Topical Report No. 34, FRI, Special Collections and University Archives, Oklahoma State Univ. Libraries, Stillwater, Okla. (Jan. 1965)
8. “1961 Annual Report,” FRI, Special Collections and University Archives, Oklahoma State Univ. Libraries, Stillwater, Okla. (1962).
9. Robinson, C.S. and E.R. Gilliland, “Elements of Fractional Distillation,” McGraw-Hill, New York (1950).
10. Billingham, J.F. and M.J. Lockett, “A Simple Method to Assess the Sensitivity of Packed Distillation Columns to Maldistribution,” Part 1, Trans. I.Chem.E., 80, Part A, p. 373 (May 2002) and Part 2, 81, Part A, p. 131 (Jan. 2003).