This Month’s Puzzler
My company has received an order to build a skid to make a mixed gel, an intermediate for a personal care product. The skid will include an agitated batch reactor where a powder is added to produce the viscous gel. This is pumped through a heat exchanger and then blended with a fragrance in a static mixer before going to an agitated storage tank. The skid boundaries were to end with feed to the storage tank but have expanded to include clean-in-place (CIP) and utilities skids; the customer insists this shouldn’t affect delivery. We’re having trouble defining boundaries between the skids and have received nothing more than some product properties, i.e., viscosity, density and heat capacity at a single temperature. What can we do to ensure this equipment works as desired?
The Delivery Date Will Change
The first part of the description of the situation presents as a classic scope control problem. Before making a proposal, the supplier should impose a requirement that the customer specifies boundaries for rates, physical properties, performance properties, and conditions. If they are not specified in advance, proposals should include a disclaimer that final schedule and cost depend upon buyer-approved boundary conditions.
If the additional skids don’t change the agreed conditions of the process skid, then the process skid delivery schedule shouldn’t change. If any part of the process skid scope changes, then delivery should be expected to change.
The classic client request when multiple skids are involved is to optimize the system to minimize investment or operating costs. Optimization implies that boundary conditions, or even equipment in the skids, can change. In this case, the client should expect delivery to change. Optimization takes time. Of course, clients don’t have to be reasonable.
The second issue is on physical properties. What you need to know depends upon processing conditions. Knowing just some of the product properties will make work difficult. At a minimum, you need to know some reaction properties, physical properties under mixing shear, and physical properties that vary with temperature.
The client may not even know the properties. A lot of work can be done with a willing assumption of some degree of technical risk and expense. Extrapolation may be cheaper and quicker than generating properties. At a minimum, this approach requires thorough documentation. Nevertheless, even guessing the properties and documenting them takes time.
Knowing how the client thinks is critical. Some clients, no matter what, will always blame you for anything that doesn’t work. You never should agree to anything but having the client completely define everything in this case. I’ve worked with this type; after negotiation the quote from the client was: “We can accept anything except an arithmetic error.” Other clients fully understand the risks involved. With this type, reasonable agreements on the way forward work well.
There’s no easy answer to your question. A lot of what your company agrees to depends upon management’s willingness to accept financial risk. In no case should unknown physical properties or desire to meet schedule be allowed to create safety, health or environmental (HSE) problems. If more time is needed for HSE, then take more time. If more information is needed for HSE, then insist upon it.
Andrew W. Sloley, principal engineer
CH2M HILL, Bellingham, Wash.
Keep Asking Questions
It will be very difficult to design the skid to the customer’s satisfaction without additional information. Put this in a memo during the bid. Some vendors will think they’re off the hook if they build precisely to the blueprints but that’s ridiculous; your company takes responsibility once you agree to build a skid.
There may be some sources to mine for your use. You can start with the material safety data sheets (MSDSs) but these usually are incomplete, rendering their usefulness doubtful. Most manufacturers don’t even bother to complete the safety portions of these vital sheets, labeling auto-ignition temperature (AIT), flash point (FP) and even normal boiling point (nBP) and vapor pressure (P*) as “N/A” or “not available.”
Other options exist. Ask the customer for contact information and a letter of introduction to its ingredient suppliers. They are required to identify fire risks if they ship a product. Another alternative is an independent laboratory. A fourth choice, which often is all that’s left, is checking textbooks and similar sources. This will be a challenge. A last option is referencing a similar compound. For example, cellulose acetate has a defined AIT and flash point in NFPA-30; cellulose ether doesn’t.
Given the nature of fragrances, I already know there’s a potential fire risk. In 1992, I identified in EPA documentation iso-amyl acetate, an ingredient in a fragrance we handled. Iso-amyl acetate has a flash point of 77°F and is a Class IC, “flammable liquid” according to NFPA-30. Obviously, this opens a can of worms.
As for identifying the fragrance for design purposes, you’re usually left with finding an analogous organic that has similar active groups (e.g., -OH, -COOH, etc.) and similar chain length — remember it must be a liquid at storage temperature! In one case, I found a “linalool” as a stand-in for a lavender fragrance.
The fragrance storage will require an emergency vent and conservation vent, which can be combined as one. Typical pressure settings for atmospheric tanks are slightly below 0.5 psig, in compliance with OSHA 1910.106(a) (2). The vacuum rating, from air in-flow, is below 0.5 oz/in2, in compliance with API- 2513 “Evaporation Loss in the Petroleum Industry.” The emergency vent setting based on external fire is higher than the air out-flow conservation setting. One vent I sized required a 4-in. inlet (refer to: http://goo.gl/CWt93u). Be careful to draw the boundary line for the skid at the exit of the conservation vent/emergency vent — you have no idea what the customer’s vent looks like, so size for the lowest pressure possible to provide ample vent pressure drop.