In this issue:
1. Project Spotlights:
* Batch Process Evaluation and Scale-up Process Design
* Refinery FCC Fractionator Performance Improvement Study
2. Technical Discussion: Relief Load Estimate for Liquids Boiling Near Critical Temperature
3. Safety Pause: “Tank Blowed Up”
Batch Process Evaluation and Scale-up Process Design
PROCESS evaluated a batch process for a new type of polymer-based encapsulated product and provided engineering support to scale up the pilot scale process to commercial scale reactors. Anomalies had been found between the results from pilot batch data and a few large scale trial batch runs from the production facility. So, first PROCESS had to evaluate existing data to determine which process operating parameters were critical to achieving acceptable product quality. The next step in the evaluation was… Read More
Refinery FCC Fractionator Performance Improvement Study
A major petroleum refinery contacted PROCESS to review operating and performance data from a fully packed refinery fluid catalytic cracking (FCC) main fractionator that was exhibiting very poor separation between its three liquid products. There were large differences in temperatures reported in different quadrants of the tower and they would flip-flop several times a month and propagate through multiple beds in the tower! PROCESS developed a HYSYS model to match the FCC performance data and began reviewing tower drawings and capabilities. Two highly loaded draw trays were found suspect.
Read more here to learn how PROCESS further analyzed the FCC and made recommendations made to improve performance.
Calculating Relief Load for Liquids Near Critical Temperature
When a liquid in a process vessel experiences excessive heat input, either due to fire, heat exchange or some other heat source, overpressure of the vessel is possible due to vapor generation from liquid boiling. The current common practice for determining the required vapor relief load for this situation is to divide the heat input by the heat of vaporization:
For the great majority of cases, this provides a reasonably conservative estimate of the vapor flowrate which needs to be relieved to prevent vessel overpressure. However, when relief conditions approach the thermodynamic critical temperature, the latent heat of vaporization rapidly decreases to zero. At boiling conditions near the critical temperature, use of the conventional method, as given in equation 1, results in relatively large relief rates.
Relief rates at near critical temperature using equation 1 are unreasonably conservative. That’s because equation 1 is only an approximation to the more general equation:
Equation 2 is the more general relationship. For conditions far from critical, the density ratio is usually much less than one and the ratio correction factor can be ignored. But for conditions near critical, the vapor and liquid densities approach each other and the correction factor becomes significant. Note also, that equation 2 should also be used for systems far from critical when the vapor density can be relatively large. This could occur in high pressure, low temperature, and high molecular weight systems.
What is the rationale behind equation 2? Leung (AIChE Journal, October, 1986, P. 1624) derived a similar relationship using a dynamic heat and material balance for the two-phase boiling system. The analysis is somewhat detailed but it can be explained quite simply.
The basis for all relief load estimates is that the volume to be relieved must equal the net volume generated. In a boiling, two-phase system, two volumes exist: the vapor volume and the liquid volume. It is important to emphasize that when a liquid boils, vapor volume is generated but liquid volume is depleted. That liquid volume can be taken up by part of the vapor without any generation of pressure. This is seen from a volume balance:
Volume to be Relieved = Vapor Volume Generated – Liquid Volume Lost
Let’s say 100 lbs of liquid is vaporized. If the vapor density is 0.1 lb/ft3, then 1000 ft3 of vapor is formed. If the liquid density is 50 lb/ft3, only 2 ft3 of liquid is depleted. For this case, the liquid volume depleted is very small compared to the vapor volume generated and can be ignored, or equation 1 applies. But near the critical point, the volume lost by the liquid can be significant compared to the volume generated from the vapor since the densities are more comparable. For this case, if the vapor density is 10 lb/ft3, then 10 ft3 of vapor is generated. And if the liquid density is 20 lb/ft3, then 5 ft3 of volume is depleted. That means 5 ft3 of depleted liquid volume is available to 5 ft3 of vapor volume and only 5 ft3 of vapor volume needs to be relieved. Mathematically:
The following table for some example molecules shows the potential reductions in estimated relief load when using the more general equation 2:
Temp | Sat. Press | Heat of Vaporization | Reduction Factor |
(°F) | (Psia) | (Btu / Lb) | |
Butane | Tcritical = | 305 °F | |
224 | 253 | 106 | 0.90 |
242 | 304 | 96 | 0.87 |
260 | 362 | 84 | 0.82 |
278 | 429 | 68 | 0.75 |
296 | 504 | 44 | 0.59 |
Decane | Tcritical = | 652 °F | |
476 | 69 | 96 | 0.96 |
512 | 98 | 88 | 0.95 |
548 | 135 | 78 | 0.92 |
584 | 181 | 66 | 0.85 |
620 | 240 | 49 | 0.72 |
Ethane | Tcritical = | 90 °F | |
8 | 248 | 149 | 0.93 |
26 | 321 | 135 | 0.90 |
44 | 407 | 119 | 0.85 |
62 | 510 | 98 | 0.78 |
80 | 631 | 65 | 0.62 |
As the critical temperature is approached, significant reductions in the calculated relief load from the conventional estimate, equation 1, results. This is shown by looking at the “Reduction Factor” in the last column. For example, for butane at 296 °F, use of the more general equation 2 results in an estimated load 0.59 times less than that estimated from equation 1. The approximate equation 1 is overly conservative as the critical temperature is approached. This could also be true for systems not near critical temperature but having high pressure, low temperature, and high molecular weight components. Subsequent reduction in piping pressure drop is even more significant since the pressure drop will be reduced by the reduction factor squared.
The use of the more general, and less conservative, equation 2 can result in significant savings in project capital cost.
Take a Pause for Safety
Below is a true story and it’s the kind of project that most in the chemical industry have dealt with. It’s just a plant expansion where a new line (pipe) is added to an existing tank…….
PROCESS’ involvement with this story was limited to technical analysis after the incident. We are sharing it with our readers in hope to bring awareness as to how important it is to understand not only the actual math behind a calculation, but also the basis for using a particular equation. Can you figure out what went wrong before the end of the story?
“Tank Blowed Up”
A facility has a solvent recovery tank that is critical to operation but it needs a new line added to it for an expansion project. This tank holds organics and water. Actually, make that a non-polar solvent and water. The plant would really like to keep the tank in operation safely while welding in the new line. So, they contact an engineering company (not named PROCESS) to calculate if it would be safe to weld on a solvent recovery tank while in operation.
The engineering company got the composition in the tank and calculated the organic vapor in the head space of the tank. They used Raoult’s, Law given as:
The engineering company determined that the vapor fraction of the solvent was below the Lower Explosive Limit (LEL) and informed the plant that it would be safe to weld on the solvent recovery tank while the tank was in operation.
Joe (not his real name) the Welder set out to weld the pipe addition to the tank. Later that day Joe went into his boss’s office and filled out the voluntary termination form and left. The form read, “I quit. Tank blowed up”. Well it took a few months to find Joe, but when he was found this in essence is what he said:
I was welding on the pipe, several feet down from the tank when I heard something like a giant bottle rocket. I looked over and the tank wasn’t there. It was arching up over my head and then it crashed into the power feeds to the plant. That’s when I decided to quit.
Luckily, Joe the Welder was not hurt. But, did you catch what happened?
The engineering company made the mistake of using Raoult’s Law for ideal solutions which underestimated the solvent vapor fraction in the head space of the tank. Raoult’s law states that the vapor fraction of each component of an ideal mixture of liquids is equal to the partial pressure of the pure component multiplied by its mole fraction in the mixture. When the solution is non-ideal, as in this case, there is a deviation from Raoult’s Law which should be accounted for by using an activity coefficient. The equation should be modified:
PROCESS’ calculation during the post-accident analysis showed that if Raoult’s Law had been properly adjusted with the appropriate activity coefficient to account for the liquid phase non-ideal behavior, the proper conclusion would have been made that the vapor pressure of the organic was above the LEL and it was not safe to weld on the tank while in operation.
Did you hear “non-polar solvent” and “Raoult’s law” and get a little squirmy? Did you hear “like to keep the tank in operation while welding” and wince? If you did, then you already know that those would be good times to take a pause for safety. Let’s review some lessons learned from this unfortunate event.
- Understand the basis for the calculation and equation that is being used. In this case an equation was used that was meant for an ideal liquid mixture when in fact there was a deviation from ideal behavior since the solvent was non-polar and mixed with water (polar).
- Have calculations (and assumptions thereupon) independently checked and verified.
- Do not let time or production pressure cloud engineering judgement.
- Always take extreme caution when performing hot work near solvents. A good engineering practice would be to not weld on equipment while it’s in operation – regardless of what the calculation says!
PROCESS wishes all of our readers a Merry Christmas and a Happy New Year!
About Us
PROCESS offers a full range of Process Design and Process Safety Services to clients around the globe. Our services include conceptual process design; simulation and modeling feasibility studies; FEL-0,1,2, and 3 (Schedule-A process design packages); front-end-engineering design (FEED); debottlenecking; process optimization; on-site operations support; process safety training; safety program development; on-site process safety auditing, performance of hazard assessments; and other services designed to meet the needs of our clients.
Contact Us at:
Process Engineering Associates, LLC
700 South Illinois Ave Suite A-202
Oak Ridge, TN 37830
info@processengr.com
Call Us: (865) 220-8722