6 replies to this topic

[daraj]

Hi,in conceptual process design(using simulator software) often we are advised to look for certain design paramters from the TEMA sheet generated to ensure the software has done a reliable design. While we look to check on clean, dirty, service coefficient, another parameter  is the pressure drops and velocities.

There is a situation where I have a converged heat exchanger model/design where everything looks good, but the velocities. The reboiler load is guhe(duty>20000kw) and hence no matter what i change i cannot seem to reduce velocities.

 

i have a couple of specific questions to ask

 

1. There are guidelines for shell and tube side velocities. But often these guidelines say, if it is a liquid it has to be say,2-5 m/s and if it is gas it has to be  below 15 m/s or something like that. But most often we dont have one single neat phase on either side. Fluids in both shell and tube sides are often undergoing phase change. while on one side it is condensing steam and on the other side it will be process fluid reboiling and changing to vapor. so how do you apply these guidelines here? these guidelines assume one phase

 

2. since this is preliminary design I dont usually go for exotic TEMA types. I stick with what is known and common like say BEM or BEU or BKT and so on. The problem is, in this case, I end up having high shell and tube side velocities. And only when I have almost 5 or 6 shells in parallel am I able to bring the velocities to 5 m/s and 20 m/s on tube and shell side respectively. But having several shells in parallel also ends up in an increaswed area/unit and thereby cost. Is this always going to be a tradeoff?

without changing TEMA type what other handles I have to keep velocities to the recomemnded levels?

 

(the pressure drop calculated vs allowed in the TEMA sheet seems to be within limits, i.e the calculated prerssure drop is well within default allowed pressure drop) 

[daraj]

if i were to rephtrase question 2 , what I want to know is as long as my calculated pressure drop is less than allowable pressure drop as reported by TEMA sheet in simulation software is itOK if I had high tube side and shell side velocities(i.e between 40 and 50 m/sec, say). or are there other issues besides pressure drop such as say erosion that place absolute uper limits on tubeand shell velocities.? interestingly the simulation softare(aspen edr) did not give any warnings on the high velocities obtained and ddi not advise me to look for alternate design. lonly thing mentioned in warnings section was about a possible vibration/resonance problem

[pavanayi]

Simulation software will try to optimize the design within the margins of each parameter that has been set. 

It will search for shell dia from x in to y in, tube lengths from x to y, velocities from x to y, pressure drop from x to y, tube passes from x to y etc etc. If it finds a solution that is within all these constraints, the software will show the calculation as converged.

The default values don't make sense always. Check all the numbers in the program options->design options to see if the limits for each parameter are valid for your case. As an example, The software might go upto 98in of shell dia, by default but you would want to keep the max value to a more reasonable number.

 

Also, before starting, ideally you should have an idea of what configuration you are looking for. This can come from past experience as well as looking at datasheets of similar exchangers. This will give you a good starting point. 

 

Indeed when the total heat transferred is >20MW, I will not expect the solution to be one shell or two shells. 

[Pilesar]

If you do not meet your velocity guidelines, consider the consequences. High velocities can cause erosion and contribute to vibration which can lead to reliability problems. High noise can be an issue. Condensing steam can start with very high velocity in the tubes and cause rapid wearing. There are ways to combat erosion: consider using a protection plate on the tubesheet or ferrules to protect the tube ends. Also consider that stainless tubes and tubesheet material is harder than carbon steel. Sometimes using multiple shells is just the best solution.

[daraj]

Hi pavanyi and pilesar,

 

Thanks.  I must emphasize that what iam doing is a mostly conceptual design exercise(say, paper exercise) with the aim of feeding into a preliminary cost estimation software (class 4 type assessment) so rigorous design is not needed.But at the same time i do not want my design to depart far from reality either. 

 

coming to veolicities, the software does not warn me on erosion/corrosion issues despite the high velocities. It only gav e a warning about a possible vibration/elastic/resonance issue. So I do not have a feel for what velocity is considered really high for my application and so on.

The calculated pressure drop is below allowed pressure drop though.The guidelinesare often for gas orliquid phase.It is frustrating to see that the guidelines dont take into accont constant phase changes happening on both sides of the exchanger.Maybe I should take the phase condition at the inlet as the actual phase and not worry about what happens later on in the course of heat exchange, in order to use these guielines?

 

I have one more update. when i changed the shell to "unbaffled" from "single segmental baffle" option, I managed to reduce velocity quite a bit without compromising too much on other parameters(like heat transfer coeficient). But is an unbaffled shell OK in reality? Having an unbaffled shell might reduce the number of shells I need in parallel to accomplish the heat duty.

I also want to know, how commin it it in inudtryto have s&t exchangers with multiple shells in series or parallel. is it always a good practice to try as much as possible to accomplish the heat transfer in one shell(provided area is less than say 1800-2000m2)? does having multiple shells in parallel pose any problem with respect to maintanance and turndown?

[Padmakar Katre]

Hi,

 

If the assignment is quite preliminary, then you need only area which you can estimate based on typical overall heat transfer coefficient, MTD (based on heat release profile from simulator) and duty required. Only area, design P,T and Material will go into costing. Detailed design you can pursue later on also.

TEMA type selection is not so straight forward and hardly required to be correct during preliminary cost estimate assignment. Thermal design is mainly an outcome of detail engineering activity.

[daraj]

Hi padmakar and others, I agree soem of it is detailed engineering. But wven while conceptual design using software i can still generate TEMA sheet that meets basic requirements sufficiently. But where Iam struggling right now, Iam using a BEM type condenser, horizontal for a low pressure column(<100mbar) and ending up with either large area and/or high tube side velocities. Ihave process fluid on tube side. I dont think changing sides matter it only shifts the problem. if it was shell side velocity/pressure drop i can try increasing shell dia or have double segmented baffles to reduce the pressure drops a bit. but what would you do to decrease the tube side velocities of process vapor, without compromising on other paramteres. This is mostly from simulation point of view. I do have converged result but finding an acceptable designseems to be difficult.