Shell Side Pressure Drop 1

[thexder00]

I've been charged with finding the pressure drop across a two pass shell side TEMA F heat exchanger with process vapor as the shell side fluid. Each pass is individually baffled with double segmental baffles. I've been trying to use Kern's equations for the pressure drop calculations but I have my doubts that these equations were meant for a 2 pass shell side.

As such, should I be considering each side as a separate heater when using Kern? Are there any methods out there for estimating pressure loss across an exchanger of this type?

Operating Pressure = 2.2 psia
Vapor Flow = 186,500 lb/hr
Condensing Area per pass 75% / 25%

Thanks in advance for any advice.

 

[tkdhwjd]

Please indicate why you are calculating the pressure drop (what are you trying to find out with this pressure drop?).
For services in vacuum, hand calculations may not give you results with sufficient accuracy. Also, please note that significatn fraction of pressure drop may occur at the inlet/outlet nozzles. Why not use HTRI to simulate dP?

 

[SeanB]

I will go one step further and ask why you don't perform a pressure survey on the exchanger and find the actual pressure drop at the current flow rate?

 

[thexder00]

The purpose for the calculation is that there seems to be a 5" Hg pressure drop across the second shell side pass as measured at the current flow rates. The customer believes this to be a design issue (as a competitor convinced him it was), although the unit ran for 10 years at design pressure. I'm simply trying to prove that at design rates, the tube pitch and shell layout are sufficient for .5 psi pressure drop.

Unfortunately, I don't have acess to HTRI at the moment, otherwise I would gladly use it. Perhaps its time to convince the higher-ups to invest.

Any further advice would be greatly appreciated.

Thanks in advance.

 

[srfish]

The Kern method of calculating shell side pressure drop will give you results that are higher than the actual pressure drop.

To improve the accuracy you need a stream analysis method or a modified Bell-Delaware method. The Heat Exchanger Design Handbook has the lattter. A publication in the July 2004 Hydrocarbon Processing publication has an article that discusses shell-side pressure drop calculations.

In addition to calculating the cross flow bundle zones, end zones, baffle windows and nozzle pressure drops, the momentum loss needs to be calculated because of the vacuum.

 

[tkdhwjd]

I must admit that I have not analyzed shell side pressure drop with a modified Bell-Delaware method. But if you are dealing with 0.5 psi or smaller, I don't think any manual calculations will give you the results with sufficient level of accuracy as pressure drop depends on so many parameters/geometry.

 

[srfish]

Please explain "Condensing area per pass 75%/25%" and "each side as a seperate heater". The modified Bell-Delaware method section that is published in the Heat Exchanger Design Handbook (HEDH)is written by J. Taborek. He is responsible for most of the HTRI program. Since you mentioned in the beginning that you were "estimating pressure loss", this may be sufficient.

 

[thexder00]

There's a longitudinal baffle in the shell side that splits condensing tubes approximately 75% on one side and 25% on the other side. These make up the two passes. Vapor enters the larger of the two passes and after initial condensing, the remaining vapors into the smaller of the two passes.

Each pass is individually baffled. Since the longitudinal baffle is approximately 80% of the length of the shell, it divides the one shell into essentially two shells, or two heaters. Since Kern, and other methods only account for single pass shell sides, I was wondering if I should estimate pressure drop in each pass by calculating them as two separate heaters using flow information based on condensing in each pass? Also, is there a large factor of error in finding an equivalent diameter (as the two passes are obviously not round) and correlating the equations to a round tube in shell heater?

Thanks to all for your insight. I'm currently looking into the Bell Delaware Method and found that my company does have a copy of the Heat Exchanger Design Handbook, which has come in very handy.

Thanks in advance

 

[thexder00]

As a quick update, you guys were absolutely correct. I used Bell-Delaware Method for pressure drop of a shell side with vapor at vacuum conditions, and it came up with 22 psi pressure drop, which is absolutely impossible since in the field the vessel ran fine for many years with an unoticeable drop. I checked and rechecked the calculations and variables and they're all spot on. Also, the shell side heat transfer coefficient was appauling (115 W/m^2 K). I'm wondering if Bell-Delaware doesn't take into account that the vapor condenses on the shell side to transfer heat, since there is no Latent Heat of Vaporization used anywhere in the calculations.

Any ideas?

 

[srfish]

What two phase pressure drop method are you using ? If you are using the homogenous flow method, it gives high pressure drops at low operating pressures.

 

[thexder00]

I believe I'm using the homogenous flow method. Neither reference to the Bell-Delaware method (Heat Exchanger Design Handbook & Wolverine Tube Heat Transfer Data Book) mention anything about 2 phase flow. Only some assumptions about vapor. Is there another source out there that will take into account 2 phase flow and condensing vapors on the shell side?