How to Design a Heat Exchanger Without Knowing Tube and Shell Diameters

[pxp004]

Lets assume an exchanger that is 100% efficient (no heat loss). Duty (Q) and LMTD is known. Your overall heat transfer coefficient (U) is dependent on your transfer coefficients on the shell and tube side, and the resistance of the pipe walls (we're going to ignore entrance effects and fouling) . In order to calculate the transfer coefficients for the tube and shell sides, you need your Reynolds number in order to calculate Nusselt Number. The Reynolds Number is based on the diameter of your pipes and shell. So from a design perspective, if

Q = U*A*LMTD​

How would you calculate the surface area needed without knowing the specs (diameters) of your tubes and shell?
And without any software such as ASPEN, HTRI, etc. The only information I have available are flowrates, temperature data, and thermal properties of the fluids.

 

[SteamKing]

As a general rule, the designer must make some assumptions about the construction of the HE in order to proceed. In other words, you can sometimes use the flow rate of the coolant along with some rules of thumb about max. velocities inside a tube to come up with an estimate of the tube ID. There are only so many tube IDs and thickness schedules to choose from.

From that information, you can calculate how many tubes are required to handle the total flow of the coolant, and then calculate the other parameters of the HE from that. Check that the HE design can produce the duty. If it can't, then you start over with different assumptions. Rinse and repeat until everything converges.

Engineers used to do this without software, but it took a while longer for things to converge into a reasonable design. To speed things along, design charts for some of this effort were prepared and used.