The main task of baffles is to direct the flow in shell and tube heat exchangers to increase the effectiveness, but also designed to support the tube bundles.

@giuma so your conclusion is "baffle plate = support plate" ?

Not really, baffles direct the fluid in the shell between the inlet and outlet ports such that the fluid expose to the bundle surface area in large extent that would improve the heat transfer between the surface area and the fluid in the shell.

@Erma,

A number of clarifications are needed from you before the context of your question can be understood. (1) What application are you talking about : refrigerant condensers or power plant condensers or regenerative feedwater heaters in power plants? The construction in each of these is different. Only small power plant condensers have a cylindrical shell. Even then, the flow of the condensed medium is in simple crossflow on the shell side. The condensers in power plants of large output (for example, 1000 MW) are necessarily of the church-window type design. In water-cooled refrigerant condensers the shell-and-tube type is used, in which the coolant may be on the shell-side or the tube-side.

(2) What fluid is on the shell (i.e., baffle) side? The two-phase condensing medium or the single-phase coolant?

As Giuma says, baffles will be required to direct the flow and support the tubes in an E-type shell, while in an X-type shell (only crossflow), they are only for supporting the tubes. In all these cases, the baffles themselves are vertical; only the baffle cut can be horizontal (up-and-down flow) or vertical (side-to-side flow). You can look up the following thesis by Googling it. Section 3 gives some details.

http://elib.uni-stuttgart.de/opus/volltexte/2011/6698/pdf/K._Mohammadi_Dissertation_Online_Publishing.pdf

In regenerative feedwater preheaters, the shell can on occasion have a longitudinal baffle (the shell is called an F shell here). The longitudinal baffle is horizontal (with the axis of the cylindrical shell being horizontal). In E-type shells, the only chance for baffle surfaces to be horizontal is if the shell in mounted vertically (a rather unusual orientation that I have not come across).

So, my question boils down to this : Does "horizontal" in your question refer to the baffle cut or baffle surface? Then one may be able to give a considered answer. 

@vijay, I make some analysis for steam power plant (100MW) with condenser type is surface condenser, two pass , reverse flow. There is a sentence, " By dividing the hang up bell type tube bundle, and installing the baffle plate nearly horizontal, the steam flow is changed to a horizontal flow with high condensation factor", could you explain to me what the meaning of that sentence?

@Erma, I too was intrigued by the slightly inclined HORIZONTAL baffle.

After giving it some thought, my take is as follows:

In condensation over a bundle of tubes, the condensate drains from the upper tubes to the lower tubes. With the additional condensate load, the condensate film gets thicker. Since the main resistance to condensation heat transfer resides in the liquid film, a higher temperature difference (Tsat – Twall)  is required to drive the heat flow through. Since the wall temperature is fixed by the coolant temperature, the only way for a higher ΔT is for the condensing pressure to rise, which implies a lower pressure differential across the turbine (and therefore less work).

If the liquid film is thin, the saturation temperature will be lower and therefore also the saturation pressure.

To keep the liquid film thin, one way would be to simply capture and drain out the condensate at intermediate levels. This can be achieved by interposing a perforated baffle, say between the tubes of the two passes. The perforations allow the uncondensed steam to flow on downwards, while the slight inclination (even one degree inclination to the horizontal would do) causes the liquid falling on the baffle to drain sideways and be taken out of the tube bundle.

In your case, “the steam flow is changed to a horizontal flow with high condensation factor”. Obviously the steam does not flow downwards, but horizontally across the tubes so that the baffle need not be perforated.

To me, it appears that this is not a favoured arrangement in the present day as tubes of the type that Wolverine Company manufactures are very efficient in draining the condensate and enhancing the heat transfer. You may Google Wolverine Tube to get more details on the profiles available.

I would have liked to add some figures. As a retired professor, I have the time but lack the energy to draw sketches, scan them and include them as attachments. Therefore, I hope that my verbal explanation is clear to you without sketches.

Horizontal baffles (1) desrease the velocity of steam/water mixture and, hence, reduce erosion of condenser tubes; (2) distribute steam/water mixture along the whole length of condenser tubes, intensifying heat exchange. They also direct condensate to the sides of condenser and, hence, reducing water film on condenser tubes and reduce aerodynamic resistance of condenser.

@vijay Thank you so much Mr Vijay, after I read the Mr. K Mohammadi, I became rather understand about "horizontal baffle", thank you for your kindness :)

@Dmitri thank you :)