you can use jackson`s relation,2002 to calculate the value of outer(ho) and inner(hi) heat transfer coefficient.Also you can use Dittus-Boelter relation whichever you can find suitable.I`am attaching a .doc file for your help.
Hi and Ho are the function of Reynolds number and prandtle number. Hence, based on the data you need to first find bulk mean temperature which will give you properties. You find the Prandtle number. based on the selected tube dimensions and flow rate, you can find Reynolds number. Finally you pick up the appropriate correlations, in the form of Nusselt number (Nu=hd/Kf), in the specified range of Re and Pr and calculate Hi or Ho.
Dittus-Boelter correlations is the most used in typical heating or cooling fluids without phase changing. In case of phase-changing fluids (condensation or vaporization) You have to use more specific correlations, like Shah or Klimenko. You can find a lot of literature on it.
I would like to add that the geometry of the heat exchanger (plate, straight tube, coiled tube ...) should be taken into account when searching for the heat transfer coefficients.
Depending on the nature of the fluid in the shell and tube sides, relevant parameters are to considered. The examples given in the Process Heat Transfer by Donald Q. Kern published by McGraw Hill can be followed.
It is a forced convection internal flow problem . depending on the type of flow Laminar or Turbulent you have to pick up Nussult Number correlation and solve for h.Use of hydraulic mean diameter,selection of properties at bulk mean temperature,consideration of viscous effects,Entrance length etc.. are critical issues.Pl go through Chapter 14 worked out example no 4 from the book "Heat Transfer A conceptual approach" written by Prof.P.K.Sarma of New age international publishers India.Or Mail to me i will forward the detailed procedure. [email protected]
Despite the large number of correlations proposed in the literature, it is difficult, in some cases, to find the appropriate correlation for calculating heat transfer coefficients. For example the case of a solid melting on the exterior surface of a coil, which is generally applied in the sulphuric acid manufacturing.
Mr. Mahesh Kumar, it is difficult to give a specific answer to your question unless you provide more information.
In a concentric tube heat exchanger, if your tube is smooth and the flow is fully developed and turbulent, hi can be got from available well-established correlations. Then it is a small matter to extract ho from an energy balance.
If you wish to develop a predictive correlation, you may use the well-known Wilson plot technique (see: The tube side heat transfer coefficient for enhanced double tube by Wilson plot analysis, Journal of Applied Science , 2011, V. 11, pp 1725-1732.)
There are situations where the inner geometry is not so simple. In such cases you can still use the Wilson technique (see: Double tube heat exchanger with novel enhancement: part II—single phase convective heat transfer, Heat Mass Transfer, (2012) 48, 1451-1462.)
Even 'well-established' correlations have an associated uncertainty. Therefore the calculated ho will have an uncertainty too, larger than for hi. Use of GA and ANN have been used to get the best pair of hi and ho. They can also be used when both values, hi and ho, are unknown and you have a single equation (heat balance). See, for example:
Simultaneous determination of in- and over-tube heat transfer correlations in heat exchangers by global regression, Arturo Pacheco-Vega, Mihir Sen, K.T. Yang, International Journal of Heat and Mass Transfer 01/2003; 46(6):1029-1040.
Like M. Raghavan, I think that in some cases where appropriate correlations are not available, the best way to determine heat transfer coefficients is the use of Genetic algorithms or other inverse thechniques. The problem is when we need to size a heat exchanger and when the heating surface is not yet known. In this case, we can determine the surface simultaneously with hi and ho.