Page 416 - Mechanical Engineers' Handbook (Volume 4)
P. 416

3 Thermal Control Techniques  405


                                                           Y     21.e                           (92)
                                                            in
                                                                  11.e
                              If the tip of the most remote fin is not adiabatic, the heat flow to temperature excess
                           ratio at the tip, which is designated as  ,

                                                                 q a                            (93)
                                                                  a
                           will be known. For example, for a fin dissipating to the environment through its tip desig-
                           nated by the subscript a:
                                                               hA                               (94)
                                                                  a
                           In this case, Y may be obtained through successive use of what is termed the reflection
                                      in
                           relationship (actually a bilinear transformation):
                                                                      (q /  )
                                                  Y        21,k 1  22,k 1  a  a                 (95)
                                                   in,k 1
                                                            11,k 1      12,k 1 (q /  )
                                                                       a
                                                                          a
                           The Cluster Algorithm. For n fins in cluster, as shown in Fig. 15b, the equivalent thermal
                           transmission ratio will be the sum of the individual fin input admittances:
                                                            Y                                   (96)
                                                          n
                                                                   n
                                                                     q
                                                                      b
                                                      e
                                                          k 1  in,k  k 1    b  k
                              Here, Y in,k  can be determined for each individual fin via Eq. (93) if the fin has an
                           adiabatic tip or via Eq. (95) if the tip is not adiabatic. It is obvious that this holds if subarrays
                           containing more than one fin are in cluster.
                           The Parallel Algorithm. For n fins in parallel, as shown in Fig. 15c, an equivalent thermal
                           admittance matrix [Y] can be obtained from the sum of the individual thermal admittance
                                            e
                           matrices:
                                                         [Y]      [Y]                           (97)
                                                                n
                                                           e
                                                               k 1  k
                           where the individual thermal admittance matrices can be obtained from
                                           [Y]    y 11  y 12         
     11    1 22  
   22      1 21


                                                                           12
                                                                   12
                                                                                12
                                                               12
                                                 y 21  y 22   1           1                     (98)
                                                                12    12    12    22
                           If necessary, [ ] may be obtained from [Y] using
                                                 [ ]      11    12         
   y 22  y 1 11
                                                                     y
                                                                          21
                                                                      21
                                                          21    22   
  y                       (99)
                                                                     y 21  y 21
                           where 
   y y    y y
                                             12 21
                                      11 22
                                 Y
   411   412   413   414   415   416   417   418   419   420   421