Page 87 - Advanced Gas Turbine Cycles
P. 87

Chapter 4.  Cycle eficiency with turbine cooling (cooling flow rates specified)   63

       B (coolant air)

            (To)xc = (Td2C + Q/rb.cF,   @o>xc  % (po)tc(1 - mzc/2).        (4.46)
       We can then proceed to determine the changes across control surface C. The final total
       temperature (To)sm has  already been obtained but  the  total pressure (po)*,,, has to be
       determined. An expression given by Hartsel [ 141 for the mainstream total pressure loss in
       this adiabatic mixing process again goes back to the simple one-dimensional momentum
       analysis given by Shapiro [ 131 for the flow through control surface C illustrated in Fig. 4.8.
       Hartsel  developed  Shapiro's  table  of  influence coefficients to  allow  for  a  difference
       between  the  total  temperature of  the  injected flow  (now  (To)xc) and  the  mainstream
       (To)xg):

            APoIPo = 11 - @o)sm~(Po)xg)l
                  = -(+YM;,m(  I  + [(To)xc~(~o)x,)l - 2Y  cos 41.         (4.47)

       Here y  is  the  ratio  of  the  velocity  of  the  injected coolant to  that  of  the  free  stream
       0, = V,/V,), Mx, the Mach number of the free stream and 4 the angle at which the cooling
       air enters the mainstream (Fig. 4.8).
          The value of y has to be determined; an approximation suggested by Hartsel is to take
             = @o)xg,  so that Vc/Vg = [(T&c/(To)x,)]'n, since the static pressures must be the
       same where the coolant enters. A sufficient approximation might be to take (To)xg as the
       exit temperature from the combustion chamber and (To)xc as the exit temperature from
       the compressor (Le. again ignoring Q in Eqs. (4.45) and (4.46)).
          A  more  sophisticated  approach  would  not  only  take  account  of  Eqs.  (4.45) and
       (4.46) to give the  two  stagnation temperatures at exit from control  surfaces A  and B,
       but  it would  also not  assume the total  pressures of  coolant and  mainstream to be  the
       same. For the first nozzle guide vane row these can be derived by accounting for losses
       as follows:
        (i)  in the mainstream (g), the stagnation pressure at delivery from the compressor less
            ApKc  in  the combustion process, and Apo in  the nozzle row itself (as in control
            surface A,  due to friction and the heat transfer away from the mainstream gas if
            included);
        (ii)  in  the  coolunr  air  stream  (c), the  stagnation  pressure  at  extraction  from  the
           compressor less a  loss  ApoD (in  the  ducting and  disks before coolant enters the
           blade itself), and Apes  (in the blading heat transfer process in  control surface B
           due to both friction and heat  transfer, if  included).
         The total pressures at X may thus be determined, as (po)xg and (P~)~,. If, as Hartsel
       implies, the mainstream Mach number at X (Mxg) is also known, which means that the
       static pressure at the mixing plane ( px) is also known, Mxc may also be determined from
       (po)x,. The two different velocities V, and V, are then obtained, together with the required
       value of y  for Eq. (4.47).
         But there is a further subtle point here in determining y, as implied by  Young and
       Wilcock. With [( p~)~Jp~] known, not only is the Mach number Mx, known but also the
       non-dimensional mass flow, { ~R(To)xc]''2/Axc(po)xc }, may be obtained. This means that
   82   83   84   85   86   87   88   89   90   91   92