| OCR Text |
Show 2.1.14 The boundary condition (11) becomes Bi(eg-es) • 2°2<e»-e;) = (es-eN) F-(FO,1) (30) by using the temperature distribution (25), where k=N denotes the isotherm closest to the surface. The dimensionless surface temperature can be iterated by this formula. If 6 >6N+1/ a n ew isotherm is generated on the particle surface (and the index N is increased by one). The temperature gradient on the particle sucface is exp{-f (Fo)g (1)} FN'1(*F o,'1 ') = f„(Fo) (31) when f (Fo)^0 and 1-exp{-fN(Fo)gN(1)} F N ( F O ' 1 ) = gTTTT (32) 'N when f(Fo)=0. It has been seen that the calculation procedure becomes unstable when AFo is too large. The stability criteria N-1 AFo < p.min{Xk+1-Xk)2 (33) k=K has been found experimentally. The value p=0,20 is used in the computer program so that the time step changes during the calculation) smaller AFo is used when the temperature gradients are large (the distance between the successive isotherms X, and X, - is small) . This calculation method is also applicable to other simpler problems in heat transfer - pure transient conduction heat transfer problems with temperature dependent material properties - drying at intense surface heat flux conditions - transient freezing of melting problems The computational advantages of the method are - the temperature (and density) dependence of the material is included in the constants Ck and Hk (and Bi and otaR/X ) that are calculated before the actual computation begins, and no |
