| description abstract | In the volatile fuel price environment of today, the quest for alternative fuels has become a heavy and long term trend in power generation worldwide. Incorporating alternative fuels in gas turbine installations raises multiple engineering questions relating to combustion, emissions, on-base and auxiliary hardware capability, safety, etc. In 2008, GE carried out a field test aimed at characterizing the combustion of ethanol in a naphtha fuelled gas turbine plant. The testing strategy has been to locally prepare and burn ethanol-naphtha blends with a fraction of ethanol increasing from 0% to nearly 100%. During the engineering phase prior to this field test, it appeared necessary to develop a sufficient knowledge on the behavior of ethanol-hydrocarbon blends in order to establish the safety analysis and address in particular the risks of (i) potential uncontrolled ignition event in the air blanket of fuel tanks and (ii) flash vaporization of potential fuel pond in a confined environment. Although some results exist in the car engine literature for ethanol-gasoline blends, it was necessary to take into account the specificities of gas turbine applications, namely, (i) the much greater potential ethanol concentration range (from 0% to 100%) and (ii) the vast composition spectrum of naphtha likely to generate a much larger Reid vapor pressure envelope as compared with automotive applications. In order to fulfill the safety needs of this field test, the “Laboratoire de Thermodynamique des Milieux Polyphasés” of Nancy, France has developed a thermodynamic model to approach the vaporization equilibria of ethanol-hydrocarbons mixtures with variable ethanol strength and naphtha composition. This model, named PPR78, is based on the 1978 Peng–Robinson equation of state and allows the estimation of the thermodynamic properties of a multicomponent mixture made of ethanol and naphtha compounds by using the group contribution concept. The saturation equilibrium partial pressure of such fluids in the various situations of relevance for the safety analysis can thus be calculated. The paper reports the elaboration of this model and illustrates the results obtained when using it in different safety configurations. | |
