Energy System Sensitivity Parameter

Abstract

In this paper, sensitivity analysis of energy systems was developed. This analysis was performed by using a quantity called sensitivity parameter, suitable for this purpose. The parameter establishes the relative influence of input variables on the calculation results. The sensitivity parameter evaluation was carried out by using a numerical procedure. This procedure is very convenient because it permits the employment of computational commercial codes. A Brayton cycle gas turbine power plant was the considered energy system. Sensitivity analysis was developed relatively to conversion efficiency and power output delivered by the system. Three gas turbine typologies were examined: heavy duty, small size and aeroderivative gas turbines. For these gas turbine typologies, in steady state working conditions, the verification case was analyzed and the corresponding input variables assumed. It resulted that the sensitivity parameter values relative to performance are very different according to the input variable considered. Furthermore, the sensitivity parameter values, for an assigned variable, are different for each of the three examined typologies of gas turbines. It was verified that percentage power output and conversion efficiency variations versus percentage variation of each variable is substantially linear for all the variables analyzed. This permits the employment of the sensitivity parameter for evaluating the percentage performance variations for assigned percentage variations of input variables. It also resulted that many of the considered variables have the same sensitivity parameter value if it is relative to power or conversion efficiency. This is important because it simplifies the sensitivity parameter calculation, particularly for the complex energy systems. Moreover, the sensitivity parameter trend versus the input variable value, is substantially the same for the three gas turbine typologies considered. In conclusion, the sensitivity parameter evaluation is interesting since it establishes how much the numerical value of a quantity resulting from the calculation is affected by a quantity assumed as input and by its numerical value.