O. Kusyy*, K. Vajen, U. Jordan

Kassel University, Institute of Thermal Engineering, Kassel, Germany
* Corresponding Author, solar@uni-kassel. de
Abstract

In this paper the application of sensitivity analysis to the investigation of solar water heating systems is considered. Two global sensitivity analysis methods are described and applied to different solar heating systems. The first one is the Morris method that only ranks parameters by importance and the second one is the Fourier amplitude sensitivity test (FAST) that quantifies the influence of the parameters on the target functions. The both methods were implemented into the GenOpt (Generic Optimization) software and coupled with the TRNSYS simulation program.

Keywords: Sensitivity analysis, Fourier amplitude sensitivity test, Morris method

1. Introduction

In recent years, many middle to large solar heating systems were installed all around Europe and especially in Germany. A proper design of such systems is decisive for their functionality. Underdimensioning or poor selection of design parameters as well as the control strategy could lead to an overall poor efficiency of the systems. During the designing process, the advanced numerical optimization methods should be used in order to find the optimal parameter values that provide the best efficiency of the system. Considering that the target functions depend on a high number of optimization parameters and that the global optimization algorithms require large number of system simulations, the task of optimization turns to be very computationally expensive. In order to decrease the number of optimization parameters and, thus, make the optimization faster, the sensitivity analysis of parameters could be used prior to optimization. Only the most influential parameters are then selected for optimization. Another straightforward application of the sensitivity analysis is analysis of uncertainties, that is, how uncertainties in parameters influence the uncertainty of the target function. Here two sensitivity analysis methods are described and applied to the analysis of two solar heating systems. For the first system the influence of the operation parameters on the cost function is investigated by the qualitative Morris method. A more comprehensive Fourier amplitude sensitivity test is applied to the investigation of the influence of the design parameters on the solar fractional savings function of the second system. In this paper only the exemplary examples of applications of the both methods are considered. The methodological application of the methods is planned but not yet realized.