Manfred Georg Kratzenberg1, Hans Georg Beyer2, Sergio Colle1, Dirk Petzoldt1,2
1Solar Energy Laboratory Federal University of Santa Catarina, Department of Mechanical

Engineering, Florianopolis, Brazil

1 Department of Electrical Engineering, University of Applied Science Magdeburg-Stendal,

Magdeburg, Germany

1. INTRODUCTION

Testing the efficiency of solar collectors is a basic pre-requisite to obtain performance characteristics for thermal solar collectors.

In order to quantify those characteristics, we must run a test to determine the efficiency curve coefficients of those solar collectors.

In an international context, standards for the respective test procedures are given by EN 12975 [ 1 ] and ISO 9806 [ 2 ]. In Brazil the standard is given by the NBR 10184. Only the EN 12975 provides different test procedures, the Steady-State Test (SST) and the Quasi-Dynamic Test (QDT). The QDT has the advantage of allowing the execution of more collector tests within the same time period, with the same test equipment and the same test facility as compared to the steady state collector test [ 9 ]. On the other hand the quasi­dynamic test requires a somewhat more demanding effort for the calculation of the collector coefficients.

There has already been made a comparison between SST and QDT [ 7 ], but there have been no statements about the uncertainties. The purpose of the present work is to discuss and evaluate of test procedures in view of the uncertainty of the determination of the collector coefficients for the QDT. The Least Square (LS) and the Weighted Least Square (WLS) regression methods applied in the trend setting QDT are presented in this paper. Uncertainties of the LS and WLS results are calculated. A methodology for the verification of real confidence limit of the LS and WLS uncertainties is presented.