Analysis of the collector test procedures for steady-state and quasi-dynamic test conditions in view of the collector coefficients uncertainties and model stability

Abstract

Outdoor collector tests are inherently performed under variable weather conditions. Whereas for the steady-state collector test SST, ISO 9806 strong restrictions are set for the weather conditions for usable data sets or samples, the ambient conditions of the quasi-dynamic collector test QDT, EN12975 are allowed to be more variable. This results in shorter collector test time, but could have drawbacks for the uncertainties caused by the reproducibility of the test results, i.e. the collector coefficients (or also called collector parameters) stability of the collector model, as well as for the estimated power and energy with this model. As the weather conditions are never the same within several tests, outdoor collector tests are not repeatable but reproducible. Like the QDT permits to use data with more variable weather conditions, it is thus may be expected, that the uncertainties of the collector parameters gained by a QDT test are superior to those from the SST test. On the other hand the model of the SST is only a reduced collector model. All optical and thermodynamic effects that appear during the application of a solar collector are not managed with that reduced model. Under this consideration it is possible, that the result of the SST collector test may estimate the produced energy with more uncertainty than the QDT. We estimate in this paper the total uncertainty and the stability of the quasi-dynamic and the steady state test methods with the objective to proof which of methods is the most reliably one. We evaluate the collector parameters and their uncertainties of a covered collector using both, the SST and QDT test methods. As basis, a large data set from 3 months of operation is applied. This set is then separated in various single data sets fulfilling either the conditions of a complete steady-state or a complete quasi-dynamic test. Hereby several sets for the case of the QDT, and one for the steady-state test one set could be identified. From each of the tests the collector parameters and their uncertainties are calculated. This allows the comparison of both, the model coefficients and their uncertainties. It is than tested with statistical methods to what extend the reduced SST model is sufficient to extract collector coefficients that are usable for the calculation of the long term energy gain of the collectors. We use as a second result statistical procedures to test whether the coefficients extracted from the QDT data sets of each of the QDT collector tests have statistical equality within 95% of confidence if we compare the same coefficients from different tests. Proofing statistical equality is in coherence with model stability of a collector model. The Energy production simulated using the SST and QDT models are compared with the measured energy during the time period of 2 month. Finally the total uncertainties for long term energy estimation of the SST and the QDT tests are quantified.