The present investigation assesses the procedures and conditions for parabolic trough solar field performance testing. It aims at criteria and guidance for solar field acceptance tests that lead to representative performance parameters on the basis of which its future (annual) yield can be predicted. It comprises the specification of test requirements in terms of measurement equipment and test conditions in view of an identification of parameters with minimum uncertainty.
A new parameterised dynamic solar field performance model is developed and validated on the basis of annual performance data generated using a detailed dynamic solar field model and meteorological data. The parameters of this new performance model comprise the basic geometry of the field, heat capacity of the headers and mean performance characteristics of installed collectors such as optical efficiency, heat loss factor, heat capacity and incident angle modifier (IAM) parameters. These performance parameters are identified from test data by means of numerical optimisation including an assessment of their uncertainties. A systematic investigation of the influence of test data on the values of identified parameters leads to basic data selection criteria in terms of required temperature levels, ranges of incident angles and irradiance level ensuring the representativeness of parameters.
An assessment of the three exemplary test scenarios of a high precision test facility, a mobile field laboratory and a set of built-in sensors yields combined measured uncertainties of ±1.8%, ± 3.5% and ±4.0% respectively (corresponding to 2? or 95%) for solar field thermal efficiency. The uncertainties of the identified parameters reflect the levels of uncertainty of the scenarios.
The annual yield is predicted using a probabilistic approach and Latin hypercube sampling to generate sets of performance parameters from their uncertainty distributions. The resulting uncertainty ranges of annual yields associated with the predictions amount to ±2-4.5% for the considered tests set-up and a coverage probability of 95%.
The new parameterised dynamic solar field performance model, the parameter identification procedure with full parameter uncertainty evaluation and the probabilistic annual simulations combine to form a new approach for acceptance testing ready for parabolic trough solar fields currently being commissioned.