Quality verification of meteorological data used for energetic simulation

Abstract

The objective of this project is to create a methodology that validates the quality of the data included in TMY (Typical Meteorological Year) files, in particular, solar irradiation data. A TMY file contains information about a typical meteorological year in a specific location. It represents the average weather conditions over periods such as 30 years and typically contains hourly values of several meteorological parameters such as solar radiation. These files are commonly used in energetic simulation, therefore their accuracy is important for energy efficiency. In order to find the most adequate method, a TMY file known to contain errors and inconsistencies is used. The TMY file taken into account represents the typical meteorological year of Leignica, a city in Poland. In first place, available methods from literature are analyzed. Three kind of tests are carried out: threshold tests that establish upper and lower limits that radiation should not exceed; step tests that establish an upper limit that the difference in radiation between two time-steps should not surpass; and finally a coherence check that verifies the geometrical relationship between radiation components. In order to assess these inconsistencies, nine indicators are created. These indicators are developed as the number of hours in the year that conditions are not met. So their effectiveness can be verified, they are applied to Leignica and other files known to contain coherent information. By assessing the indicators obtained, there is no significant evidence that the data in the Leignica TMY file contains more errors than data in the other files taken into account. Furthermore, there are three indicators that have no apparent utility, since they are null for all or almost all locations (indMinV , indGR;Gd and indST;Gb). For this reason, it is possible to conclude that ordinary tests in the bibliography do not point out many of the usual deviations in TMY files such as the Leignica file taken into account here. In second place, a different methodology is carried out: a Matlab function is developed. This function automatically flags inconsistencies, so they can be later subjected to visual control. This new method does not use theoretical values but synthetic data obtained with Meteonorm in the form of a TMY file. The objective is to compare the two TMY files. It is necessary to comment that Meteonorm files are not necessarily perfect but they provide a coherent estimation of what to expect in an specific location. To begin with, it is possible to check the consistency of sunrise and sunset times, which cannot be done through the tests found in literature. This is one of the principal advantages of this method. In this case, five indicators are created: three for radiation values that are based on the difference between two curves CDF (Cumulative Distribution Function), and two for sunrise and sunset hours that represent the number of days in a year that there is a lag. These new indicators are applied to the same files as before so they can be compared. These tables show that indicators based on the analysis of synthetic data effectively provide information about the validity of the TMY data. Normal beam radiation, Gbn is indeed higher than expected in the Leignica file, and this fact is effectively remarked by indGbn, as well as differences in the times of sunrise and sunset in the TMY file in comparison with the synthetic data. It is clear that the obtained indicators consistently show whether the information contained in the TMY file is valid or not.