The theoretical analysis will be based on the average data recorded at the Tartu-Toravere Meteorological Station (TOR) of the Estonian Meteorological and Hydrological Institute [4] (http://www. emhi. ee/index. php? ide=8,74). This station is located in the south-eastern part of Estonia (58.25° N, 26.5° E, 70 m a. s.l.). It is a high-quality radiation monitoring station in the Baseline Surface Radiation Network. Radiation data, measured continuously since 1955, characterize the whole region from Scotland up to the north-west Russia. The modest actinometrical resource at TOR ~980 kWh-1m-2 is prevailingly concentrated on the summer season and can be converted (to electricity or heat) only partially due to the selectivity of solar collectors on azimuth. The analysis will cover six summer months from April to September, when radiation is technologically useable. We will ignore the efficiency of energy conversion, but will consider the incident angle modifier k0 for the beam component. Incident angle modifier [5] is a continuous variable that depends on the incident (attack) angle of the beam component 0.

k9 = 1 — tan1/r (<9/2), r є {0.25 — 0.4}.

Also, the effective (convertible) beam irradiance is the product of the “natural” (actinometrical measured) irradiance and the said incident angle 9 modifier. The diffuse component will be approximated with its isotropic model. It means that the performance of the module due to the diffuse component does not depend on its direction, but on the volume of the “visible” sky sphere, characterized by the angle ^x in Fig. 1. This is a simplification that involves some error. Declination 5 has the step of one month. Clock angle ю has the step of 10 minutes to be matched with the recording interval at the experiment1 to study the transient processes. Accordingly, the original hourly dataset [4] was interpolated into a 10-minute dataset.

The location and time variables assess the zenith angle 9z [6] and together with the construction variables the current incident angle 9. In our examples, construction variables are: the initial tilt angle (mostly p0 = 45°), initial azimuth (y0 = 0°), and the deflection angle x =var. Incident angle modifier k9 (yr) depends on the relative azimuth of the sun yr, i. e. the difference between the azimuth of the sun as and the azimuth of the projection of the gradient line of the deflected module on the horizontal plane. GbT is the calculated beam irradiance on the deflected module to be corrected due to the incident angle modifier.

A tilted module can “see” only a part of the half-sphere of the sky, limited due to the tilt angle of the deflected module Px. In the farm its visible sector is additionally limited because of the neighboring module shading effect. Therefore, the corrected diffuse component Gd has to be

The experiment was made at Tallinn University of Technology 59.4° N, 23.7° E, 30 m a. s.l.

limited proportional to the “visible” sector yx Fig. 1. The sum of both of the corrected (beam and diffuse) component values on the deflected module is the effective global irradiance GTx, which is converted into electricity (and heat). A flow-chart of the simulation model is shown in Fig. 2. Due to predicted volume of the present paper we cannot discuss some practical problems, which involve some additional limitations.