**3. Experimental Set-Up**

Figure 5 sketches a realistic situation in which, during the day, a shadow overlays different panels. As a consequence, the shadowed panels disturb those connected in parallel: a decrease in voltage of shadowed panel involves a decrease of not shadowed ones and a rise of current for the power balance. This rise of current is not constantly probable: if the not shadowed panels are in the area of maximum power point, any decrease of voltage decreases the power.

**Figure 5.** Probable shadow projections in the studied situation.

Figure 6 shows the effect of a shadow due to the presence of a pole by considering a PV DRS scheme. The hypothesis is that each panel has three lines of cells and is connected to the DRS. DRS is connected to a two channels inverter.

**Figure 6.** DRS in a PV plant.

Reconfiguration performances were tested with a prototype DRS developed at the University of Palermo and working on a twelve panels system. DRS acquires the state of every panel with a sensing system (voltage, current and temperature).

Figure 7 shows the experimental DRS system. Twelve panels (PMMP 215 W, VMPP 28.27 V, IMPP 7.59 A, Voc 36.37, Isc 8.21 A, Conergy, (Hamburg, Germany) have been connected individually to the DRS. Panels located at latitude: 38◦ 5.9'; longitude: 13◦ 20.6', measurement carried out in the morning of 27 April 2016, from 10:00 to 13:00, with incident radiation between 1.0 and 1.1 kW / m<sup>2</sup> and temperature between 20–25 ◦C.

Without shadows, each panel works in the same point: the DRS generates basic topology, in this case, two parallel identical strings of six modules in series. To perform the behavior of the DRS, different resistive loads have been considered and Table 9 lists the different working conditions.

In order to test the DRS, an artificial shadow has been created. Figure 8 displays three shadow cases for a panel connected in series with five others. The artificial shadow cuts one, two or three lines of modules, dropping the performance of the panel and of the string. Each stoppage of line requires the action of the bypass diode, and a successive voltage decrease of the panel. The shadowed panel

has labelled with the number 6, so V1–V5 are the voltages of normally irradiated panels and V6 is the shadowed one; DRS evaluates the power of each panel and connects the panels into the string.

**Figure 7.** Experimental system with panels individually connected to DRS.



**Figure 8.** Shaded panel case 1, 2 and 3. In case 1 shadow can vary from 225 to 450 cm2; which corresponds to one interruption of a line of cells; in case 2 shadow varies from 450 to 900 cm2, which corresponds to two interruption of lines; case 3 corresponds to the interruption of the panel.

Table 10 recapitulates the behavior of the DRS with shading and without; P1–P5 are the powers of the not shadowed panels, P6 is the power of the shadowed one in three cases, I is the current of the system in the different cases taken into account


**Table 10.** Electrical characteristics of the string in different shadow conditions.

Figures 9 and 10 show respectively the voltage-current and voltage-power profiles of the panel with three working lines of cells, two working lines of cells and only one working line. Blue curves describe the working points without any shading. If a line of cells is shaded the orange curve has to be considered, maintaining a similar current and with a new voltage. When a shadow covers two lines of cells from the blue curve the gray curve has to be considered, maintaining a similar current and with a new voltage.

**Figure 9.** Interpolated V-I curves of the panel with different shadows. Working lines have been described.

**Figure 10.** Interpolated V-P curves of the panel with different shadows. Working lines have been described.

New voltages indicate the new power conditions. The DRS is able to regroup similar irradiated panels and/or exclude the densely shaded panels. The different operation is due to the DRS architecture and the algorithm implemented. A not smart DRS can only exclude shaded panels, a high-performance DRS relocates them on suitable dynamic arrays.
