The peak load in Riyadh occurs during the hot season, and in general between noon and 4 pm. The power plant and the transmisson and distribution lines are then under maximum load. Also, the peak load on the utility grid in Riyadh coincides with the peak solar-insolation, which results in peak power-generation in the PVGC system. Thus, the PVGC system provides an excellent means of reducing the peak load imposed on the grid. The PVGC system does not need electrical batteries for the storage of electrical energy because the utility grid acts as the store for the PVGC system. Normally, the size of storage batteries of the PV stand-alone system should be sufficient for more than 3 days supply in Riyadh. Therefore, the cost of the PVGC system is 40% less then the PV stand-alone system, of equivalent power supply as the PVGC system. The 40% decrease in the initial cost is due to the cost of storage batteries (capacity, type and quality).

The power of the PVGC system is 6 kWp under standard conditions, i.e. 1000 W/m², 25°C and 1.5 AM. The PV generator is divided into 6 arrays, each of 1 kWp. The PV arrays are wired to supply dual voltages to the inverter, i.e. positive, zero and negative. The inverter has a built-in automatic controller, which is suitable for different methods of operation and protection of the PV arrays, inverter and the parallel operation with the utility grid. Several parameters are recorded by the data-logger in order to obtain a performance analysis of the PVGC system. These parameters include solar radiation incident on horizontal and tilted planes, ambient and PV module temperatures, voltages and currents for the positive and negative arrays, inverter input and output voltage, current and power of the inverter, and current supplied to the load and the utility grid, i.e. value and direction of flow of the current between PV generator, utility grid and load. To simulate various possible operating conditions, the load current has been varied.

During the period of the peak load imposed on the utility grid, all the electrical equipment, i.e. power generator, transformer, and transmission and distribution lines, is fully loaded, and the rate of energy loss is a maximum in the grid equipment. Therefore producing the electrical power at a location near to the load, i.e. via the PVGC system, will reduce the percentage loading upon the equipment and hence the amount of energy lost in the equipment of the utility grid is less.

The PVGC system does not need electricity-storage batteries, because the electrical power generated by the PV array is, automatically, either consumed by a local load or fed to the utility grid or shared by the load and the grid.

The results obtained from the simulation of the behaviour of the PV array,under various conditions, have been studied and compared with experimental data. Many software packages are available for the simulations of PV array behaviour but the predictions from various packages differ.

Also the degree of agreement between the predictions from the simulation and the experimental data depend on the accuracy of the values of the input parameters for the software, e.g. for the dust accumulation on the collectors surfaces: the latter is a function of the tilt angle, the wind speed and direction, and the frequency of cleaning of the PV arrays. Others input parameters to the simulation software have important effects on the output results of the simulation of PV array systems behaviour, namely solar radiation (and hence the tilt of the array, and the solar spectrum variation – hourly, daily and monthly), ambient temperature, operating temperature of the PV array and its tilt angle (fixed or tracked on a single axis). The values of the measured data for the PV array depend on the accuracy of the equipment used, i.e. sensors and transducers, which are used to measure various values of voltage, current, solar radiation and temperature. So the accuracy of the measuring equipment used should be checked periodically in order to reduce the errors. The accumulation of dust depends on the tilt angle and the climatic condition of the location of the PV array. Dust accumulation can reduce the output power from the array by 5% to more than 25% for low and high accumulations of dust, respectively. The assumed 5% loss of power due to dust accumulation on the PV array surface can be reduced to less than 1% at the Solar Village by frequent cleaning of the PV array. The reference conditions for the comparison of various cases is the case numbered 1 in Table 1, i.e. a fixed tilt angle of 25° (namely the latitude of the Solar Village), and 5% loss of power due to dust accumulation on the PV arrays. The other parameters are kept constant, i.e. albedo =20%, mismatch losses between various arrays =3% and wiring losses of the PVGC system =2%.

From Table 1, the following conclusions can be drawn relative to the reference conditions (i.e. case 1):

• PV arrays, with a fixed tilt angle, which is equal to site latitude (case 1), give a higher output per year when compared with arrays having other values of the fixed tilt-angle, either smaller or larger in magnitude.

• PV arrays with a single-axis tracker along with fixed tilt angle (case 5) give about 22.2% more output than achieved with the reference systems (case 1).

• PV arrays with single tracker along with monthly best tilt angle (case 8), give about 25.8% more output power than achieved with the reference system (case 1).

• When the PV array is left without cleaning for a long time, the dust accumulation can reduce the output of the reference system (case 1) by 21.05%. Therefore, because the cost of the PV system is relatively high as compared with other sources of power, it is advisable to clean the PV array each time it is required. Cleaning can be accomplished quickly and is not a tedious task.

• Normally the single-axis tracker system (case gives a high output-power, but with 25% loss of power due to dust accumulation, i.e. case number 16. This will result in less output power than the reference condition (case 1) by 0.9%. Therefore the cleaning of the PV array, whenever required, is an economic and efficient way to get the maximum possible outputs from the PV system.

This was studied under various conditions, such as with respect to variation of the tilt angle, load power and the cleaning of the PV array surface. The tilt angle was varied according to the season. The power and the duration of operation of the load were also varied. The weekly profile of the load is ON for 5 days (i.e. during the day) and is OFF for 2 days.

Using Eq. (1), several cases are considered as follows:

• When the load is OFF, therefore, P_load=0, and P_inv=?Pgrid: this means that the output power of the inverter is given to by the grid.

• When the load is ON but P_inv=0, i.e. very low incident solar radiation is received on the surface of the PV array, then P_load=P_grid, which means that the load is supplied by the grid.

• When the load is ON but P_inv is less than P_load, e.g. P_load=4 kW and P_inv=2 kW, after substitution in Eq. (1), it can be seen that P_grid=2 kW, which means that load power is supplied from the inverter and the grid. Fig. 2 shows the conditions which are discussed above: as an example at 7:00 am, the load is supplied only from the grid, while after 7:30 am, the load is supplied from the PV generator, which is also supplying the energy to the grid. Also, it is clear from Fig. 2 that starting from noon, the PVGC is supplying current to satisfy the load and is feeding the grid. Therefore, the PVGC system is reducing the peak load imposed on the grid. Fig. 3 shows the performance of the PVGC system. It is clear from the figure that, during the month of September 1996, all the energy generated by the PV-inverter is supplied to the grid, which means that the load is OFF. But in September 1997, the load was increased, so that it consumed all the energy generated by the PV-inverter: therefore, P_grid=0, which means that the grid neither supplies nor receives energy. When P_grid=0, the inverter of the PVGC system must be internally protected against the case of islanding. The performance of the inverter was excellent during every islanding test.