A small 50 Hz grid system of ten buses with overcurrent protection system is constructed in ETAP as in Figure 1. There are three buses at extra high voltage level developed in ring main system. Four transmission lines are developed in which one is transmitting 132 kV while the rests are for 275 kV. The current and power flow simulations are based on Newton-Raphson method. There are two clustered loads of voltage 33 kV and 0.23 kV. A distribution feeder is developed with three buses consisting of single-phase loads as shown in Figure 2.

The specifications of the transmission lines in the grid system are tabulated in Table 1 and that of the underground cables in the distribution feeder are shown in Table 2.

The loads of the distribution feeder at different hour of the day as shown in Figure 3 is based on the scaleddown version of real-time power demand of New South Wales on June 28, 2012. Bus 1 and Bus 2 of the distribution feeder has equal amount of loads.

There are two cases which are used in illustrating the decrease in output power of generator when there is solar power injection. The parameters of the power generation and loads are tabulated in Table 3. The solar power injection into the grid system in the first case is shown in Table 4.

In Figure 4, the output power from the swing generator is observed to reduce with the injection of solar power into the grid system. In every scenario, the output power from the swing generator reduces in proportion to the increase in solar power injection. Hence, at similar loadings, some of the gas turbine generators can be halted in operation. This reduces the combustion of fossil fuels and less greenhouse emission to the environment. The curve showing the output power of swing generator as derived in ETAP deviates away from the anticipated curve. The anticipated curve predicts the reduction of output power from swing generator equal to the amount of solar power injection. The deviation between the anticipated curve and the power curve of swing generator is the transmission losses.

Since the anticipated curve parameter is higher than the power curve of the generator, it is concluded that the transmission losses are reduced with solar power injection. Case 2 illustrates in details the reduction in transmission lines losses with increase in solar power injection into the grid system. Figure 4 shows the reduction in

Table 3. Power generation and demands of Case 1.

Table 4. Solar power injection of Case 1.

output power from swing generator after injection of solar power in different scenarios.

In Case 2, the power generation and demands are the same as in Table 3. However, the generation of MVARcontrolled generator is manually reduced in ETAP based on the exact amount of solar power injection into the grid. Table 5 shows the active solar power injection in different scenario simulated.

Figure 5 shows the output power from the swing generator in Case 2 and the anticipated output power. Figure 6 shows the manual reduction of the output power of MVAR-controlled generator which is inversely proportional to the level of solar power injection into the grid.

It is predicted that by having solar power supplying to the loads, the amount of power output from MVARcontrolled generator can be reduced by the same amount and hence, the swing generator output power will be constant. However, in Case 3, the output power of swing generator is not constant but reduces with increasing solar power injection which is shown in Figure 5. The deviation of the power curve of swing generator and the anticipated curve is explained by the power loss in the transmission lines. With no solar power injection in the grid, the amount of output power from swing generator is equal to the total power demand of the loads and also the power loss in the transmission lines.

With solar power injection into the grid, the power across the tranmission lines is lesser and hence, lesser

Table 5. Solar power injection of Case 2.

transmission losses which lead to less power generation required from the swing generator.

With the power demand as shown in Figure 3 for each bus in the distribution feeder, the bus voltage at different hour of the day was obtained through simulation in ETAP. Without solar power injection, the bus voltages are well within the acceptable range of operating voltage at every hour of the day. Figure 7 shows the bus voltages with and without solar power injection in the distribution feeder.

In the first case, the amount of solar power injected to each bus of the distribution feeder is illustrated in Figure 8. The solar power injection in Figure 8 is the scaleddown version of solar power generation at Kalgoorie, Western Australia. Equations (1)-(4) were used in calculating the altitude and azimuth angle.

where

β = altitude angle,

= azimuth angleL = latitude,

= solar declination on 21^st day of the monthH = hour anglen = first day of each month.

Next, the hourly clear-sky insolation on PV array on 21^st day of each month of a year for horizontal collector was calculated using Equations (5)-(12).

where I_B = beam of radiationA = “apparent” extraterrestrial fluxk = optical depthm = air-mass indexI_BH = direct beam radiation for horizontal collectorβ = altitude angleI_DH = diffuse radiation on horizontal collectorC = sky diffuse factorI_c = total radiation on collector.

After calculation of the hourly clear sky insolation on the PV array on 21^st day of the month, the obtained results were adjusted to practical value using Equation (13).

The measured average daily insolation is the practical data provided by National Climate Centre, Australia. In January, the measured average daily insolation is 7.861 kWh/m².

Next, the total clear sky insolation was multiplied by the value of k. The area of the module was assumed. The total power at terminals and the power loss in each degree rise in temperature were calculated using Equations (14)-(16).

where T_amb = 33.7˚C (assumed for January)NOCT = Nominal Operating Cell Temperature = 45˚CS = 1 kW/m².

The power loss due to dirt and mismatch for horizontal collector was determined using Equations (17) and (18) calculated the final maximum power.

Based on the total final maximum power determined for every hour of the day in January, the results were scaled down and shown in Figure 8.

The bus voltages after injection of solar power, as shown in Figure 7, illustrate a rise in the bus voltages. However, the bus voltages remain within the acceptable range and demonstrates a better voltage regulation compared to when there is no solar power injection into the distribution feeder.

In the second case, the solar power injection of Bus 1 and Bus 2 remain constant. As for Bus 3, there is a large increase in solar power injection as shown in Figure 9.

The third case is the amount of solar power injection of all buses are scaled by factor of 18 of Case 1. The operating bus voltages were measured in ETAP and shown in Figures 10 and 11 respectively for Case 2 and 3.

In both Case 2 and Case 3, it is observed that the voltage at Bus 3 in between 11:30 am to 12:30 pm has increased to more than the maximum allowable voltage of 440 V. Both cases illustrate the overvoltage issue due to large-scale solar power injection at low power demand.

In uni-directional power flow grid system, the protection systems are set by considering the full load current associated with each relay and the time discrimination of the

relays. In simulating Case 1 which different amount of solar power is injected into the grid, the loads and MVAR-controlled generation of Case 1 are as shown in Table 6.

Figure 12 shows an example scenario where the reverse power flows from 0.4 kV bus to 33 kV bus in supplying of power to the loads at 33 kV due to injection of large amount of solar power at the 0.4 kV bus. The figure also includes the location of the two affected relays, Relay 33 and 11.

Table 6. Power generation and demands of Case 1.

Figures 13 and 14 illustrate the amount of current flowing through the primary side of the current transformer associated to Relay 33 and 11 respectively and their respective relay pickup current.

At active solar power injection of 45 MW and above, the current flowing through the primary side of current transformer associated with Relay 11 and 33 exceeds the pickup current. This is as shown in Figures 13 and 14 respectively. The loads at 0.4 kV bus is lesser than that of in 33 kV bus. In conventional grid system, the pickup current is set to meet the full load current of the lesser low-voltage loads and the associated fault current.

Hence, large amount of solar power injection into the grid and reverse power in supplying to 33 kV loads has possibility of exceeding the pickup current of the relays.

Under this condition, there is no reverse power flow to the generator but circuit breaker is tripped as signaled by either Relay 33 or Relay 11. Hence, at active solar power injection of 45 MW and above, there is false operation of overcurrent protection system which is designed for uni-directional power flow grid system.

In solving the overvoltage issue without reducing the amount of solar power injection at the low-voltage network, one potential solution is to use bi-directional inverters in interfacing the battery banks with the buses in the distribution feeder. The bi-directional inverters and battery banks are to be installed by the grid operator to ensure the bus voltages are within the acceptable range.

In ETAP, the bi-directional inverter is illustrated using an inverter and a charger with combination to a double-throw switch. In between 10 am to 4 pm, the solar power injection in the distribution feeder is high but low power demand. Hence, at this time interval, the bi-directional inverter works as a charger, taking portion of the power from the AC buses and charging the battery banks. During night time, the bi-directional inverter works as an inverter, converting DC power of the battery banks to supply to AC buses for the loads. The charging mode of the bi-directional inverter is shown in Figure 15.

The parameters of the single battery bank used in ETAP are shown in Table 7.

By charging the battery banks using bi-directional inverter from 10 am to 4 pm, the new bus voltages measured in ETAP remain in the acceptable range. The resulted new bus voltages are as tabulated in Table 8.

In order to mitigate the risk of false operation of Relay 33 and Relay 11 due to reverse power flow to 33 kV bus under normal grid condition, the relay current settings are adjusted based on the maximum bi-directional current flow across the primary side of the associated current transformer, whichever direction where the current is greater. Considering the maximum active solar power injection is 60 MW, the new relay current settings are tabulated in Table 9 with comparison to the previous relay current settings.

Hence, the pickup current of Relay 11 and 33 are increased. With the new pickup current, Relay 33 and 11 will not operate at maximum active solar power injection of 60 MW. Hence, there is no possibility of false tripping due to reverse power flow to 33 kV loads. However, all relays have to be re-graded to ensure of the correct operation time during fault.