Keeping the grid stable is often synonymous with keeping frequency within a narrow band. Deviations manifest as changes in the voltage and microgrids entirely powered by distributed solar generators are more sensitive to the issue than utility grids fed by a multitude of power sources.
Researchers Han Min Htut and Wijarn Wangdee, of King Mongkut's University of Technology North Bangkok, in Thailand, have tackled the issue and proposed a new inverter control strategy. Their findings were published in Engineering Journal as the article Virtual Oscillator Control of Multiple Solar PV Inverters for Microgrid Applications.
A gradual increase in power electronics-interfaced generation methods on the grid has led to a paradigm shift in how grids respond to disturbances. A microgrid powered by very high penetration of small scale solar will have to face the challenge of stable operation, as all those maximum power point trackers (MPPTs) ensure attaining such an outcome is complicated. In the set-up the Thai-based scholars envisioned for their tests, the PV generation sources were connected to the grid with DC-DC boosters. Shading or other changes in irradiation can alter the input voltage for the DC-DC booster, subsequently changing output voltage.
Virtual oscillator control
The academics suggested use of a modified virtual oscillator control (VOC) and a cascaded sliding mode control (SMC) would help optimize microgrid management strategies. When PV output power is higher than the combined loads in the grid, inverters will not use their maximum power point trackers. However, they will switch back to using the algorithm when the power supply dips below demand. The control strategy enables stable operation of 100% solar microgrids even in islanding mode, without requiring energy storage to stabilize voltage frequency.
To achieve that, the Thai group proposed a hybrid controller with a switch between a ‘fast' MPPT and a slower one for microgrid-integrated solar. In their setup, the power electronics feature a single controller regulating DC-link voltage and MPPT autonomously, without any need for system reconfiguration. Effectively, the two-stage converters can decrease DC-link voltage if the PV capacity cannot meet its droop control command – in which algorithms consider active power frequency and manage the active power output as a function of frequency deviation. Droop control lifts the nadir, the point of widest frequency deviation, and improves the recovery process for grid frequency when large loads are connected.
According to the researchers, the VOC approach provides droop control-like features for voltage and frequency regulation but does so at a smaller computational burden due to its instantaneous measurements, which allow for a faster and better response of dynamic control strategies.
The virtual oscillator uses a control logic that measures inverter output filter inductor current with the output being used to generate the pulse-width-modulation signals. A negative damping resistor and a resonant circuit set the system frequency and a non linear cubic voltage-dependent source sustains oscillation. Controlling the voltage means the cascaded sliding mode control uses two control loops – an inner current loop and the outer voltage one. The system works to force parameters onto predefined values and keep them there.
The researchers tested the new decentralized control methods for single and three-phase inverters to ensure load power-sharing between devices and frequency synchronization of multiple voltage source inverters operating in parallel.
Validated
In a power-hardware-in-the-loop environment, and with Matlab and Simulink's help, the scholars validated their control strategy under different grid conditions. Two PV sources were scaled to 15 and 30 kW capacities and the modeled microgrid was exposed to challenging grid conditions including 100% islanding mode, unbalanced sharing of loads between two power sources, a seamless transition from grid-connected to islanding, and rapid changes to irradiation during islanding.
In one test, for example, the team modeled the microgrid load was a constant 30 kW for 10 seconds. For the first three seconds, one of the two PV systems received irradiation of 1 kW/m2, that then fell to 600 W/m2 and finally 300 W/m2. The other PV system received abundant solar irradiation. The test results demonstrated the control strategy allowed for autonomous switching between load power-sharing and maximum power point modes without reconfiguring the inverter control scheme.
The researchers said their approach could also reduce the utilization rate and scale of energy storage in a microgrid as it would be required solely for nighttime load shifting rather than also for providing stability.
