The MG frequency and voltage control mechanisms described in Sects. 12.3.1 and
12.3.3 ensure continuous power balancing and frequency restoration to nominal values during islanding operating conditions. However, the effectiveness of the strategies will depend on several operating conditions, such as:
• MG storage capacity, which is essential to ensure primary frequency regulation.
• Controllable MS reserve capacity, in order to perform secondary frequency regulation.
• Non-controllable MS power production, which acts as a negative load in the system.
• MG flexible load, including the EV connected to the LV network.
• MG load in comparison to total generation.
The information sent by the smart meters to the MGCC can be used to characterize the MG operating conditions and coordinate the available resources in order to improve the MG resilience to severe disturbances. Updating the MG emergency operation strategies according to the MG actual operating state will allow:
• Minimizing the amount of load to be disconnected.
• Minimizing the time during which loads are disconnected.
• Ensure that the MG has sufficient storage capacity to ensure primary frequency control following a given disturbance.
• Maintain frequency excursions within admissible limits.
The algorithm presented in Fig. 12.17 will run online at the MGCC in order to analyze the MG operating conditions and then identify the most appropriate actions to ensure a secure islanding. It consists in four main steps, namely:
(1) Characterize the MG operating state based on the information sent by smart meters. The algorithm determines the MG storage capacity, microgeneration reserve capacity, power flow between the MG and the MV network and load (divided between the controllable and non-controllable load).
(2) Determine the severity of the disturbance. If the MG is operating interconnected to the main grid, the occurrence of an unplanned islanding will cause an active power unbalance (Pdist) equal to the power exchanged with the MV network. If the MG is operating islanded, the power unbalance will result from the changes in loads or generation. It is important to notice that planned islanding events are not a key concern, since adequate control action can be taken in order to balance the MG load and generation prior to islanding, thus minimizing the associated transient phenomena.
Fig. 12.18 MG simplified dynamic model to run at the MGCC
(3) Determine the amount of load to shed. In case the MG does not have enough reserve capacity, it is necessary to exploit emergency responsive loads and shed some part of the MG load in order to ensure power balance.
(4) Evaluate the security of the MG during an unplanned islanding. The algorithm determines if the MG has enough storage capacity to ensure power balance during the time required for restoring the frequency to nominal values or if the frequency goes out of admissible limits. Otherwise, it might be necessary to shed some load temporarily.
In Step (3) the amount of load to shed (DPshed) is determined as in Eq. (12.5), based on the MG reserve capacity (R) and on the active power unbalance (Pdist) estimated in Step (2),
DPshed = R — Pdist (12.5)
During islanding conditions, if the MG lacks energy to ensure power balance, the system will collapse. The energy balance in the storage unit (E) can be determined by Eq. (12.6).
E = Edp — Ems — Eve — Erl (12.6)
where EDP is the energy resultant from the power unbalance after the MG disturbance, EMS is the energy provided by the MS (both controllable and noncontrollable) and EVE is the energy resultant from the response of the EV to the frequency deviation. ERL results from: (1) the energy not supplied to responsive loads, which were disconnected due to the emergency state of the MG or (2) the energy to be supplied to dump loads, in case the disturbance leads to an excess of generation regarding the MG load.
Based on Eq. (12.6) it is possible to evaluate the energy required to ensure secure islanded operation. However, the energy injected/absorbed by the MG storage unit will depend on the response of the controllable MS such as SSMT or fuel cells (for example Solid Oxide Fuel Cells—SOFC). Since these units present a non-linear power response, the energy required to ensure MG robust operation following a disturbance cannot be accurately estimated using simplified linear models.
In order to overcome this difficulty, a simplified MG dynamic model that is represented in Fig. 12.18 was adopted. The model consists on a single equivalent bus, considering only the load, generation and storage resources. The model adopted neglects the presence of the LV network and the dynamics of the power converters, which are faster than the MG dynamic behavior. The storage unit and EV are represented by their external P-f control loop. TdP is the delay related to the response of voltage source inverters and Tinv is the delay of the EV grid coupling inverter . Loads and non-controllable MS are represented as constant active power.
The load shedding and the generation emergency dispatch are used as inputs of the MG dynamic model in order to evaluate the energy balance within the MG and the expected frequency deviation for a given period. As outputs, the model provides the total energy injected by the storage units and the MG frequency response. Based on these values, the algorithm then verifies if the MG storage units have sufficient capacity to ensure power balance and if the minimum frequency does not violate the admissible frequency limits (fmin). If these conditions are violated, it is possible to follow an iterative procedure to determine an additional amount of load to be disconnected and compensate the slow response of some MS to power control signals. The frequency threshold for its activation can also be identified, based on the frequency response obtained from the MG simplified model.
The proposed approach is intended to support MG islanding operation during short periods of time (i. e. less than 1 h). For larger time frames of operation in islanding conditions, complementary approaches need to be considered, involving forecasting of loads with different degrees of flexibility (including EV), as well as forecasts for renewable based microgeneration.