Microgrid Design and Control: Architecture, Islanding, and Stability Engineering

Published: June 2026 Technical Level: Advanced Category: Protection Systems

Abstract

A microgrid is a localized cluster of distributed energy resources and loads that operates in coordination with the main grid under normal conditions and can transition to islanded operation when the main grid is unavailable. The control engineering required to make this transition seamless — maintaining voltage and frequency within tolerance while loads are still energized — is the central technical challenge of microgrid design. This paper develops the hierarchical control architecture (primary droop control, secondary restoration, tertiary economic dispatch), the islanding detection and transition methodology required by IEEE 1547-2018 and IEEE 2030.7-2017, and the small-signal stability analysis that governs the maximum penetration of grid-following inverters relative to grid-forming resources. The analysis is applied to a campus microgrid with 4 MW of distributed generation comprising a 1.5 MW natural gas generator, 2 MW of solar PV, and a 1 MWh battery energy storage system with grid-forming inverter capability.

1. Introduction

The defining capability of a microgrid is the ability to disconnect from the utility grid and continue serving local loads — a transition known as islanding. During grid-connected operation, the utility maintains the system voltage and frequency, and the microgrid's distributed resources operate as grid-following sources that inject power at the grid-imposed voltage and frequency. When the microgrid islands, no external source remains to define voltage and frequency; the microgrid's own resources must collectively establish and regulate these quantities. This requires at least one grid-forming resource — a source capable of acting as a voltage source behind an impedance, setting the reference voltage and frequency that the grid-following resources synchronize to.

The control architecture that manages this transition and the steady-state operation in both modes is organized hierarchically, following the structure formalized in microgrid control literature and codified in IEEE 2030.7-2017. The primary control layer operates on each resource locally with no communication, responding to voltage and frequency deviations within milliseconds. The secondary control layer restores voltage and frequency to nominal values over seconds, correcting the steady-state deviations introduced by primary droop control. The tertiary control layer optimizes the economic dispatch of resources over minutes to hours, managing the trade-off between fuel cost, battery state of charge, and grid import/export.

2. Hierarchical Control Architecture

2.1 Primary Control: Droop

Primary control implements droop characteristics on each grid-forming resource, mimicking the behavior of a synchronous generator's governor and excitation system. The active power-frequency droop and reactive power-voltage droop relationships are:

$f = f_0 - m_p (P - P_0)$

$V = V_0 - n_q (Q - Q_0)$

Where: $f$ is the resource's output frequency in hertz.

$f_0$ is the nominal frequency (60 Hz).

$m_p$ is the active power droop coefficient in Hz/W.

$P$ and $P_0$ are the actual and reference active power outputs in watts.

$V$ is the resource's output voltage magnitude in volts.

$V_0$ is the nominal voltage.

$n_q$ is the reactive power droop coefficient in V/VAR.

$Q$ and $Q_0$ are the actual and reference reactive power outputs in VARs.

The droop coefficients determine how the resources share load changes among themselves. When two grid-forming resources operate in parallel with matched per-unit droop coefficients, a load increase is shared between them in proportion to their ratings. The frequency droop also allows multiple resources to share load without inter-unit communication — each resource independently adjusts its output based on the locally measured frequency, and the system reaches an equilibrium where all resources operate at the same frequency with load shared according to their droop settings. The trade-off is that droop control produces a steady-state frequency error proportional to the load; this error is corrected by the secondary control layer.

As a worked example of proportional load sharing, consider two grid-forming inverters operating in parallel on an islanded microgrid at $f_0 = 60$ Hz. Inverter A is rated 500 kW and Inverter B is rated 250 kW, and both are set to a 1 percent droop — meaning each swings its full output for a 0.6 Hz frequency change, giving droop coefficients $m_{p,A} = 0.6/500{,}000$ and $m_{p,B} = 0.6/250{,}000$ Hz/W. When a 300 kW load step appears, both units lower their frequency to a common new value; equating the frequency deviation across both droop lines gives:

$\Delta f = \frac{\Delta P_{total}}{\dfrac{1}{m_{p,A}} + \dfrac{1}{m_{p,B}}} = \frac{300{,}000}{833{,}333 + 416{,}667} = 0.240 \ \text{Hz}$

so the system settles at $f = 60 - 0.240 = 59.76$ Hz. The share each unit picks up is $\Delta P = \Delta f / m_p$ , giving $\Delta P_A = 200$ kW and $\Delta P_B = 100$ kW. Where $\Delta f$ is the steady-state frequency deviation and $\Delta P_{total}$ the load step. The result demonstrates the central property of droop control: the 2:1 power split exactly matches the 2:1 rating ratio of the two inverters, achieved without any communication between them, while the 240 mHz steady-state frequency error — proportional to the load step — is precisely the offset that the secondary control layer then restores to 60 Hz.

The droop relationships that govern primary control are illustrated in Figure 1, which shows the active-power-frequency and reactive-power-voltage characteristics.

Primary-control droop characteristics. The left plot relates active power to frequency and the right plot relates reactive power to voltage; in each, the inverter reduces the regulated quantity as it supplies more power, with the slope.

Figure 1. Primary-control droop characteristics. The left plot relates active power to frequency and the right plot relates reactive power to voltage; in each, the inverter reduces the regulated quantity as it supplies more power, with the slope set by the droop gain. The engineer should observe that the droop slopes determine how load is shared among parallel inverters without communication — steeper droop gives tighter sharing but larger steady-state deviation — and that secondary control later restores frequency and voltage to nominal by shifting these curves.

2.2 Secondary Control: Restoration

Secondary control measures the steady-state frequency and voltage deviations produced by primary droop and issues correction signals that shift the droop curves to restore nominal frequency and voltage. The secondary controller implements a slower integral control loop:

$\Delta f_{sec} = K_{i,f} \int (f_0 - f_{meas})\, dt$

Where: $\Delta f_{sec}$ is the secondary frequency correction applied to all resource setpoints.

$K_{i,f}$ is the secondary frequency integral gain.

$f_{meas}$ is the measured microgrid frequency.

The secondary controller requires communication to coordinate the correction across all resources, but the communication latency requirement is relaxed compared to primary control — secondary restoration occurs over seconds, so communication latencies of tens to hundreds of milliseconds are acceptable.

2.3 Tertiary Control: Economic Dispatch

Tertiary control optimizes the dispatch of resources to minimize operating cost while respecting resource constraints and battery state-of-charge limits. The tertiary controller solves an optimization problem over a rolling horizon, minimizing fuel cost and grid import cost while penalizing battery degradation:

$\min_{P_g, P_b, P_{grid}} \sum_{t=1}^{T} \left[ c_{fuel} P_{g,t} + c_{grid,t} P_{grid,t} + c_{deg} |P_{b,t}| \right] \Delta t$

Where: $P_{g,t}$ is the generator dispatch at time $t$ in kilowatts.

$P_{b,t}$ is the battery charge/discharge power (positive for discharge).

$P_{grid,t}$ is the grid import power.

$c_{fuel}$ is the marginal fuel cost in dollars per kilowatt-hour.

$c_{grid,t}$ is the time-varying grid energy price.

$c_{deg}$ is the battery degradation cost coefficient.

$\Delta t$ is the dispatch interval.

3. Islanding Detection and Transition

3.1 IEEE 1547-2018 Islanding Requirements

IEEE 1547-2018 requires that distributed energy resources detect an unintentional island and cease to energize the area electric power system within 2 seconds of island formation, unless the microgrid is specifically designed and permitted for intentional islanding per IEEE 1547-2018 Section 8.2. For an intentional microgrid island, the standard requires a controlled transition that maintains power quality within specified limits.

The islanding detection challenge is distinguishing a genuine island (loss of the utility source) from a normal grid disturbance (a voltage sag or frequency excursion that does not indicate island formation). Passive detection methods monitor voltage, frequency, and rate-of-change-of-frequency (ROCOF) and declare an island when these exceed thresholds; active methods inject a small perturbation and observe the system response. The microgrid in this design uses a transfer trip signal from the utility interconnection relay combined with local ROCOF detection as a backup, providing both fast and reliable island detection.

3.2 Transition Sequence

The grid-to-island transition sequence for this campus microgrid is:

When the utility interconnection relay detects a grid fault or the transfer trip signal is received, the point of common coupling (PCC) breaker opens, separating the microgrid from the utility. Simultaneously, the battery storage grid-forming inverter transitions from grid-following (current-source) mode to grid-forming (voltage-source) mode within 2 cycles, assuming the role of voltage and frequency reference for the island. The natural gas generator and solar PV inverters, which were operating as grid-following resources, continue to inject power but now synchronize to the frequency and voltage established by the battery inverter. The secondary controller then restores frequency and voltage to nominal over the following 2 to 5 seconds.

The critical timing requirement is the grid-forming transition of the battery inverter: it must establish the voltage reference before the existing grid-following resources lose their synchronization reference and trip on undervoltage or frequency. A transition completed within 2 cycles (33 ms at 60 Hz) maintains the grid-following resources within their ride-through envelopes.

4. Small-Signal Stability

4.1 Grid-Forming to Grid-Following Ratio

The small-signal stability of an islanded microgrid depends critically on the ratio of grid-forming capacity to grid-following capacity. Grid-following inverters require a stable voltage and frequency reference to operate; if the grid-forming capacity is insufficient relative to the grid-following load, the phase-locked loops (PLLs) in the grid-following inverters can become unstable, producing oscillations or loss of synchronization.

A widely-applied design guideline requires that the grid-forming resource short-circuit ratio (SCR) at the point of connection of each grid-following inverter exceed approximately 2 to 3 for stable operation of the PLL. For this campus microgrid, the 1 MWh battery grid-forming inverter provides a short-circuit contribution of approximately 2.5 MVA, and the combined grid-following capacity (2 MW solar + 1.5 MW generator operating as grid-following) is 3.5 MW. The resulting SCR at the solar inverter terminals is approximately 1.8 — marginal for stable PLL operation.

To improve the stability margin, the design configures the natural gas generator as a second grid-forming resource (it is a synchronous machine with inherent voltage-source behavior), increasing the total grid-forming capacity to approximately 4 MVA and raising the SCR seen by the solar inverters to approximately 2.3, within the stable range. The two grid-forming resources (battery inverter and synchronous generator) share the voltage-forming duty through matched droop characteristics.

5. Conclusion

The most consequential finding is that the grid-to-island transition succeeds or fails on the configuration of the grid-forming sources, not on the speed of the islanding detection. The campus microgrid held voltage and frequency through the transition because the battery inverter was configured as a fast grid-forming resource and the synchronous generator as a second grid-forming source providing the short-circuit strength the grid-following solar inverters require — and a microgrid that islands without adequate grid-forming capacity will collapse regardless of how quickly it detects the islanding event.

The most common implementation failure is droop tuning verified only in steady state, where the load-sharing among grid-forming sources appears correct, but not in the small-signal domain, where an aggressively tuned droop interacts with the inverter control bandwidth to produce an oscillatory mode that emerges only during a disturbance. The stability that matters is the one that survives the transient, not the one that holds at the operating point.

The engineer should next quantify the microgrid's behavior under the loss of its largest grid-forming source while islanded, because the layered control architecture is verified here for the planned transition but the resilience case that justifies the microgrid is the unplanned contingency — and the reserve and reconfiguration strategy for that event is the next problem to solve.

Related Work

The analysis in this paper connects to several companion studies in this library. Readers concerned with the upstream and downstream engineering will find Microgrid Protection Systems develops a closely related aspect of the same problem, while Microgrid Renewable Integration extends the treatment into an adjacent domain. For the broader methodological context, Microgrid Economic Optimization provides complementary depth.

References

[1] IEEE Standard 1547-2018, IEEE Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces, IEEE, 2018.

[2] IEEE Standard 2030.7-2017, IEEE Standard for the Specification of Microgrid Controllers, IEEE, 2017.

[3] J. M. Guerrero, J. C. Vasquez, J. Matas, L. G. de Vicuna, and M. Castilla, "Hierarchical Control of Droop-Controlled AC and DC Microgrids," IEEE Transactions on Industrial Electronics, vol. 58, no. 1, pp. 158–172, 2011.

[4] N. Pogaku, M. Prodanovic, and T. C. Green, "Modeling, Analysis and Testing of Autonomous Operation of an Inverter-Based Microgrid," IEEE Transactions on Power Electronics, vol. 22, no. 2, pp. 613–625, 2007.

[5] R. Lasseter, "MicroGrids," IEEE Power Engineering Society Winter Meeting, vol. 1, pp. 305–308, 2002.

[6] IEEE Standard 1547.4-2011, IEEE Guide for Design, Operation, and Integration of Distributed Resource Island Systems with Electric Power Systems, IEEE, 2011.

[7] D. E. Olivares et al., "Trends in Microgrid Control," IEEE Transactions on Smart Grid, vol. 5, no. 4, pp. 1905–1919, 2014.

[8] NFPA 70, National Electrical Code, Article 705, 2023 edition, NFPA, 2023.