Microgrids

Integrating renewable generation and storage while designing, modelling, and controlling resilient DC and AC microgrids.

Microgrid Solutions

A microgrid is a bounded set of sources, storage and loads that can operate connected to the utility or on its own. That single capability — islanding — is what makes the architecture choice consequential: everything about control, protection and power quality follows from which bus the energy actually sits on.

DC Microgrids

Efficient DC power distribution systems with advanced control strategies for optimal energy management and reliability.

AC Microgrids

Grid-connected and islanded AC microgrid architectures with seamless transition capabilities and power quality management.

Hybrid AC/DC Architectures

Integrated systems combining the benefits of both AC and DC distribution for maximum efficiency and flexibility.

Microgrid integration

Distributed energy and storage

A microgrid brings energy resources with very different electrical characteristics into one distribution and control system. Wind, solar and fuel cells supply energy; storage manages the timing mismatch between generation and demand. Each resource needs a power-conversion, control and protection interface matched to its behaviour.

01 · Renewable generation

Wind-energy integration

A wind turbine produces AC whose voltage and frequency move with rotor speed. A full-scale or doubly-fed converter decouples that mechanical speed from the microgrid, enabling maximum-power tracking while meeting active-power, reactive-power and fault-ride-through requirements. The rectified link can feed a DC bus, or a grid-side inverter can synchronise it to an AC bus.

Native output
Variable-frequency AC
Typical interface
AC–DC–AC converter
Control focus
MPPT, pitch and ride-through
02 · Renewable generation

Solar-PV integration

A PV array is natively DC, with an operating point that moves with irradiance and temperature. On a DC bus, a DC–DC stage performs maximum-power-point tracking and bus matching; on an AC bus, an inverter also manages synchronisation, power factor and power quality. The energy manager coordinates curtailment, forecast error and the storage available to smooth production.

Native output
Variable DC
Typical interface
DC–DC or DC–AC
Control focus
MPPT, curtailment and bus support
03 · Dispatchable generation

Fuel-cell integration

A fuel cell is a dispatchable DC source, but stack voltage changes with current and its reactant and thermal systems respond more slowly than electrical loads. A boost-derived DC–DC interface regulates voltage and limits current, while a battery or supercapacitor absorbs fast power steps. Dispatch must balance fuel use, stack life and the microgrid's power request.

Native output
Low-voltage DC
Typical interface
Isolated or non-isolated boost
Control focus
Current limits, dynamics and life
04 · Flexibility resource

Energy-storage integration

Storage is the microgrid's bidirectional flexibility layer: it absorbs renewable surplus and releases energy during deficits, islanding transitions and load steps. Batteries provide minutes-to-hours energy, while high-power devices such as supercapacitors handle faster transients. Their bidirectional converters can also regulate the bus, form the grid and support black start, subject to state-of-charge, thermal and ageing limits.

Power direction
Bidirectional charge/discharge
Typical interface
Bidirectional DC–DC or DC–AC
Control focus
SoC, grid forming and ageing

Three architectures, three different problems

The bus is the architecture. Choosing DC, AC or both decides how many conversion stages sit between a panel and a load, what quantity carries power-balance information, and what you have to do to clear a fault.

DC microgrid — one bus, one control variable DC bus ±380 V / 800 V PV Battery AC grid DC loads AC loads
Fewer conversions, and nothing to synchronise. Every source and load meets on one DC bus, so there is no frequency, no phase angle and no reactive power to manage — bus voltage alone carries the power-balance information, which is what makes DC droop control so simple. The hard parts move elsewhere: interrupting DC fault current with no natural zero crossing, and keeping the bus stable against constant-power loads, whose negative incremental impedance actively works against the source filters.
AC microgrid — mature equipment, harder control AC bus grid PCC PV Battery Genset / CHP AC loads DC loads
Everything already exists — protection, switchgear, transformers, motors. The cost is that power balance now shows up as frequency and voltage, and every source has to agree on both. At least one converter must be grid-forming — setting the voltage and frequency the others follow — and the static transfer switch at the PCC has to island and resynchronise without the phase step that trips everything downstream.
Hybrid AC/DC microgrid — the interlinking converter is the design AC sub-grid 400 V / 50 Hz DC sub-grid ±380 V Interlinking converter bidirectional power sharing Genset / CHP AC loads Wind Utility grid PCC PV Battery EV / DC loads
Let each element sit on the bus it natively belongs to. PV, batteries and DC fast charging are DC by nature; motors, transformers and existing distribution are AC by nature — so a hybrid architecture stops paying conversion stages just to force everything onto one bus. The whole cost concentrates in one place: the interlinking converter. It has to share power between the two sub-grids, hold each one's droop characteristic, and decide which side loses support first during a fault or an island.

What DC gets right

PV, batteries, fuel cells, LED lighting, variable-speed drives and EV charging are all natively DC. Putting them on a DC bus removes a conversion stage each way, and with it the losses and the hardware.

There is no frequency, no phase angle, no reactive power and no synchronisation. Power sharing needs one variable — bus voltage.

What DC makes hard

DC fault current has no natural zero crossing, so interruption needs solid-state or hybrid breakers and deliberate fault-current limiting, with selectivity designed rather than inherited.

Tightly regulated converters look like constant-power loads. Their negative incremental impedance destabilises the bus unless source impedance and bus capacitance are designed against it.

Where AC still wins

Protection practice, switchgear, transformers, metering and motor loads all already exist and are certified. Voltage transformation is a passive component.

The cost is control: frequency and voltage both carry power balance, at least one converter has to be grid-forming, and islanding and reconnection have to happen without a phase step.

Control and operation

A microgrid has no infinite bus to lean on. In island mode the converters themselves have to set voltage and frequency, share load without fighting each other, and hand back to the utility cleanly when it returns.

Droop: power sharing without communication Vnom Vmin P Vbus slope = −Rd DC — voltage droop V = Vnom − Rd P 50 Hz fmin P f slope = −mp AC — frequency droop f = f0 − mp P primary droop after secondary restoration
Droop trades a steady-state error for autonomy. Each converter watches only a local quantity — bus voltage on DC, frequency on AC — and shares power in proportion to its own slope, with no communication at all. The price is the deviation itself: more power always means lower voltage or lower frequency. Secondary control then shifts the whole characteristic back up to nominal (dashed).
Hierarchical control: responsibility split by timescale Tertiary — dispatch minutes–hours economic optimisation · grid exchange · state of charge Secondary — restoration seconds removes the droop deviation · resynchronisation Primary — droop and inner loops milliseconds local and autonomous · power sharing · grid-forming or following setpoints down faster and more autonomous at the bottom, slower and more global at the top
The layers are decoupled by timescale, and that is the whole design. Primary control has to act within milliseconds, so it may use only local measurements and cannot wait for a communication link. Secondary control pulls the deviation back out over seconds. Tertiary control decides over minutes what to buy, sell, charge or discharge. Putting a responsibility on the wrong layer — a communication-dependent loop doing work that should be locally autonomous — is the most common way a microgrid control scheme fails.

Grid-forming vs grid-following

A grid-following converter needs an existing voltage to lock a PLL onto — perfect when the utility is present, useless the instant it is not. A grid-forming converter behaves as a voltage source behind an impedance and establishes the reference itself. An islandable microgrid needs at least one, and usually needs them to share.

Islanding and resynchronisation

Detecting the island, transferring without a load interruption, then matching voltage, phase and frequency before reclosing. Passive detection has a non-detection zone; active methods inject a small perturbation. The transfer is where most commissioning problems appear.

Inertia that is not there

A converter-dominated microgrid has almost no rotating inertia, so the rate of change of frequency after a disturbance is high. Virtual synchronous machine and synthetic inertia schemes emulate it, trading storage headroom and bandwidth for a slower, more forgiving frequency response.

Protection with limited fault current

Converters cannot supply the multiples of rated current that overcurrent protection assumes. Islanded settings differ from grid-connected ones, and coordination often needs directional, differential or communication-assisted schemes rather than time grading.

Power quality and harmonics

Switching converters supply the harmonics, and weak islanded networks have high grid impedance, so distortion and resonance interact. Filter design and converter output impedance shaping stop being independent problems.

Energy management

Above the fast loops, the dispatch problem: forecast, state of charge, degradation cost, tariffs, and the exchange schedule with the utility. This layer decides economics; the layers below decide whether the system stays up.

The solid-state transformer — a power router

An SST is the clearest answer to the hybrid question. Instead of a line-frequency transformer plus separate converters bolted around it, one multistage device provides the medium-voltage connection, the isolation and every port the microgrid needs — with controllable power flow between them.

The solid-state transformer as a power router Solid-state transformer MV front end isolated, HF LV output DC link DC link MV AC grid LV AC loads DC bus / storage DC port straight off the link
One device, several ports, and control over which way power goes. A line-frequency transformer changes voltage passively by turns ratio and does nothing else. The three stages of an SST give medium-voltage connection, high-frequency isolation and low-voltage output — and the middle DC link is already a DC port, so PV and storage can hang off it directly and skip an entire conversion stage. From the microgrid's point of view it is not a transformer at all; it is a routing node that decides where power goes.

What it replaces

The line-frequency transformer. Isolation moves to a high-frequency transformer inside the isolated DC–DC stage, which is why the magnetics shrink by an order of magnitude in volume.

What it adds

Bidirectional power flow, voltage regulation independent of turns ratio, a native DC port at the internal link, reactive support and harmonic control at the MV side, and ride-through behaviour that a passive transformer cannot offer.

What it costs

Series-connected medium-voltage cells, isolation coordination, cell-voltage balancing, distributed control, and efficiency that must beat a 99 %-efficient passive component to justify itself.

The full series works through it stage by stage: applications and motivation · architecture · AC–DC front end · DAB isolation stage · DC–AC output stage · modular integration and testing.