Renewable Energy, Storage, Microgrids and Grid-Forming Control

Introduction and Context

Electric power systems are changing from predominantly synchronous-generator networks into heterogeneous cyber-physical systems containing photovoltaic (PV) arrays, wind turbines, battery energy-storage systems (BESSs), power-electronic loads, electric vehicles, and controllable demand. Power electronics is the enabling technology: it converts variable renewable energy into grid-compatible electricity, regulates storage, decouples generation from grid frequency, and allows local microgrids to operate either connected to or separated from the utility.

This transformation creates a fundamental control problem. A conventional synchronous generator contributes rotating inertia, synchronizing torque, voltage support, fault current, and frequency-sensitive governor response. A conventional grid-following converter instead measures an existing grid voltage through a phase-locked loop (PLL) and injects commanded current. When converter penetration becomes high, the assumption of a strong voltage waveform may no longer be valid. Grid-forming (GFM) control addresses this challenge by commanding converter voltage magnitude and angle so that the converter can establish or support the electrical waveform.

The subject therefore combines energy conversion, control theory, network dynamics, protection, optimization, and operational governance. A technically successful renewable plant is not merely efficient; it must also be stable, dispatchable within its limits, secure, interoperable, and compliant with grid codes.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Explain the power-electronic interfaces used by solar, wind, storage, and microgrid resources.
2. Quantify renewable power, battery state of charge, round-trip efficiency, and converter ratings.
3. Distinguish grid-following, grid-supporting, and grid-forming control.
4. Analyze microgrid operating modes, transitions, protection, and energy management.
5. Select suitable storage technologies for technical and economic requirements.
6. Evaluate a practical microgrid using performance, reliability, safety, and regulatory criteria.
7. Identify research challenges associated with low-inertia systems and inverter-based resources.

1. Renewable-Energy Conversion Interfaces

1.1 Photovoltaic generation

A PV cell is a semiconductor p–n junction whose current depends on irradiance, temperature, and terminal voltage. Cells are connected in series and parallel to form modules and arrays. Since the maximum-power point changes continuously, a DC–DC converter commonly performs maximum-power-point tracking (MPPT), while a DC–AC inverter supplies the grid or local AC bus.

The instantaneous electrical power is

where P is power in watts, V is terminal voltage in volts, and I is current in amperes. For an array with Nₛ series modules and Nₚ parallel strings, ideal voltage and current scale approximately as V_array ≈ NₛV_module and I_array ≈ NₚI_module. Practical values are reduced by wiring, temperature, mismatch, conversion, and degradation losses.

A simplified energy estimate for a period is

where P_r is rated DC power, H_POA is equivalent peak-sun-hours on the plane of array, PR is the performance ratio, and E_PV is energy. The formula assumes that H_POA is expressed in hours at 1 kW/m² and that PR aggregates temperature, mismatch, soiling, wiring, availability, and conversion losses. For example, a 1 MW plant receiving 5.2 peak-sun-hours per day with PR = 0.82 produces approximately 4.264 MWh per day under the stated conditions.

1.2 Wind-energy conversion

Wind power depends strongly on wind speed:

where ρ is air density in kg/m³, A is rotor swept area in m², v is wind speed in m/s, and C_p is the power coefficient. The equation describes aerodynamic power extracted from the wind, not necessarily electrical output. Betz’s law limits an ideal actuator disk to C_p ≤ 16/27 ≈ 0.593; real turbines operate below this limit. Gearboxes and generators may be omitted in direct-drive architectures.

Wind turbines commonly use doubly fed induction generators or full-scale converter interfaces. A full converter electrically decouples generator frequency from grid frequency and provides broad control of active and reactive power. This flexibility is useful for fault ride-through and ancillary services, but the converter and its thermal system must be rated for the required operating envelope.

1.3 Converter architectures

Important architectures include:

The DC-link capacitor provides short-term energy buffering. Its energy is

where C_dc is capacitance in farads and V_dc is DC-link voltage in volts. This stored energy is small compared with battery energy, but it helps decouple fast source-side and grid-side power variations. DC-link control must respect semiconductor voltage ratings, capacitor ripple-current limits, and permissible voltage excursions.

2. Energy Storage: Models, Sizing and Applications

2.1 Why storage is required

Renewable generation is variable, while demand and network constraints are not. Storage can shift energy, smooth ramps, provide frequency response, support black start, reduce curtailment, defer network upgrades, and maintain critical loads during outages. No single technology optimizes all services.

A useful classification is:

2.2 State of charge and efficiency

State of charge (SOC) is the available stored energy relative to usable rated energy. A simplified discrete model is

where SOC is dimensionless, P_c and P_d are charging and discharging powers in watts, η_c and η_d are charge and discharge efficiencies, ∆t is the time step in hours if power and energy units are consistent, and E_n is nominal usable energy. Charging and discharging are normally mutually exclusive in an operational model. SOC must remain within prescribed minimum and maximum limits, for example SOC_min ≤ SOC ≤ SOC_max.

Round-trip efficiency is

where E_in is energy delivered to storage during charging and E_out is energy recovered during discharge. If η_c = 0.96 and η_d = 0.96, the idealized conversion-only round-trip efficiency is η_rt = η_cη_d = 0.9216, or 92.16%, before auxiliary and standby losses.

Battery ageing arises from calendar effects, temperature, average SOC, depth of discharge, current rate, and cycle count. A technically valid dispatch algorithm therefore includes temperature limits, current limits, thermal management, cell balancing, state-of-health estimation, and fault isolation. SOC estimation may combine coulomb counting with an observer or equivalent-circuit model; coulomb counting alone accumulates sensor and model errors.

2.3 Storage sizing example

Suppose a critical load requires 500 kW for 4 hours. The required delivered energy is 2 MWh. If the design permits a usable SOC window of 80%, inverter and battery discharge efficiency is 95%, and a 10% design margin is adopted, nominal battery energy is approximately

Thus, a nominal battery of about 2.9 MWh is required, subject to manufacturer limits, temperature derating, ageing reserve, and auxiliary consumption. The inverter must also sustain at least 500 kW, with additional capability if motor starting, fault support, or transient overload is required.

3. Microgrids and Their Operating Modes

A microgrid is a bounded electrical system containing loads and distributed energy resources, with defined electrical points of common coupling (PCC), capable of coordinated operation in grid-connected and islanded modes. Its assets may include renewable generation, storage, dispatchable generators, controllable loads, protection equipment, and supervisory controls.

3.1 Hierarchical control

A practical hierarchy is:

• Primary control: fast local voltage, current, frequency, and power-sharing response; it acts without depending on high-bandwidth communications.
• Secondary control: restores voltage and frequency deviations caused by primary droop and coordinates multiple inverters.
• Tertiary control: optimizes import/export, fuel use, storage, reserve, and economic dispatch while observing network constraints.

For active-power frequency droop, a common relationship is

where ω is commanded angular frequency, ω₀ is nominal angular frequency, P is measured active power, P₀ is a reference operating point, and m_p is the frequency-droop coefficient in radians per second per watt. For reactive-power voltage droop,

where V is voltage magnitude, Q is reactive power, and n_q is the voltage-droop coefficient. Droop enables decentralized sharing but intentionally permits steady-state deviations; secondary control removes these deviations.

Power sharing is affected by line impedance, voltage measurement location, converter rating, droop slopes, and saturation. Pure droop does not guarantee proportional sharing in low-voltage networks with unequal feeder impedances. Virtual impedance, power-line compensation, adaptive droop, or network-aware secondary control may be required.

3.2 Grid-connected and islanded operation

In grid-connected mode, the utility generally establishes voltage and frequency, while the microgrid controls power exchange at the PCC. In islanded mode, at least one voltage-forming resource must establish the AC waveform. The transition must be coordinated with anti-islanding protection, breaker status, phase alignment, load shedding, and energy reserves.

A seamless transition requires that the islanded voltage reference be synchronized with the utility before opening the PCC breaker, or that a controlled resynchronization process be used before reconnection. Reconnection requires acceptable voltage magnitude, frequency, phase angle, phase sequence, and slip rate, according to applicable interconnection rules.

3.3 Microgrid energy management

An energy-management system schedules resources over a horizon using forecasts and constraints. A simplified objective is

where the sum is over time intervals t; C terms are costs, P_grid is grid import, P_fuel is dispatchable generation, P_b is battery power, and P_curt is curtailed renewable power. This objective is illustrative: a production system may include reserve value, start-up costs, demand charges, emissions, uncertainty, and reliability penalties. Power balance must hold at each interval:

All terms must use the same sign convention and units. Optimization does not replace protection or fast control; it supplies slower set points to devices whose local controls enforce electrical safety.

4. Grid-Following and Grid-Forming Control

4.1 Grid-following control

A grid-following inverter behaves principally as a controlled current source. A PLL estimates grid angle θ, transformations convert measured quantities to a rotating reference frame, and inner current controllers regulate d- and q-axis currents. Under suitable conditions, active power is controlled mainly through the d-axis current and reactive power through the q-axis current, although exact signs depend on the chosen reference convention.

Grid-following control performs well in strong grids with a stable voltage waveform. Weak grids create difficulties: PLL interaction with network impedance, voltage distortion, control coupling, delayed angle estimation, and loss of synchronism. The short-circuit ratio is often used as an indicative strength metric, but it is not a complete stability measure, especially in multi-inverter networks.

4.2 Grid-forming control

A grid-forming inverter regulates an internal voltage magnitude and angle and exchanges current according to network conditions. It can establish voltage in an island, share load with other forming units, and provide fast frequency response. Common implementations include:

• Droop-based GFM: frequency and voltage references vary with measured P and Q.
• Virtual synchronous-machine control: emulates selected inertial and damping behavior.
• Matching or power-synchronization control: relates converter frequency to DC-side or power variables.
• Direct voltage control: regulates the voltage waveform through cascaded voltage and current loops.

A simplified virtual inertia equation is

where M is a virtual inertia coefficient, P_ref is commanded input power, P_e is electrical output power, D is a damping coefficient, and ω₀ is nominal frequency. This is a control law, not a claim that the converter contains physical rotating inertia. Its validity is limited by stored energy, current limits, measurement delay, and the controller’s ability to maintain stability.

A GFM inverter cannot supply unlimited fault current. Semiconductor current limits may cause voltage reduction, current prioritization, or control-mode transitions during faults. Consequently, protection schemes designed for high-current synchronous sources may fail to detect or selectively clear inverter-fed faults.

4.3 Current limiting and overload management

Current limiting is central to GFM design. During a severe voltage sag, the controller may prioritize active current, reactive current, negative-sequence current, or voltage support according to grid-code and protection requirements. The selected priority changes real-power delivery, DC-link behavior, thermal stress, and transient stability. Limiters should be designed as part of the nonlinear control system rather than appended after nominal-loop design.

5. Protection, Stability and Power Quality

Converter-dominated networks require coordinated analysis across several time scales:

Important performance measures include voltage deviation, frequency nadir, rate of change of frequency, harmonic distortion, flicker, unbalance, availability, curtailed energy, unserved energy, efficiency, and battery degradation. Harmonic compliance must be evaluated at the point of common coupling and under relevant operating combinations, not only at nominal loading.

Protection must address islanding, overcurrent, differential faults, ground faults, reverse power, loss of synchronism, arc flash, and communications failure. Adaptive protection, traveling-wave methods, differential protection, and converter-aware relays may be necessary. Cybersecurity is part of protection: compromised set points or communications can create unsafe electrical states even when hardware is functioning correctly.

6. Applied Case: Illustrative Remote Health-Care Microgrid

Background

Consider an illustrative islanded-capable microgrid serving a remote health-care facility. The critical load is 300 kW, the average daily energy demand is 4.8 MWh, and the site has 500 kW of PV. A diesel generator rated at 400 kVA provides extended backup. The design objective is to reduce fuel use while maintaining vaccine refrigeration, medical equipment, communications, and emergency lighting during utility outages.

Problem

PV output varies rapidly with cloud cover. The diesel generator cannot operate efficiently at very low loading, while the critical load must not experience unacceptable voltage or frequency excursions. The system must also black-start after a prolonged outage and reconnect safely when the utility returns.

Analysis and approach

A 1.5 MWh nominal lithium-ion battery is selected initially. With a usable SOC window of 80% and a discharge path efficiency of 95%, usable delivered energy is approximately

This supports a 300 kW critical load for approximately 3.8 hours if no PV contribution exists and auxiliary losses are neglected. Since the design requirement is not necessarily to supply the full average load, noncritical loads are shed through a prioritized load-management scheme.

A 500 kVA bidirectional inverter operates as GFM during islanding. PV inverters operate mainly as grid-following units when the GFM inverter establishes the voltage. The supervisory controller maintains a minimum SOC reserve for forecast uncertainty and starts the diesel generator when SOC, predicted irradiance, or load reserve reaches a threshold. During black start, the battery inverter energizes the bus, sequentially connects essential loads, and then permits PV synchronization.

Outcome

The architecture permits renewable utilization during normal operation, fast battery response during cloud transients, and diesel operation near an efficient loading range during long deficits. The expected benefits are lower fuel consumption, reduced starts, and improved outage resilience. Commissioning must nevertheless verify fault clearing, neutral and grounding behavior, generator–inverter interactions, harmonic performance, transfer switching, and safe reconnection.

Learning points

1. Storage energy and inverter power are separate design quantities.
2. GFM capability is essential for island formation but does not remove current limits.
3. Load prioritization may be more economical than oversizing storage.
4. Forecast uncertainty requires SOC reserve and dispatch flexibility.
5. Protection and commissioning are as important as energy calculations.

7. Design Methodology and Best Practices

A robust design process should:

1. Define critical loads, operating modes, PCC requirements, and unacceptable events.
2. Characterize renewable profiles, load diversity, fault levels, ambient conditions, and communications.
3. Select converter topology and control mode according to grid strength and islanding requirements.
4. Size both energy and power, including degradation, temperature, reserve, overload, and auxiliary loads.
5. Simulate steady state, RMS dynamics, EMT transients, faults, black start, reconnection, and communications loss.
6. Specify interoperable interfaces, time synchronization, logging, and cybersecurity controls.
7. Validate controllers using controller-hardware-in-the-loop or power-hardware-in-the-loop where risk justifies it.
8. Commission progressively: factory tests, site tests, staged energization, island tests, protection tests, and emergency procedures.

AI and data-driven methods can improve irradiance forecasting, anomaly detection, battery state estimation, and predictive maintenance. They should be deployed with uncertainty bounds, human oversight, secure data pipelines, explainable alarms, and fail-safe fallback controls. An AI prediction must never be allowed to bypass deterministic protection or operating limits.

8. Challenges, Governance and Future Directions

Major challenges include weak-grid stability, uncertain protection behavior, converter interactions, black-start coordination, battery degradation, supply-chain constraints, recycling, fire safety, and inconsistent interconnection requirements. Grid codes increasingly specify fault ride-through, reactive current injection, frequency response, voltage support, active-power ramp rates, and model validation. Engineers must therefore understand both technical standards and the authority having jurisdiction.

Environmental and ethical evaluation should include the full life cycle of batteries and power-electronic equipment, mineral sourcing, worker safety, end-of-life recovery, land use, noise, and distributional impacts. Resilience investments should not systematically exclude rural or low-income communities. Governance must define responsibility for firmware updates, remote access, model changes, data ownership, and emergency control authority.

Research directions include multi-inverter stability theory, reduced-order yet validated models, grid-forming behavior under severe faults, coordinated transmission–distribution control, long-duration storage, solid-state transformers, wide-bandgap semiconductors, and secure autonomous microgrid operation. The central research challenge is to obtain fast decentralized response without sacrificing system-wide coordination and verifiable safety.

Chapter Summary

Renewable resources require power-electronic interfaces because their natural electrical characteristics do not directly match grid requirements. PV and wind converters regulate energy capture, DC-link voltage, active power, reactive power, and grid interaction. Storage adds temporal flexibility, but its design must distinguish power rating from energy capacity and must account for SOC limits, efficiency, ageing, thermal management, and safety.

Microgrids coordinate distributed resources and loads across grid-connected and islanded modes. Primary droop control supports local sharing, secondary control restores voltage and frequency, and tertiary management optimizes economic and reliable operation. Grid-following inverters depend on an existing waveform, whereas grid-forming inverters establish or support voltage and frequency. GFM controls improve operation in weak grids and islands but remain constrained by stored energy, semiconductor current limits, protection requirements, and nonlinear interactions.

Successful deployment requires integrated electromagnetic, dynamic, optimization, protection, cybersecurity, regulatory, and life-cycle analysis. Simulation, hardware validation, careful commissioning, and transparent governance are essential. The future power system will not be defined by renewable generation alone, but by how intelligently converters, storage, loads, networks, and human institutions are coordinated.

Key Terms

Renewable-energy converter: Power-electronic interface that converts variable renewable output into usable electrical power.

Maximum-power-point tracking: Control that operates a PV array near its instantaneous maximum-power point.

Performance ratio: Ratio representing the actual PV energy yield relative to an ideal reference after aggregated losses.

Power coefficient: Fraction of wind aerodynamic power extracted by a turbine rotor.

DC link: Intermediate DC electrical stage connecting source-side and grid-side converters.

State of charge: Stored usable energy expressed relative to a specified energy capacity.

Round-trip efficiency: Ratio of energy recovered from storage to energy supplied to it.

Depth of discharge: Fraction of available battery capacity removed during a discharge.

Microgrid: Bounded electrical system capable of coordinated grid-connected and islanded operation.

Point of common coupling: Electrical point where a facility or microgrid connects to the utility.

Droop control: Local control law that deliberately relates frequency or voltage to active or reactive power.

Grid-following inverter: Converter that synchronizes to and injects controlled current into an existing grid voltage.

Grid-forming inverter: Converter that regulates an internal voltage waveform and can establish grid voltage and frequency.

Virtual inertia: Control-emulated inertial response based on measured power and frequency dynamics.

Primary control: Fast local control responsible for immediate stabilization and sharing.

Secondary control: Coordinating control that restores frequency and voltage references.

Tertiary control: Supervisory optimization of dispatch, reserves, and power exchange.

Fault ride-through: Ability to remain connected and support the grid during specified voltage or frequency disturbances.

Curtailment: Deliberate reduction of available renewable generation.

Black start: Restoration of an electrical system without relying on an energized external grid.

Review Questions

Short Answer Questions

1. Why is an MPPT controller used in a PV system?
2. Distinguish battery power rating from battery energy rating.
3. What is the principal difference between grid-following and grid-forming control?
4. Why does primary droop control produce steady-state frequency deviation?
5. State two technical challenges caused by high penetration of inverter-based resources.

Long Answer Questions

1. Explain the architecture and control functions of a PV-plus-storage microgrid.
2. Derive and interpret the simplified SOC update equation, including its assumptions and constraints.
3. Compare lithium-ion, flow-battery, pumped-hydro, and hydrogen storage for grid applications.
4. Discuss primary, secondary, and tertiary microgrid control with reference to power sharing and economic dispatch.
5. Explain why protection and fault-current behavior must be reconsidered in grid-forming, converter-dominated networks.

Analytical / Application Questions

1. A critical load of 250 kW must be supplied for 6 hours. The battery usable SOC window is 75%, discharge efficiency is 92%, and a 15% design margin is required. Calculate the minimum nominal battery energy, neglecting auxiliary losses, and explain additional factors that could increase the selected rating.
2. Design a control and operating sequence for a campus microgrid that must transition from grid-connected operation to islanded operation and later resynchronize. Identify measurements, control layers, protection actions, and failure contingencies.
3. A weak-grid PV plant exhibits oscillatory power and PLL angle behavior after a nearby fault. Propose a structured investigation using network-strength assessment, EMT simulation, controller review, hardware-in-the-loop testing, and possible grid-forming remedies.

Multiple Choice Questions with Answers

Question 1

For a PV array rated at 1 MW with 5.2 peak-sun-hours per day and a performance ratio of 0.82, what is the approximate daily energy production?

A. 2.46 MWh

B. 4.26 MWh

C. 5.20 MWh

D. 6.34 MWh

Correct Answer: B. 4.26 MWh

Question 2

In the wind-power equation P_w = ½ρAv³C_p, which variable has the strongest idealized dependence on wind speed?

A. Air density

B. Swept area

C. Wind speed

D. Power coefficient

Correct Answer: C. Wind speed

Question 3

If charge and discharge efficiencies are each 0.96, what is the conversion-only round-trip efficiency?

A. 0.4800

B. 0.9216

C. 0.9600

D. 1.9200

Correct Answer: B. 0.9216

Question 4

What is the principal function of a grid-forming inverter during islanded operation?

A. Only measure utility frequency

B. Establish and regulate the local voltage waveform

C. Disconnect all renewable generation

D. Operate exclusively as a passive rectifier

Correct Answer: B. Establish and regulate the local voltage waveform

Question 5

Which control layer normally restores frequency and voltage deviations caused by primary droop?

A. Switching control

B. Primary control

C. Secondary control

D. Mechanical protection

Correct Answer: C. Secondary control

Question 6

A battery has a nominal energy rating of 1.5 MWh, a usable SOC window of 80%, and discharge efficiency of 95%. Neglecting auxiliaries, how much energy can it deliver?

A. 0.80 MWh

B. 1.14 MWh

C. 1.50 MWh

D. 1.88 MWh

Correct Answer: B. 1.14 MWh

Question 7

Why can a conventional overcurrent protection scheme be ineffective for some inverter-based resources?

A. Inverters always produce infinite fault current

B. Inverters may limit fault current to values near normal operating current

C. Inverters cannot measure voltage

D. Inverters eliminate all electrical faults

Correct Answer: B. Inverters may limit fault current to values near normal operating current

Question 8

What is the main purpose of a DC-link capacitor in a converter system?

A. Provide long-term seasonal energy storage

B. Provide short-term energy buffering between converter stages

C. Replace all semiconductor switches

D. Eliminate the need for control systems

Correct Answer: B. Provide short-term energy buffering between converter stages

Question 9

Which statement best describes grid-following control?

A. It establishes an independent voltage waveform without measuring the grid

B. It synchronizes to an existing voltage waveform and controls injected current

C. It operates only mechanical generators

D. It cannot exchange reactive power

Correct Answer: B. It synchronizes to an existing voltage waveform and controls injected current

Question 10

Which factor is especially important when sizing a battery beyond the basic energy calculation?

A. Ignoring temperature because batteries are ideal

B. Ageing, SOC limits, auxiliary consumption, and reserve requirements

C. Using only the inverter nameplate power

D. Assuming renewable output is constant

Correct Answer: B. Ageing, SOC limits, auxiliary consumption, and reserve requirements