Grid Integration, Power Quality and Smart Grids

Introduction and context

Electric power systems were traditionally designed around large synchronous generators connected to transmission networks and passive distribution loads. Solar photovoltaic (PV), wind, battery storage, electric vehicles, flexible loads and digital controls are changing that architecture. Renewable generators are often geographically dispersed, connected at distribution voltage, interfaced through power electronics and governed by variable weather conditions. Their integration is therefore not simply an energy-supply problem; it is a power-system planning, protection, control, market and communications problem.

Grid integration means connecting a resource so that it delivers energy while maintaining acceptable voltage, frequency, stability, power quality, protection performance and operational security. A connection that is technically possible at one operating point may become unacceptable when several plants operate simultaneously, demand is low, a transmission line is out of service or a fault occurs. The maximum additional generation that can be connected without violating defined limits is commonly called hosting capacity.

This chapter examines the principal technical impacts of renewable resources, the functions of grid codes, operational tools such as forecasting and curtailment, and the architecture of smart grids. The emphasis is on analysis suitable for electrical engineers who must evaluate connection studies, specify controls, interpret power-quality measurements and participate in system planning.

Learning objectives

By the end of this chapter, a student should be able to:

1. Explain how inverter-based renewable generation affects voltage, frequency, stability, fault current and harmonics.
2. Apply basic voltage-drop, short-circuit, power-factor and hosting-capacity reasoning.
3. Describe grid-code requirements, forecasting, curtailment and ancillary services.
4. Compare demand response, virtual power plants, smart meters and advanced distribution-management systems.
5. Evaluate protection, transmission and distribution planning issues in renewable-rich networks.
6. Identify limitations, governance risks and good engineering practice in smart-grid deployment.

1. The changing grid architecture

A modern grid contains four interacting layers:

Synchronous generators inherently store kinetic energy and contribute short-circuit current. Most PV plants and many modern wind turbines are inverter-based resources (IBRs). An IBR does not directly impose an electromechanical voltage waveform; it measures the grid and synthesizes current through a converter. In grid-following control, a phase-locked loop tracks an existing voltage waveform. In grid-forming control, the converter regulates an internal voltage angle and magnitude and can establish or support a waveform, subject to its energy source and current limits.

The distinction is important. A strong grid has high short-circuit strength relative to the converter rating, so phase tracking is generally easier. A weak grid has larger voltage changes for a given current and may produce control interactions, resonance or instability. A useful screening index is the short-circuit ratio:

where S_sc is the three-phase short-circuit apparent power at the point of common coupling (PCC), in VA or MVA, and S_r is the rated apparent power of the plant. SCR is dimensionless. It is a screening measure rather than a complete stability criterion: it does not by itself represent network impedance angles, nearby converters, controls or harmonic resonances.

2. Voltage behaviour and reactive power

2.1 Voltage rise in distribution feeders

A distribution feeder with resistance and reactance experiences a voltage change when real and reactive power flow through it. For a balanced three-phase feeder, a commonly used approximate relation is:

Here, ∆V is the line-to-line voltage-magnitude change in volts, R and X are the total positive-sequence resistance and reactance in ohms, P and Q are three-phase real and reactive power flows in watts and var, and V is the nominal line-to-line voltage in volts. The approximation assumes a small angle and small voltage difference, a balanced system and a specified sign convention. In per-unit form, the same relationship is approximately ∆V_pu ≈ R_puP_pu+X_puQ_pu.

For a feeder exporting PV power, the sign of P reverses relative to normal load flow. Because distribution feeders often have appreciable resistance, active-power export can raise the local voltage. In underground or high-voltage networks, reactance may dominate and reactive-power control becomes particularly influential.

Worked example. A three-phase 400 V feeder has total R=0.08 Ω and X=0.06 Ω. A PV system exports P=60 kW at unity power factor, so Q=0. The approximate voltage rise is

The approximate rise is 12 V, or 3.0% of 400 V. A detailed study would use phase-domain or load-flow modelling, conductor temperature, unbalance, transformer tap position and minimum-load conditions.

2.2 Reactive-power support

An inverter can absorb or supply reactive power within its apparent-power rating. The capability relationship is:

where S is apparent power in VA, P is real power in W and Q is reactive power in var. If a 100 kVA inverter produces 80 kW, its theoretical reactive capability, ignoring thermal and control limits, is

Power factor is

for sinusoidal voltage and current, with displacement and distortion effects treated consistently. Grid codes may require fixed power factor, a volt-var curve, watt-var control, or voltage-reactive droop. Controls must include dead bands and ramp limits to avoid oscillatory interaction between many devices.

2.3 Voltage unbalance and flicker

Single-phase rooftop PV distributed unevenly across phases can produce negative-sequence voltage. Unbalance causes additional heating and torque pulsations in rotating machines. Rapid changes in irradiance, wind or converter output may cause voltage fluctuations and flicker, especially on weak feeders. Mitigation includes phase balancing, feeder reinforcement, reactive support, energy storage, inverter ramp-rate control and coordinated voltage-regulator settings.

3. Frequency, inertia and stability

The instantaneous balance between generation and demand determines frequency. A simplified aggregate swing equation for a coherent synchronous area is:

where H is the inertia constant in seconds, ω_s is synchronous angular frequency in rad/s, ω is system angular frequency, P_m is mechanical input power, P_e is electrical output power, and D represents frequency-sensitive damping. The equation is a simplified model, normally expressed on a common power base. It applies to small deviations and does not replace a multi-machine dynamic model.

When conventional units are displaced, aggregate physical inertia may fall. A given loss of generation can then produce a larger rate of change of frequency (RoCoF). However, inverter-based resources can provide fast frequency response by changing active power, if headroom, energy storage and controls are available. A PV plant operating at maximum available power cannot provide sustained upward response unless it is deliberately curtailed or paired with storage.

Frequency services include:

• Inertial or synthetic response: rapid power injection or absorption following a frequency change.
• Primary frequency response: autonomous response through a droop characteristic.
• Secondary regulation: central or automatic restoration of frequency and interchange schedules.
• Tertiary reserve: slower redispatch and replacement of used reserves.

A frequency-droop relation can be written as

where R is the droop coefficient in Hz per unit power, ∆f is frequency deviation and ∆P is the power response. Actual implementations impose dead bands, saturation, response times and energy limits.

Small-signal stability concerns whether small disturbances decay. Transient stability concerns the ability to remain in synchronism after severe disturbances. Voltage stability concerns whether acceptable voltage can be maintained as loading and reactive demand change. Converter controls add fast dynamic modes, so stability studies must include control models, current limits, phase-locked loops, plant controllers and interactions with network impedance.

4. Fault levels and protection

4.1 Fault-current characteristics

Synchronous generators can supply several times rated current for cycles, with a decaying subtransient and transient response. Inverters commonly limit current to approximately a controlled multiple of rated current and may change the phase of that current to support voltage. Their fault contribution is therefore smaller, controlled and dependent on firmware and grid voltage.

For a balanced three-phase bolted fault, a basic short-circuit estimate is:

where I_sc is fault current in amperes, V_LL is prefault line-to-line voltage in volts and Z_th is the Thevenin positive-sequence impedance in ohms. The equation assumes a balanced fault, a linear Thevenin equivalent and RMS quantities. Asymmetrical peak current, DC offset, fault resistance, negative-sequence effects and inverter current limiting require more detailed analysis.

Renewables can both increase and decrease protection performance. Added generation may raise fault current at some locations, but an inverter's limited contribution may reduce fault-current magnitude from downstream faults compared with a synchronous plant. Reverse power flow can also change relay reach and coordination. Conventional overcurrent protection may become insensitive, directional elements may be necessary, and islanding detection must be coordinated with reconnection rules.

4.2 Protection principles

Protection design should consider:

1. Selectivity: only the faulted section should be disconnected where practical.
2. Sensitivity: the relay must detect the minimum credible fault current.
3. Speed: interruption must limit equipment damage and instability.
4. Dependability and security: the system should operate for internal faults but not for external disturbances.
5. Communications dependence: pilot and differential schemes require reliable, secure channels.

Anti-islanding protection is required when a distributed resource could energize a de-energized utility section. Passive methods monitor voltage, frequency and phase; active methods perturb operating conditions. No method is universally perfect, particularly when generation and load are closely matched. Intentional islanding, if permitted, requires dedicated controls, sectionalizing, grounding, black-start capability and clear operational authority.

5. Harmonics and power quality

Power quality includes voltage magnitude, frequency, waveform distortion, unbalance, rapid voltage changes, interruptions and transients. Inverters use switching devices and filters that can generate high-frequency components. Harmonics may also arise from transformers, motor drives and resonant interactions.

The voltage total harmonic distortion is commonly defined as

where V₁ is the RMS fundamental voltage and V_h is the RMS voltage component at harmonic order h. A current THD expression is analogous. The relevant limit depends on the applicable grid code or power-quality standard, the PCC and the measurement method; limits should not be assumed to be identical at every voltage level.

Resonance occurs when network inductance and capacitance exchange energy at a frequency near a harmonic component. Adding capacitor banks, cables or filters can shift resonance. Harmonic studies should therefore use frequency-dependent network impedance, converter emission spectra, background distortion, operating configurations and equipment tolerances. Mitigation includes tuned or broadband filters, active filtering, appropriate switching strategies, damping, phase distribution and limits on injected harmonic current.

6. Hosting capacity and connection assessment

Hosting capacity is not a universal constant. It depends on the planning criteria, technology, location, topology, operating assumptions and mitigation options. A screening process generally includes:

• thermal loading of lines, cables and transformers;
• steady-state voltage and regulator operations;
• reverse power flow and transformer tap constraints;
• fault levels and protection coordination;
• flicker, harmonics and unbalance;
• thermal and transient stability;
• communications, grounding and safety;
• normal, contingency and restoration conditions.

A deterministic hosting-capacity study tests defined scenarios, such as minimum load with maximum PV and maximum load with maximum export. Probabilistic studies represent weather, demand, outages and correlations. The result should state whether it is a thermal, voltage, protection or stability limit, because the remedy differs.

Possible mitigation measures include active-power curtailment, volt-var control, dynamic operating envelopes, reconductoring, transformer upgrades, on-load tap control, storage, network reconfiguration and flexible interconnection agreements. An apparently inexpensive mitigation must be evaluated over its lifetime, including lost energy, control complexity, maintenance and customer impacts.

7. Grid codes and interconnection requirements

A grid code is a set of technical requirements for users connected to a power system. It typically specifies:

• voltage and frequency operating ranges;
• fault ride-through and dynamic reactive-current response;
• active-power ramp rates and reconnection delays;
• reactive-power capability and voltage control;
• frequency response and active-power control;
• power-quality emissions;
• telemetry, models, testing and cybersecurity;
• performance during abnormal and islanded conditions.

Fault ride-through requirements prevent unnecessary tripping during short voltage disturbances. A plant may be required to remain connected and inject reactive current, subject to converter current limits. Compliance should be demonstrated through validated simulation models, factory tests, commissioning tests and monitored operational data. Grid codes are jurisdiction-specific; engineers must consult the current code of the relevant transmission or distribution system operator rather than rely on generic values.

8. Forecasting, dispatch and curtailment

Wind and solar forecasts support unit commitment, reserve procurement, congestion management and maintenance scheduling. Forecasts may be physical, statistical or hybrid. Useful outputs include point forecasts, prediction intervals, ramp-event probabilities and spatially aggregated forecasts.

Forecast error can be represented as

where y_t is observed production and ŷ_t is forecast production at time t. Mean absolute error is

where n is the number of observations. Errors should be normalized carefully, for example by plant capacity, and evaluated for different horizons and weather regimes. A forecast that is accurate on average may still fail to identify rare but operationally important ramps.

Curtailment is a reduction in available renewable output. It may be required for thermal congestion, voltage control, minimum synchronous generation, frequency security, system restoration or negative-price conditions. Fair rules require transparent priority, compensation or market arrangements where applicable, auditable dispatch instructions and consideration of distributed customers. Curtailment should not be treated as a substitute for all network investment: excessive curtailment can waste clean energy and weaken investment signals.

9. Ancillary services and flexibility

Ancillary services maintain reliability beyond the delivery of scheduled energy. Renewable plants, storage and aggregated loads can provide regulation, reserves, voltage support, black start or restoration assistance when technically capable. A service specification must define the response direction, activation time, duration, accuracy, availability, telemetry and recovery requirement.

Battery storage is particularly useful because it can respond rapidly, but its state of charge limits sustained service. Co-optimizing energy and reserves avoids promising the same capacity twice. Inverter-based resources can provide voltage support without producing real power, although reactive capability may vary with active-power output and thermal conditions.

10. Demand response and virtual power plants

Demand response changes electricity consumption in response to prices, incentives or direct control. It can be:

• direct load control, such as dispatching water heaters;
• price-based response, in which customers respond to tariffs;
• interruptible or contractual response, with defined availability;
• automated flexibility, coordinated through building or industrial controls.

A demand-response baseline estimates what consumption would have been without activation. Baseline error creates settlement and fairness problems. Customer comfort, industrial production, rebound consumption and privacy must be included in program design.

A virtual power plant (VPP) aggregates geographically dispersed PV, batteries, electric vehicles and flexible loads. Its aggregator forecasts availability, schedules resources and sends dispatch commands while respecting local network constraints. A VPP is not a physical generator; its deliverability depends on communication, customer participation, distribution capacity and forecast accuracy. Distribution-level locational constraints should be represented rather than assuming that all aggregated flexibility is interchangeable.

11. Smart meters and smart-grid control

Smart meters provide interval energy measurements and may support remote connection, outage detection, time-of-use tariffs and distributed-energy coordination. Phasor measurement units provide time-synchronized voltage and current phasors at higher reporting rates and are valuable for wide-area monitoring, although their cost and communications requirements differ from those of ordinary meters.

A modern control hierarchy may include:

Figure: A simplified smart-grid control hierarchy with measurement and feedback across levels.

The distribution management system (DMS) supervises feeders, switches, voltage regulators and outages. A distributed-energy-resource management system (DERMS) coordinates PV, storage and flexible devices. The energy-management system (EMS) operates the transmission system. Clear interfaces are needed to prevent conflicting commands, especially when a local voltage objective conflicts with a system-level reserve instruction.

Cybersecurity is a reliability requirement. Threats include unauthorized switching, false measurements, ransomware, supply-chain compromise and denial of service. Good practice includes least-privilege access, network segmentation, encryption, secure updates, event logging, anomaly detection, recovery exercises and tested manual fallback. Artificial intelligence can assist load forecasting, anomaly detection and predictive maintenance, but its outputs require validation, explainability appropriate to the risk, data-quality controls and human accountability. An opaque model should not be allowed to make an irreversible protection decision without defined safeguards.

12. Transmission and distribution planning

Transmission planning evaluates resource location, bulk transfers, contingencies, stability, congestion and interconnection queues. Renewable zones may be remote from load centres, requiring new lines, reconductoring, dynamic line ratings, phase-shifting transformers or controllable power flow. Planning should compare network expansion with storage, demand flexibility, alternative generation and expected curtailment.

Distribution planning historically forecast one-way peak demand and sized equipment accordingly. High DER penetration requires multi-directional power-flow analysis, minimum-load scenarios, phase imbalance analysis, hosting-capacity maps and coordination with transmission constraints. Asset decisions should consider electrification of transport and heating, climate-related hazards, resilience and equity.

A useful planning principle is to distinguish reliability, resilience and adequacy. Adequacy concerns having sufficient resources over time; reliability includes security against disturbances; resilience concerns preparation for, absorption of and recovery from low-probability high-impact events. Renewable integration plans should address all three rather than use annual energy as the only metric.

13. Illustrative case study: PV growth on a rural feeder

Background. A 12.47 kV rural feeder supplies 4 MW of minimum daytime load. A developer proposes 6 MW of PV at the remote end. The feeder currently has a 5 MW substation transformer and a conventional overcurrent protection scheme.

Problem. At minimum load, the proposed plant could export approximately 2 MW toward the substation. The utility is concerned about voltage rise, reverse power flow, relay coordination and loss of visibility during cloud-driven ramps.

Analysis. A preliminary load flow shows that the remote-bus voltage reaches the upper planning limit during maximum irradiance. A short-circuit study finds that inverter-limited fault current does not increase the upstream fault level enough to operate the existing remote relay reliably. Harmonic measurements on the feeder are acceptable before connection, but a frequency-scan study identifies a possible resonance after the plant's filter and cable capacitance are included. A contingency study also shows that an outage of the parallel feeder would overload the transformer during certain export conditions.

Approach. The interconnection package includes a plant controller with volt-var capability, a maximum export limit during specified transformer conditions, high-speed communications, directional protection and revised relay settings. The utility installs feeder voltage sensors, upgrades the substation transformer protection, validates the inverter model and requires a ramp-rate limit during abnormal network conditions. A battery is not initially installed because the export limit and network upgrade are less costly; the agreement allows future storage to provide additional flexibility.

Outcome. The project becomes connectable subject to staged commissioning. Operational data are used to verify voltage, harmonics, protection response and forecast performance. The export limit is reviewed after one year using measured feeder loading rather than being treated as permanent.

Learning points. Hosting capacity is constrained by several mechanisms simultaneously; inverter controls can solve voltage problems but cannot automatically solve protection or thermal problems; and a connection agreement should specify measurable performance, communications and review conditions.

14. Challenges, limitations and governance

Technical solutions have boundaries. Volt-var control may increase losses or cause control interactions. Curtailment can reduce revenue and renewable utilization. Batteries degrade and require fire-safety management. Advanced communications create cyber and privacy exposure. Forecasting improves decisions but cannot eliminate uncertainty. Dynamic ratings depend on weather sensors and conservative failure modes.

Regulatory governance must define who owns data, who may control customer equipment, how flexibility is compensated and how performance is verified. Smart-meter data can reveal occupancy patterns; collection should be limited to legitimate purposes, protected from unauthorized access and governed by transparent retention policies. Access to distributed-energy participation should not depend exclusively on expensive automation that excludes low-income customers. Tariffs and programs should address both system value and distributional effects.

Environmental and social assessment also matters. Renewable integration requires minerals, land, transmission corridors and end-of-life management. A technically optimal solution may be unacceptable if it ignores affected communities, indigenous rights, worker safety or lifecycle impacts.

15. Best-practice engineering workflow

1. Define the PCC, voltage level, operating modes and applicable grid code.
2. Collect validated network, equipment, weather, load and protection data.
3. Study normal operation, credible contingencies, minimum load, maximum generation and restoration states.
4. Perform load flow, short-circuit, harmonics, protection, dynamic and EMT studies at the level required by the risk.
5. State assumptions, uncertainty ranges, model versions and acceptance criteria.
6. Compare reinforcement, controls, storage, curtailment and flexible-connection alternatives over lifecycle cost and performance.
7. Validate models through factory, commissioning and field tests.
8. Monitor actual performance, review settings after network changes and maintain a secure change-management process.

Chapter Summary

Renewable integration changes voltage flow, frequency response, stability, fault-current behaviour, protection coordination and waveform quality. Distribution voltage is influenced by both real and reactive power, while inverter apparent-power capability limits simultaneous active and reactive support. Reduced synchronous generation can reduce physical inertia, but appropriately controlled inverters and storage can provide fast frequency response. Fault studies must represent current-limited inverter behaviour rather than assume synchronous-generator characteristics.

Hosting capacity is a location- and criterion-dependent result involving thermal, voltage, protection, harmonic and stability constraints. Grid codes specify operating, fault ride-through, reactive-power, frequency-response, communication and testing requirements. Forecasting, curtailment, ancillary services, demand response and VPPs provide operational flexibility, but each has uncertainty, energy or participation limits. Smart meters, ADMS, DERMS, secure communications and carefully governed data enable smart-grid coordination. Sound planning combines transmission and distribution analysis with reliability, resilience, equity, cybersecurity and lifecycle considerations.

Key Terms

Hosting capacity: Maximum DER capability connectable while defined technical limits remain satisfied.

Inverter-based resource: A generator or storage device connected through power-electronic conversion.

Point of common coupling: Electrical location where a customer or plant connects to the utility network.

Short-circuit ratio: Ratio of short-circuit apparent power to resource rated apparent power.

Grid-following inverter: Converter that synchronizes to an existing grid voltage waveform.

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

Reactive power: Component of AC power associated with energy exchange between electric and magnetic fields.

Voltage ride-through: Required continued operation during specified voltage disturbances.

RoCoF: Rate of change of system frequency after a disturbance.

Ancillary services: Services that support frequency, voltage, reserves, restoration or system security.

Fault ride-through: Ability to remain connected and meet specified response requirements during faults.

Harmonic distortion: Waveform deviation caused by components at integer multiples of the fundamental frequency.

Anti-islanding protection: Measures preventing unauthorized energization of an isolated network.

Demand response: Deliberate change in electricity consumption in response to a signal or control.

Virtual power plant: Coordinated aggregation of distributed generation, storage and flexible loads.

Smart meter: Meter capable of communicating interval measurements and receiving authorized commands.

DERMS: System for monitoring and coordinating distributed energy resources.

Curtailment: Reduction of otherwise available generation output.

Dynamic rating: Equipment rating adjusted using measured or forecast operating conditions.

Prediction interval: Range intended to contain a future outcome with a stated probability.

Review Questions

Short Answer Questions

1. Why can a PV plant cause voltage rise on a rural feeder?
2. Distinguish grid-following and grid-forming inverter control.
3. What is the purpose of fault ride-through?
4. Why does inverter fault current complicate overcurrent protection?
5. Name three functions of a DERMS.

Long Answer Questions

1. Explain the voltage, frequency and stability impacts of high renewable penetration.
2. Describe a complete hosting-capacity assessment and its possible mitigation measures.
3. Explain the role of grid codes in renewable interconnection and compliance testing.
4. Compare forecasting, curtailment, storage and demand response as flexibility tools.
5. Discuss cybersecurity, privacy and equity issues in smart-grid deployment.

Analytical / Application Questions

1. A 400 V feeder has R=0.05 Ω, X=0.04 Ω and a PV export of 50 kW at unity power factor. Estimate the voltage rise using the approximate three-phase relation and state its limitations.
2. A 10 MVA inverter-based plant is connected at a bus with 80 MVA short-circuit power. Calculate the SCR and discuss what additional information is needed before judging stability.
3. A feeder's proposed PV project passes its thermal study but fails protection sensitivity during a remote fault. Recommend and justify at least three technically distinct remedies.

Multiple Choice Questions with Answers

Question 1

Which condition most directly tends to increase voltage rise from PV export on a distribution feeder?

A. High feeder resistance and low local demand

B. High synchronous inertia

C. Low inverter apparent-power rating only

D. A larger short-circuit ratio

Correct Answer: A. High feeder resistance and low local demand

Question 2

A 100 kVA inverter produces 80 kW. Ignoring other limits, its maximum reactive-power magnitude is approximately:

A. 20 kvar

B. 36 kvar

C. 60 kvar

D. 180 kvar

Correct Answer: C. 60 kvar

Question 3

What is the principal purpose of fault ride-through requirements?

A. To force all plants to disconnect immediately during faults

B. To prevent unnecessary disconnection during specified disturbances

C. To eliminate all harmonic current

D. To increase transformer copper losses

Correct Answer: B. To prevent unnecessary disconnection during specified disturbances

Question 4

For a balanced three-phase fault, the basic estimate I_sc = V_LL/(√3 Z_th) assumes:

A. A balanced fault and a linear Thevenin equivalent

B. A single-phase fault with zero voltage

C. An unlimited inverter current

D. A purely mechanical generator model

Correct Answer: A. A balanced fault and a linear Thevenin equivalent

Question 5

Which service is most closely associated with rapid active-power adjustment following a frequency deviation?

A. Primary frequency response

B. Meter certification

C. Harmonic filtering only

D. Transformer oil testing

Correct Answer: A. Primary frequency response

Question 6

A virtual power plant is best described as:

A. A new physical power station at one location

B. An aggregation coordinating distributed resources as a portfolio

C. A protection relay for a transmission line

D. A passive electricity meter

Correct Answer: B. An aggregation coordinating distributed resources as a portfolio

Question 7

Which issue can result from adding inverter filters and cable capacitance to a network?

A. Network resonance at a harmonic frequency

B. Guaranteed elimination of voltage unbalance

C. Infinite fault current

D. Removal of all forecast error

Correct Answer: A. Network resonance at a harmonic frequency

Question 8

If a plant has 80 MVA short-circuit power at its PCC and is rated 10 MVA, its SCR is:

A. 0.125

B. 8

C. 70

D. 800

Correct Answer: B. 8

Question 9

Which planning scenario is especially important for detecting PV-related voltage rise?

A. Maximum PV output and minimum daytime load

B. Zero PV output and maximum night load only

C. Maximum rainfall with no electrical model

D. Only the annual average load

Correct Answer: A. Maximum PV output and minimum daytime load

Question 10

Which practice is most appropriate for trustworthy AI-assisted grid operation?

A. Allowing an unvalidated model to trip protection automatically

B. Ignoring data quality because models learn automatically

C. Using validation, human accountability and secure fallback procedures

D. Publishing all customer meter data without restrictions

Correct Answer: C. Using validation, human accountability and secure fallback procedures