{
  "id": "microgrid",
  "title": "Microgrid",
  "lang": "en",
  "source_lang": "en",
  "summary": "A microgrid is a coordinated local electricity system that supplies defined loads and can operate independently of a wider grid.",
  "category": "technology",
  "subcategory": "energy",
  "type": "concept",
  "aliases": [
    "micro-grid"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/microgrid/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/microgrid/",
      "json": "https://aiwiki.page/zh/microgrid.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A microgrid is a group of interconnected electrical loads and distributed energy resources within a defined electrical boundary that can be controlled as a single system. It may operate connected to a wider electrical grid, independently as an electrical island, or exclusively as an isolated network. Its defining feature is coordinated control of local supply and demand—not simply the presence of rooftop solar panels, batteries, or standby generators.\n\nMicrogrids serve installations ranging from individual buildings and university campuses to industrial sites, villages, and islands. They can maintain selected services during utility outages, integrate renewable energy, and coordinate electricity with heating, cooling, and transport. Their performance depends on generation capacity, stored energy, network design, control systems, and operating arrangements. A microgrid is not necessarily renewable, continuously self-sufficient, or capable of an interruption-free transition into islanded operation."
    },
    {
      "heading": "Definition, boundaries, and electrical architectures",
      "text": "A microgrid has an identifiable electrical boundary enclosing the resources and loads under its control. For a grid-connected installation, the interface with the utility is commonly called the point of common coupling or point of interconnection. Switching equipment at this interface allows the local network to separate from the utility. A microgrid may have several interconnection points, but their coordinated operation must preserve a well-defined boundary and prevent unintended energization of external circuits.\n\nSize alone does not determine whether a system is a microgrid. A small building with coordinated generation, storage, loads, and islanding controls can qualify, while a large collection of independently operated generators may not. There is no universally accepted upper capacity limit. Projects range from kilowatt-scale installations to networks supplying many megawatts. The term minigrid is also used, especially for rural electrification, although its meaning varies by jurisdiction and development program.\n\nIn an alternating-current microgrid, sources and loads connect through an AC distribution network, usually at conventional utility frequencies of 50 or 60 hertz. Larger installations commonly use three-phase electric power, and different voltage levels may be linked by an electrical transformer. AC architecture is compatible with established distribution equipment, rotating generators, and most existing buildings. It nevertheless requires control of frequency, voltage, reactive power, and synchronization between sources.\n\nA direct-current microgrid distributes electricity through one or more DC buses. Photovoltaic arrays, batteries, electronic equipment, and some vehicle-charging systems can connect through DC conversion stages, potentially avoiding repeated AC–DC conversions. Voltage regulation and fault interruption remain essential, however. DC faults lack the natural current zero crossings of AC systems, making interruption more demanding. A DC–DC converter may provide voltage adaptation, controlled power transfer, or electrical isolation, depending on its topology.\n\nHybrid AC–DC microgrids connect AC and DC subnetworks through bidirectional converters. Their advantages depend on the actual mix of sources and loads rather than on an inherent superiority of either architecture. Physical layouts may be radial, ring-shaped, or meshed; radial systems are comparatively simple to protect, while alternative paths can improve service continuity at the cost of more complex switching and protection. A networked microgrid links two or more controllable microgrids so that they can exchange power or support one another under specified operating conditions.\n\nA virtual power plant is a related but distinct concept. It aggregates distributed resources for coordinated scheduling or market participation, potentially across widely separated locations. Those resources need not form a local electrical island. A microgrid instead has a physical network boundary and a local operating responsibility, even when it also participates in a broader aggregation."
    },
    {
      "heading": "Generation, storage, and controllable demand",
      "text": "Microgrids combine resources with different operating characteristics. Dispatchable generation can adjust output when required, subject to fuel, maintenance, and technical constraints. Common examples include a diesel engine driving an electrical generator, natural-gas reciprocating engines, small gas turbines, and fuel cells. These resources can provide sustained electricity when fuel remains available, but minimum loading, start-up time, ramp limits, emissions, and maintenance requirements affect how they are scheduled.\n\nCombined heat and power installations recover useful heat from electricity generation. At hospitals, universities, and industrial facilities, recovered heat can supply hot water, steam, or absorption cooling. Their economics depend strongly on matching thermal production to useful demand. A generator that is efficient when both electricity and heat are used may perform poorly when heat is rejected. Some plants also require auxiliary electricity to start, so the presence of cogeneration does not automatically establish independent restart capability.\n\nSolar energy and wind power reduce fuel consumption but introduce weather-dependent production. A photovoltaic array built from solar cells supplies DC electricity, typically connected to an AC network through a power inverter. Solar output follows irradiance and is affected by cloud cover and temperature; wind generation depends on local wind conditions and turbine operating limits. Their contribution to annual energy can be large even when their dependable output during a particular emergency is limited.\n\nEnergy storage separates the timing of production and consumption. Batteries can respond rapidly to power imbalances, provide reserve, absorb excess generation, and support restoration. A lithium-ion battery is widely used because of its efficiency, response speed, and modularity. Other options include lead–acid batteries, flow batteries, flywheels, and thermal storage. A battery management system monitors and constrains battery operation, but it is distinct from the controller that dispatches the complete microgrid.\n\nStorage power and storage energy describe different capabilities. A battery rated at 1 MW and 4 MWh can nominally discharge at rated power for four hours, before accounting for usable state-of-charge limits, conversion losses, auxiliary consumption, and degradation. A high power rating helps manage abrupt disturbances, while a large energy inventory supports prolonged operation. Neither rating alone establishes how long an entire site can remain islanded.\n\nDemand is also a controllable resource. Demand response shifts or reduces consumption through equipment such as refrigeration systems, pumps, air conditioning, and electric-vehicle chargers. Thermal inertia and stored water can permit temporary interruption without immediately losing the underlying service. Loads are often assigned priorities: life-safety equipment and essential communications receive preference over discretionary charging or comfort cooling. Such classification must reflect actual operating requirements; a small auxiliary load may be essential to the operation of a much larger critical system."
    },
    {
      "heading": "Grid-connected operation, islanding, and restoration",
      "text": "During grid-connected operation, the wider network normally provides a strong reference for voltage and frequency. The microgrid may import electricity, export surplus generation where permitted, reduce peak demand, or maintain storage for a possible outage. Its controller must respect interconnection limits, equipment ratings, and contractual restrictions. Some installations prohibit export altogether, requiring active control of generation and storage when local production approaches local consumption.\n\nThe instantaneous active-power balance can be expressed, with imports taken as positive, as\n\n\\[P_{\\mathrm{gen}}+P_{\\mathrm{storage}}+P_{\\mathrm{grid}} = P_{\\mathrm{load}}+P_{\\mathrm{loss}},\\]\n\nwhere positive \\(P_{\\mathrm{storage}}\\) denotes discharge. In islanded operation, \\(P_{\\mathrm{grid}}=0\\). Generation, storage, and demand must then remain balanced internally. A persistent deficit cannot be corrected by control software alone: it requires additional generation, storage discharge, or load reduction. Reactive-power balance and voltage constraints must also be satisfied in an AC system.\n\nIslanding can be planned or unplanned. A planned separation allows the controller to reduce power flow across the interconnection, establish appropriate reserve, and prepare local voltage-forming resources. An unplanned separation follows an upstream fault or disturbance and may occur while the microgrid is importing substantial power. Local resources must rapidly replace that import or disconnect enough load to avoid collapse. The transition may be effectively seamless for some loads, involve a short interruption, or require a complete shutdown followed by restart.\n\nA key distinction is between intentional island operation and accidental islanding. Utility interconnection protection generally seeks to prevent generators from continuing to energize an uncontrolled portion of a disconnected distribution system. A properly designed microgrid instead creates a deliberate island behind an established isolation boundary. Its protection and control arrangements must ensure that local generation cannot backfeed utility lines that are expected to be de-energized.\n\nBlack start is the ability to restore an electrical system without an external energized supply. A black-start resource must establish the initial voltage and, for an AC system, frequency. Restoration then proceeds by energizing buses, starting auxiliary equipment, connecting additional generation, and adding loads in manageable blocks. Transformer magnetizing inrush, motor-starting current, and the simultaneous return of thermostatically controlled loads can exceed normal operating demand, making restoration more demanding than simply matching steady-state power ratings.\n\nReconnection requires the island and utility to be compatible before their circuits are joined. Voltage magnitude, frequency, phase sequence, and phase angle must fall within acceptable limits. A synchronizing controller adjusts the island and authorizes closure of the interconnection switch. After reconnection, power transfers are changed gradually to avoid equipment overload or abrupt disturbance. A standby generator connected through an open-transition transfer switch may provide valuable backup, but it does not necessarily offer the coordinated multi-resource operation associated with a microgrid."
    },
    {
      "heading": "Control hierarchy and power-electronic behavior",
      "text": "Microgrid control operates across several timescales. Converter switching and current regulation act rapidly; primary controls stabilize immediate disturbances; secondary controls correct deviations; and supervisory scheduling allocates resources over minutes, hours, or longer. This hierarchy connects control theory with power electronics and operational planning. Implementations differ, and some combine functions rather than assigning them to separate devices.\n\nIn an AC system dominated by conventional synchronous generators, rotating machines contribute physical inertia and establish voltage through their excitation systems. Power-electronic resources behave according to their control algorithms and current limits. A grid-following inverter typically measures an existing voltage waveform and injects controlled current relative to it. A grid-forming inverter establishes a voltage waveform and can support an island without relying on a strong external reference, provided its source, controls, and protection are designed for that duty.\n\nPrimary control often uses droop characteristics to share load among sources without requiring continuous centralized communication. A simplified active-power–frequency relationship is\n\n\\[f=f_0-m_P(P-P_0),\\]\n\nwhere \\(f_0\\) and \\(P_0\\) are reference values and \\(m_P\\) is the droop coefficient. An increase in active-power output is associated with a reduction in frequency reference, allowing parallel sources to share changing demand. A similar relationship can link voltage and reactive power. Conventional decoupled droop relationships work most directly under suitable network-impedance assumptions; resistive low-voltage feeders may require modified controls.\n\nSecondary control restores quantities displaced by primary droop, such as frequency and bus voltage. It can be centralized, with one controller coordinating resources, or distributed, with local controllers exchanging information. Communication delays, packet loss, measurement errors, and unequal line impedances can affect performance. Stable operation therefore requires attention to interactions among control loops, not merely correct operation of each device in isolation.\n\nThe supervisory energy-management system determines schedules and operating reserves. It uses forecasts of demand, renewable production, electricity prices, and equipment availability to select generator commitments, battery charging, and flexible-load actions. Mathematical optimization can minimize operating cost while enforcing fuel limits, battery constraints, emissions requirements, and islanding reserves. Model predictive control repeatedly updates decisions as new measurements and forecasts become available, implementing only the near-term portion of each calculated schedule.\n\nA major technical challenge is the interaction of heterogeneous equipment. Two inverters that operate satisfactorily in separate tests may interact poorly through a weak network because their voltage controls, synchronization systems, or filters overlap dynamically. Grid-forming capability also does not imply unlimited fault current or unlimited energy. The inverter remains constrained by semiconductor ratings and the battery or generator behind it. Detailed simulation, controller testing, and commissioning must therefore consider both system-level behavior and device limits."
    },
    {
      "heading": "Protection, power quality, and communications",
      "text": "Electrical protection must distinguish faults from normal operating changes and isolate the smallest practical affected area. Conventional distribution protection often relies on large fault currents supplied by the utility and on predictable current direction. Microgrids complicate both assumptions. Power may flow in either direction, and islanded operation can produce much lower fault current than grid-connected operation, especially when inverter-based resources dominate.\n\nA circuit breaker or fuse selected for utility-fed fault levels may not coordinate correctly when the microgrid is islanded. Inverters often limit output current to protect their switching devices, so an overcurrent relay may receive insufficient evidence of a fault. Possible responses include directional protection, differential protection comparing currents at the boundaries of a protected zone, communication-assisted tripping, and adaptive settings that change with operating mode. Each approach introduces trade-offs in cost, communication dependence, and sensitivity.\n\nGrounding requires particular care. Separating from the utility can alter the path for earth-fault current or remove the normal neutral reference. Generator winding arrangements, transformer connections, grounding resistors, and switched neutral conductors must be coordinated across grid-connected and islanded modes. Incorrect arrangements can impair fault detection or create unsafe touch voltages. DC networks add issues including pole-to-earth faults, insulation monitoring, and rapid discharge from connected capacitors.\n\nPower quality concerns extend beyond average voltage and frequency. Nonlinear loads and converter switching can produce harmonics, while motor starts and abrupt load changes can cause voltage dips. Unbalanced single-phase loads stress three-phase equipment and complicate converter control. Filters, voltage regulation, load balancing, and suitable reserve improve performance, but their interactions require analysis. A microgrid designed to supply ordinary building loads may not meet the tighter requirements of a precision manufacturing process.\n\nCommunications connect meters, protective devices, controllers, and operator interfaces. Supervisory control and data acquisition provides monitoring and remote operation, while local controllers execute time-critical functions. A robust design distinguishes functions that need communication from those that must continue safely when communication fails. Primary voltage and frequency support generally cannot depend on a remote cloud service remaining continuously available.\n\nCybersecurity is part of electrical reliability because unauthorized control commands can change breaker positions, generator output, or protection settings. Measures include network segmentation, authenticated access, controlled software updates, event logging, and restricted privileges. Recovery procedures also matter: equipment may need to operate locally while supervisory systems are unavailable. Testing should verify electrical responses to failed links, stale measurements, inconsistent device states, and loss of the main controller, rather than assuming communications are always correct."
    },
    {
      "heading": "Reliability, resilience, and critical services",
      "text": "Reliability and resilience describe related but different objectives. Reliability concerns the continuity and adequacy of service under expected conditions and ordinary contingencies. Resilience emphasizes the capacity to withstand, adapt to, and recover from disruptive events, including prolonged utility outages or damage to local infrastructure. A microgrid can improve both, but only for the loads physically connected to an operable part of its network.\n\nCritical-load identification is therefore central. A hospital may need electricity for intensive-care equipment, ventilation, communications, sterilization, and water pumping, while a community shelter may require refrigeration, lighting, heating or cooling, and device charging. Dependencies extend beyond electrical nameplates. A functioning generator does not preserve a service if floodwater disables its switchgear, a fuel pump lacks power, or the telecommunications connection required by an operational system has failed.\n\nAutonomy is an energy-and-logistics problem, not simply a generation-capacity problem. A useful first approximation for battery-only operation is\n\n\\[T \\approx \\frac{E_{\\mathrm{usable}}}{P_{\\mathrm{average}}},\\]\n\nwhere \\(E_{\\mathrm{usable}}\\) is available stored energy and \\(P_{\\mathrm{average}}\\) is average supplied load. Actual duration depends on load variation, conversion efficiency, temperature, reserve requirements, and battery condition. Renewable generation may extend operation, but its contribution during a particular event depends on weather and equipment availability. Fuel-based generation introduces dependence on stored inventories, delivery access, and functioning fuel infrastructure.\n\nResilience assessment often uses outage scenarios rather than a single design duration. Simulations can vary outage start time, duration, renewable production, equipment failures, and load priorities. Relevant measures include unserved energy, hours of critical-load service, and the probability of sustaining specified functions for a given period. Two systems with the same annual renewable output can have very different resilience if one regularly reaches its minimum battery charge just before an outage.\n\nEquipment placement and physical diversity are important. Elevating switchgear above flood levels, separating redundant circuits, protecting fuel supplies, and avoiding a single vulnerable communications path may provide more benefit than adding generation alone. Routine testing is also essential because islanding switches, starting systems, batteries, and protection settings may remain largely unexercised during normal grid-connected operation. A successful transfer test under light load does not establish performance under the highest credible demand.\n\nA uninterruptible power supply and a microgrid can complement one another. The former bridges very short interruptions for sensitive equipment, while the latter supports a larger site over longer periods. Even a microgrid with rapid islanding may retain local uninterruptible supplies for loads that cannot tolerate any discontinuity. The required service level determines whether the design prioritizes interruption duration, total energy availability, redundancy, or some combination of these."
    },
    {
      "heading": "Economics, ownership, and environmental performance",
      "text": "Microgrid economics combine several value streams. Normal operation may reduce electricity purchases, lower billed peak demand, improve use of on-site generation, or earn compensation for services supplied to the wider grid. Islanding capability adds a different benefit: avoiding some consequences of outages. These values are not automatically additive because the same equipment may be needed for competing purposes at the same time.\n\nA battery illustrates this conflict. Discharging during an expensive tariff period can reduce an electricity bill, but it leaves less energy available if an outage begins immediately afterward. Maintaining a high reserve improves readiness while reducing opportunities for price arbitrage. Similarly, providing utility frequency support may consume inverter capacity that would otherwise handle local disturbances. Dispatch optimization must include these opportunity costs and any contractual commitments.\n\nCapital expenditure includes more than generators and storage. Engineering studies, switchgear, protection changes, controllers, communications, civil works, permits, and commissioning can represent substantial costs. Existing buildings may require expensive modifications to separate critical loads from noncritical circuits. Over the operating life, fuel, staffing, inspections, replacement parts, battery degradation, and software support affect total cost. Projects are commonly evaluated using discounted cash flow, but resilience benefits depend on uncertain outage frequency and the consequences of service interruption.\n\nOwnership arrangements vary. A facility may own and operate the complete system, or a third party may finance equipment and sell electricity or resilience services through a long-term agreement. Utilities can own microgrids serving selected customers where regulation permits. Community projects may involve municipal bodies, cooperatives, or multiple private owners. These structures determine responsibility for dispatch, maintenance, fuel procurement, cybersecurity, and restoration decisions.\n\nEnvironmental performance depends on what the microgrid builds, how it operates, and which external generation it displaces. Renewable generation and efficient cogeneration can reduce fuel use, while diesel operation produces local air pollutants as well as greenhouse gases. Battery storage shifts energy rather than creating it; its emissions effect depends partly on the electricity used for charging and the generation displaced during discharge. Manufacturing impacts and equipment replacement also belong in a life-cycle assessment.\n\nElectrification can create additional interactions. Heat pumps may reduce fuel consumption for heating but increase winter electrical demand. Vehicle charging can provide flexible demand, while bidirectional vehicles can potentially supply buildings when compatible equipment, operating agreements, and vehicle availability permit. The grid connection, local network, and protection must accommodate these flows. A system described as “net zero” on an annual energy basis may still import electricity during critical hours and require substantial backup during extended low-renewable periods."
    },
    {
      "heading": "Development, standards, and representative applications",
      "text": "Locally generated electricity predates large interconnected grids. Early power stations supplied nearby customers, while remote settlements and industrial facilities continued using isolated networks after centralized utility systems expanded. The modern microgrid concept developed around a different technical objective: treating multiple distributed resources and loads as a coordinated entity capable of intentional islanding and controlled interaction with a larger network.\n\nThe Consortium for Electric Reliability Technology Solutions, established in the United States in 1999, played an influential role in formalizing this approach. Work associated with Robert H. Lasseter emphasized autonomous local control, distributed generation, and stable transitions between interconnected and islanded operation. A CERTS microgrid test bed at American Electric Power’s Dolan Technology Center in Ohio demonstrated coordinated operation of multiple resources and helped establish experimental methods for evaluating the concept.\n\nCampus and institutional systems became important applications because they combined concentrated demand, identifiable ownership, and existing energy infrastructure. The University of California, San Diego developed a widely studied campus system combining cogeneration, solar generation, storage, and sophisticated energy management. Princeton University’s campus microgrid supplied selected campus loads while islanded during Hurricane Sandy in October 2012, illustrating the use of an established institutional energy system during a regional disruption.\n\nThe Sendai microgrid in Japan became another prominent case following the March 2011 earthquake and tsunami. Its differentiated supply arrangements supported designated loads during the disruption and demonstrated the importance of designing around specific service requirements. Remote-community applications have different priorities: renewable generation and storage can reduce dependence on delivered diesel fuel, but controls must account for limited technical support, seasonal demand, and difficult supply chains.\n\nStandards distinguish equipment interconnection from microgrid-level control. IEEE 1547-2018 addresses the interconnection and interoperability of distributed energy resources with associated electric power system interfaces. It is not, by itself, a complete design specification for an islanded microgrid. IEEE 2030.7-2017 specifies functions for microgrid controllers, while IEEE 2030.8-2018 addresses testing of those controllers. The IEC 62898 series provides guidance on microgrid planning and operation. Actual installations must also comply with applicable electrical codes, utility requirements, and local safety rules.\n\nTesting methods increasingly connect real controllers or power equipment to simulated networks. Hardware-in-the-loop simulation permits evaluation of faults, communication delays, transitions, and abnormal operating conditions without exposing a complete installation to every disturbance. Power-hardware-in-the-loop methods extend this approach to physical electrical devices. Such tests are especially useful when different manufacturers’ equipment must operate together, although field commissioning remains necessary to verify wiring, settings, grounding, and actual protection behavior."
    },
    {
      "heading": "Research directions and unresolved engineering problems",
      "text": "High shares of inverter-based generation are changing the central stability problem. Traditional methods developed for networks dominated by synchronous machines do not always describe fast converter interactions adequately. Research examines how grid-forming controls share loads, respond to faults, limit current, and recover after large disturbances. A controller must remain stable not only in normal operation but also when current limiting changes its effective behavior.\n\nProtection remains a closely connected challenge. A low-current fault may resemble a heavy load, while power-electronic controls can alter fault waveforms within milliseconds. Proposed techniques use differential measurements, high-frequency signals, traveling-wave information, or coordinated device status. Their practical value depends on dependable detection, manageable communication requirements, and compatibility with existing equipment. Protection must also remain selective as sources connect, disconnect, and change operating modes.\n\nUncertainty affects both planning and operation. Renewable forecasts, customer demand, fuel availability, component failures, and outage duration are imperfectly known. Stochastic and robust optimization methods represent these uncertainties in different ways, balancing expected cost against adverse outcomes. Machine learning can assist forecasting and equipment monitoring, but operational control still requires enforceable electrical constraints and validated behavior outside the conditions represented in historical data.\n\nNetworked microgrids introduce additional coordination problems. Neighboring systems may exchange emergency support, share reserve, or restore a damaged distribution area in stages. These benefits require compatible protection, clear operating authority, and feasible power flows through the connecting network. Independent owners may also have conflicting priorities: one facility may need to conserve fuel while another seeks additional supply. Market rules and contracts cannot replace the electrical checks needed before systems are interconnected.\n\nLong-duration operation raises questions beyond fast control. Increasing battery capacity can extend autonomy, but different storage technologies, dispatchable low-carbon generation, thermal storage, and flexible demand may be better suited to particular durations and climates. Seasonal energy adequacy differs fundamentally from second-to-second stability. A microgrid can regulate frequency perfectly yet still exhaust its energy supply during a prolonged weather event.\n\nInteroperability and maintainability are continuing deployment issues. Open interfaces can reduce vendor lock-in, but a common communications protocol does not guarantee that devices interpret operating states or limits identically. Equipment replacement over a system’s life may introduce new controls and firmware into an older installation. Research and standardization therefore address reproducible testing, explicit capability descriptions, secure update procedures, and methods for verifying that a modified microgrid retains its intended operating envelope."
    }
  ],
  "links": [
    "electrical-grid",
    "renewable-energy",
    "three-phase-electric-power",
    "electrical-transformer",
    "dc-dc-converter",
    "diesel-engine",
    "electric-generator",
    "gas-turbine",
    "fuel-cell",
    "combined-heat-and-power",
    "solar-energy",
    "wind-power",
    "solar-cell",
    "power-inverter",
    "energy-storage",
    "lithium-ion-battery",
    "lead-acid-battery",
    "flow-battery",
    "battery-management-system",
    "demand-response",
    "air-conditioning",
    "electric-vehicle",
    "black-start",
    "control-theory",
    "power-electronics",
    "mathematical-optimization",
    "model-predictive-control",
    "circuit-breaker",
    "scada",
    "cybersecurity",
    "infrastructure",
    "greenhouse-gas",
    "heat-pump",
    "machine-learning",
    "vendor-lock-in"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}