How Does the Power Grid Match Supply and Demand Every Second? | How It Works

What is the short answer?

An electricity grid has very little storage compared with what it delivers, so the power plants feeding it must produce exactly what everyone is using at that same instant. Grid operators watch one number to know whether they are getting it right: the frequency of the alternating current. When demand runs ahead of supply, the spinning generators slow down and the frequency falls; when supply runs ahead, it rises. Operators keep it within a narrow band using forecasts made a day ahead, automatic controls that act in seconds, reserve plants that start within minutes and, as a last resort, cutting off some customers to save the rest.

Definition: Grid frequency is the number of complete cycles the alternating current makes each second, measured in hertz. The current reverses direction twice in each cycle. It is 50 hertz in India, Europe and most of Asia and Africa, and 60 hertz in North America.

Why must supply match demand at every moment?

Most power stations make electricity with large generators that spin in step with the grid. In a 50 hertz grid, every one of those machines turns in lockstep, like riders on one long tandem bicycle. When someone switches on a kettle, that extra load acts like a small hill. The generators slow slightly, and the frequency dips, until more fuel, water or steam is fed in.

The frequency is therefore a live measure of balance across the whole grid. The same reading appears in a control room in Mumbai and at a substation hundreds of kilometres away, because the whole synchronous area turns together. India has run one national synchronous grid since 31 December 2013, when the southern region was joined to the rest.

Japan shows how deep these choices run. Its eastern half, including Tokyo, uses 50 hertz and its western half, including Osaka, uses 60 hertz. Power can cross between the two halves only through special converter stations.

Storage is growing, but it still holds only a small fraction of daily demand in most countries. So the job remains balancing, not storing, and grid codes set the band within which the frequency must stay. In India, the Central Electricity Regulatory Commission's grid code sets that band at 49.90 to 50.05 hertz.

What happens, step by step, to keep the grid balanced?

1. Forecast demand. A day ahead, operators predict use for each block of the next day from weather forecasts, the calendar and past patterns. A heatwave, a cold snap or a big televised match changes the forecast.
2. Schedule generation. Plants are scheduled, often through an electricity market, to meet that forecast at the lowest cost, with extra capacity held back as reserves.
3. Inertia absorbs the first shock. If a large plant trips, the energy stored in every spinning generator slows the fall in frequency during the first few seconds. No one has to act for this to happen; it is physics.
4. Primary response acts in seconds. Governors on generators sense the falling frequency and open their steam or water valves automatically. In Continental Europe, this reserve must be fully active within 30 seconds.
5. Secondary control restores the frequency in minutes. A central computer, using automatic generation control, raises output at chosen plants until the frequency returns to normal and power flows between regions match their schedules.
6. Tertiary reserves replace what was used. Within 15 minutes or so, operators start other plants or buy power from neighbours, so the fast reserves are free for the next fault.
7. Load shedding is the last line. If the frequency keeps falling past a set point, relays automatically disconnect blocks of customers. Losing some supply for a while is better than the whole grid collapsing.

How does this work with real numbers?

Take an illustrative grid with 50,000 megawatts of generation running at 50 hertz. Suddenly, a 1,000 megawatt power station trips. The figures below are invented to show the reasoning, but the method, from the standard swing equation, is the one engineers use for a first estimate.

The energy stored in the spinning machines is measured with an inertia constant, H, in seconds. A grid with many large coal, gas, hydro and nuclear units might have H of about 5 seconds. The starting rate at which the frequency falls is roughly the lost power times the normal frequency, divided by twice H times the total generation:

Rate of fall = (1,000 × 50) ÷ (2 × 5 × 50,000) = 50,000 ÷ 500,000 = 0.1 hertz per second.

Table: How inertia changes the first seconds after the same 1,000 megawatt loss

In practice, primary response starts within seconds and the fall slows long before 10 seconds pass. The table shows why low inertia worries operators: the same fault leaves half as long for reserves to act.

Real grids size their reserves to the biggest single loss they expect. In Continental Europe, which runs as one synchronous area from Portugal to Turkey, Ukraine and the Baltic states, the reserve that acts within 30 seconds is sized for a sudden loss of 3,000 megawatts.

Note: Demand helps a little on its own. Many motors draw slightly less power when the frequency drops, which slows the fall. Engineers call this load self-regulation.

How are solar, wind and batteries changing the balancing job?

Solar panels, batteries and most modern wind turbines connect through inverters and other power electronics, which turn their output into alternating current at grid frequency. These separate any spinning parts from the grid, so they add little or no natural inertia. As they replace coal and gas plants, grids lose part of their first line of defence.

Their output also depends on the weather. A cloud bank or a calm evening can remove thousands of megawatts within hours, so forecasting the weather has become part of forecasting supply.

Grids are adapting in several ways:

• Batteries. A battery can respond in a fraction of a second. The Hornsdale Power Reserve in South Australia, opened in December 2017, showed how quickly large batteries can steady the frequency after a fault.
• Grid-forming inverters. New inverter designs can imitate the steadying effect of a spinning machine, and grid codes are starting to require them.
• Pumped storage. Water is pumped uphill when power is cheap and released through turbines when demand peaks.
• Demand response. Factories, water pumps and electric vehicle chargers can be paid to cut or shift their use for a while.
• Interconnectors. Links between countries or regions let one area lean on another when its own supply is short.

Where does the balance break down?

Large blackouts are rare, and most begin with several problems at once. Five cases show the main ways a grid fails:

• United States and Canada, 14 August 2003. Lines in Ohio sagged into overgrown trees, and a failed alarm system meant operators did not see the trouble build. The cascade cut power to about 50 million people in eight states and Ontario.
• India, 30 and 31 July 2012. On successive days, the northern grid and then the northern, eastern and north-eastern grids collapsed, cutting about 36,000 megawatts of load on 30 July and about 48,000 megawatts on 31 July. The government's enquiry committee found no single cause: some northern utilities had drawn more power than scheduled while several key lines were out of service.
• Great Britain, 9 August 2019. After a lightning strike, an offshore wind farm and a gas plant tripped within seconds, a loss larger than the reserves were sized for. The frequency fell below 48.8 hertz, and automatic load shedding cut supply to about 1 million customers. The rest of the grid survived, as the design intended.
• Texas, United States, February 2021. A winter storm froze gas wells and power plants just as heating demand soared. The grid operator, ERCOT, ordered rotating outages of up to 20,000 megawatts, cutting power to millions of customers, and later said the grid had been minutes from a total blackout.
• Spain and Portugal, 28 April 2025. Almost all of mainland Spain and Portugal lost power at 12:33 Spanish time (11:33 in Portugal). The European grid operators' expert panel found that voltage rose faster than the system could control it, alongside oscillations and plant trips, and that a chain of plants then disconnected. Its October 2025 report called it the first known blackout caused by overvoltage.

The lesson across these cases is that balance depends on more than frequency. Voltage, line limits, maintenance, weather planning and clear communication between control rooms all matter.

Where can you go deeper in World 101 and the board notes?

This page explains one mechanism. The World 101 subject Energy Systems and the Power Grid covers how electricity is made, moved and priced around the world.

Two related subjects take it further. Renewable Energy looks at solar, wind and hydro power and their limits. The Energy Transition looks at how countries are moving away from coal, oil and gas, and what that asks of their grids.

For the physics behind the grid, see the Alternating Current notes for CBSE Class 12 Physics and ISC Class 12 Physics, which explain frequency and transformers. Electrical Power and Household Circuits for ICSE Class 10 Physics follows the power from the station to the home.

Sources

The figures and cases on this page come from these regulators, grid operators and official inquiries:

• European Commission Regulation (EU) 2017/1485, the guideline on electricity transmission system operation: the 3,000 megawatt reference loss and the 30-second activation time for Continental Europe.
• Central Electricity Regulatory Commission, Indian Electricity Grid Code Regulations, 2023: the frequency band of 49.90 to 50.05 hertz.
• U.S.-Canada Power System Outage Task Force, "Final Report on the August 14, 2003 Blackout in the United States and Canada", April 2004.
• Ministry of Power, Government of India, report of the Enquiry Committee on the grid disturbances of 30 and 31 July 2012, August 2012.
• Ofgem, report on the 9 August 2019 power outage, January 2020.
• Federal Energy Regulatory Commission and North American Electric Reliability Corporation, report on the February 2021 cold weather outages in Texas and the South Central United States, November 2021.
• ENTSO-E expert panel, factual report on the 28 April 2025 blackout in Spain and Portugal, October 2025, and final report, March 2026.
• Ministry of Power, Government of India, "Creation of National Grid": the southern region joined the national synchronous grid through the 765 kV Raichur to Sholapur line on 31 December 2013.
• Hitachi, "Completion of Chubu Electric Power Grid Hida converter station equipped with Hitachi's frequency converters", 1 April 2021: the converters that link Japan's 50 hertz east and 60 hertz west.
• ENTSO-E, "Continental Europe successful synchronisation with Ukraine and Moldova power systems", 16 March 2022, and "ENTSO-E confirms successful synchronization of the Continental European electricity system with the systems of the Baltic countries", 9 February 2025.
• Australian Energy Market Operator, "Initial operation of the Hornsdale Power Reserve Battery Energy Storage System", April 2018, on the battery that began operating in December 2017.
• National Grid ESO, "Technical Report on the events of 9 August 2019", 6 September 2019.
• ERCOT, review of the February 2021 extreme cold weather event, presentation to the Board of Directors, 24 February 2021.
• P. Kundur, Power System Stability and Control, McGraw-Hill, 1994: the swing equation behind the worked example.

Glossary

• Frequency: the number of cycles per second of alternating current, which falls when demand exceeds supply and rises when supply exceeds demand.
• Synchronous area: a region in which all generators turn in step at the same frequency, such as India's national grid or Continental Europe.
• Inertia: the energy stored in spinning generators, which slows the change in frequency in the first seconds after a fault.
• Reserves: generation or demand held ready to respond when the balance is upset, sorted by how fast they act.
• Load shedding: cutting supply to some customers on purpose to stop the whole grid collapsing.
• Inverter: an electronic device that turns the output of solar panels or batteries into alternating current at grid frequency.

Common errors and misconceptions

• Misconception: Power stations store electricity and release it when people need it. Correct: Most electricity is used the instant it is made; the grid balances supply and demand continuously.
• Misconception: A blackout always means there were too few power plants. Correct: Many blackouts start with faults on lines, control failures or voltage problems, even when enough capacity exists.
• Misconception: Load shedding is always a failure of the system. Correct: Automatic under-frequency load shedding is a designed defence; routine rota cuts caused by shortages are a sign of too little supply.
• Misconception: Solar and wind make a grid impossible to balance. Correct: They make balancing harder, but batteries, better forecasts, new inverters and interconnectors are filling the gap.

Key takeaways

• A grid must produce exactly as much electricity as is being used, at every instant, because it stores very little.
• Frequency is the grid's balance gauge: it falls when demand exceeds supply and rises when supply exceeds demand.
• Inertia, then automatic primary response, then secondary and tertiary reserves restore the balance after a fault.
• Load shedding is the last automatic defence, cutting some customers to save the rest of the grid.
• Solar and wind add little or no natural inertia, so grids are adding batteries, grid-forming inverters and interconnectors.
• Major blackouts usually combine several failures, such as line faults, weather and missed warnings.

Test yourself

What does a falling grid frequency tell an operator?

It shows that demand is running ahead of supply, so the spinning generators are slowing and more power is needed quickly.

Why does inertia matter in the first seconds after a power station trips?

The energy stored in spinning generators slows the fall in frequency, buying time for automatic reserves to respond.

What is the frequency band set by India's grid code?

The Central Electricity Regulatory Commission sets it at 49.90 to 50.05 hertz around the normal 50 hertz.

Why do solar panels and wind turbines reduce a grid's inertia?

They connect through power electronics, which separate any spinning parts, such as wind turbine blades, from the grid frequency, so they add little or no natural inertia.

What happened to the frequency in Great Britain on 9 August 2019?

It fell below 48.8 hertz after two plants tripped, and automatic load shedding cut supply to about 1 million customers.