How grid battery storage works
Grid batteries store energy rather than creating a primary energy source. Follow an LFP cell through modules, racks, protected DC collection and a bidirectional converter to the AC grid, then check where energy is lost.
In 60 seconds: how it works
A grid battery stores electricity when there is plenty and releases it when it is needed. It does not make energy, it moves it in time.
- Thousands of battery cells store energy chemically. They are grouped into modules and racks.
- To charge, a bidirectional converter changes the grid’s AC power into DC power for the batteries.
- To discharge, the same converter changes the batteries’ DC power back into AC for the grid.
- A battery management system watches voltage and temperature to keep the cells safe.
- A transformer connects the system to the grid. Some energy is lost as heat in each direction.
Want the engineering detail? Explore the 3D model, then read the engineering notes and the types and variations.
By Power Stations Worldwide · Related reading updated · Preparation and review scope. Educational illustrations; no independent engineering peer review is claimed.
Explore the equipment in 3D
Rotate equipment, follow energy flows and inspect each component. Read the steps below at any time.
Choose a step to read the process. Activate the 3D model or start a guided tour to use the animation controls.
LFP/graphite storage with three independent protected DC groups and bidirectional single-stage PCS units. Nine containers, enlarged cell/module/rack and three-leg PCS cutaways; nominal 2 m/unit and compressed site spacing. One fixed-OCV charge-integration model is illustrative, not a calibrated battery, transient, thermal, fire-safety or structural design. Separate simulation and display clocks. Motion is slowed for inspection. Coloured overlays reveal flow through enclosed equipment; gold pulses show net electrical energy transfer, not individual electrons. Cyan DC cues specifically indicate conventional-current direction, including the opposite return on DC−; gold AC/grid cues indicate mean energy-transfer direction. See the labelled legend and circuit below. Both are visual aids, not particle-speed measurements.
How it works
- Electricity for charging
- AC/DC power conversion
- Electrochemical storage
- Controlled discharge
- AC export to the grid
A battery plant moves electrical energy between times of use. Charging drives electrochemical changes inside the cells; discharging releases some of that stored energy through an external circuit. Ions move through the electrolyte while electrons travel through the electrical connections. The battery is therefore a store of chemical potential energy, rather than a tank of electric current or a new primary energy source. A bidirectional power-conversion system provides the interface between the DC battery and the AC network.
Two ratings describe different capabilities: megawatts measure the rate of charging or discharge, while megawatt-hours measure stored or deliverable energy. Their ratio gives a nominal discharge duration, subject to operating limits. State of charge, losses and auxiliary consumption affect what is available in practice. Control systems monitor cells and coordinate the plant. An operator may reserve energy for a later demand peak or change power rapidly for grid services, but every discharge is limited by the energy and equipment available.
Follow both directions of the battery power path
Selected architecture: LFP/graphite storage, three independent DC groups, three bidirectional single-stage PCS groups and a common AC collector. There is no separate DC/DC stage. All equipment counts, voltage references and efficiencies are illustrative, not specifications of a real installation. [B1][B2][B6][B7]
3D overlay legend: cyan DC cues indicate conventional current, in amperes: during discharge it leaves on DC + and returns on DC −; electrons move oppositely. Gold AC/grid cues indicate mean energy-transfer direction, in watts, not instantaneous alternating current. DC polarity and AC phase labels supplement colour. Particle speed is an inspection aid, not electron speed, ion speed, amperage or MW.
Choose equipment in the diagram controls or 3D model. Selection highlights the same stage and explanation; it never changes your operating command.
Energy, not a tank of current
Charge is tracked in coulombs (C); 1 ampere-hour (Ah) = 3600 C. Energy is tracked in joules (J); 1 megawatt-hour (MWh) = 3.6 × 10⁹ J. A megawatt (MW) is a rate of energy transfer. [B1]
Closed enclosures normally contain the equipment. The enlarged cell, rack and PCS inspection views are teaching cutaways, not instructions to open an energized installation.
Download the editable Blender battery sceneLive measurement points & one checked balance
Positive current and power mean discharge towards the grid; negative mean charge. A very small discharge can still produce net grid import because auxiliaries are larger than export. OCV means open-circuit voltage; here its fixed 768 V value is an assumption, not a measured LFP curve.
- Battery terminal voltage
- 768.000 V
- Aggregate DC current
- 0.000 A
- Current per PCS group
- 0.000 A
- Current per matched rack
- 0.000 A
| Measurement / loss boundary | MW |
|---|---|
| Stored chemical-power proxy (positive draw) | 0.000000 |
| DC terminals (positive discharge) | 0.000000 |
| PCS AC output | 0.000000 |
| After local transformers | 0.000000 |
| After main transformer, before auxiliaries | 0.000000 |
| Net grid / PCC (positive export) | -0.030000 |
| Cell I²R heat | 0.000000 |
| PCS conversion loss | 0.000000 |
| Local transformer loss | 0.000000 |
| Main transformer loss | 0.000000 |
| PCC auxiliary consumption | 0.030000 |
Instantaneous balance residual: . Stored-energy draw = net PCC export + all losses (signed, also valid while charging).
| Integrated quantity | MWh |
|---|---|
| Grid import | 0.000000 |
| Grid export | 0.000000 |
| Cell heat | 0.000000 |
| PCS losses | 0.000000 |
| Local transformer losses | 0.000000 |
| Main transformer losses | 0.000000 |
| Auxiliary energy | 0.000000 |
| Change in stored-energy proxy | 0.000000 |
Ledger residual: . Initial storage + import = final storage + export + every loss. No round-trip claim is shown for an unfinished cycle.
Documented reduced calculation model
All displayed values and 3D flow directions use this single model. It integrates charge exactly for each constant-command interval, splitting a step at a SOC bound before adding idle auxiliary energy. It is not a validated electrochemical, switching or network time-domain simulation. [B1][B5]
- SOC z = q / (3600 C_Ah); dq/dt = −I
- q: remaining charge in C; C_Ah: aggregate capacity in Ah; t: simulated seconds; I: aggregate terminal current in A, positive discharge. Coulombic efficiency is assumed unity, but energy efficiency is not.
- V_t = V₀ − IR; P_DC = V_t I; P_cell heat = I²R
- V_t: terminal voltage (V); V₀: fixed OCV (768 V); R: aggregate matched resistance (0.001 Ω); P: power (W). Current uses the continuous low-current quadratic root I = 2P_DC / (V₀ + √(V₀² − 4RP_DC)).
- U = V₀q; −dU/dt = P_DC + I²R
- U: stored-energy proxy (J). This identity and SOC/energy proportionality require the fixed-OCV assumption. Real OCV, resistance, temperature, capacity and cell balance vary.
- P_out = ηP_in (discharge); |P_in| = |P_out| / η (charge)
- η: assumed efficiency, applied separately at PCS (0.97), local transformer (0.995) and main transformer (0.995). Each boundary's loss is a nonnegative difference. Net PCC power subtracts the separate 30 kW auxiliary branch.
- η_round trip = E_PCC export / E_PCC import
- E: integrated energy, measured at the same PCC over a completed charge/discharge cycle returning to its initial q. A one-way power ratio is not round-trip efficiency. [B2]
Perfectly matched, fixed-capacity LFP groups; no self-discharge, ageing, leakage, hysteresis or polarization dynamics. AC current, reactive power, voltage regulation, harmonics, faults, thermal capacity, fire propagation and structural loads are not calculated. No standards compliance is claimed.
Worked operating point — explicitly invented parameters
At 50% SOC, command 3 MW DC discharge: I = 3926.322932 A, V_t = 764.073677 V; stored-energy draw 3.015416 MW = 3 MW DC + 0.015416 MW cell heat. After the PCS and two transformers: 2.88097275 MW. Subtract 0.030 MW auxiliaries: 2.85097275 MW PCC export.
One simulated hour leaves SOC at 36.366934264%. Recharging at 3 MW DC to the exact original 50% takes 3636.81118235 seconds: export 2.85097275 MWh / import 3.18619497708 MWh = 89.4789167% illustrative complete-cycle efficiency. It is neither a real installation rating nor a universal LFP efficiency.
Default usable energy between 10% and 90%: 17.69472 MWh, from a 22.1184 MWh fixed-voltage reference capacity. An approximate duration must account for the starting SOC, resistive heat, conversions and auxiliary boundary.
All nine stages — readable without WebGL
1. AC interface and grid boundary
A battery site exchanges electricity with the grid through switches and transformers. Charging imports; discharging can export. Idle still uses electricity for auxiliary equipment. [B6][B7]
Engineering detail
PCC means point of common coupling: the common import/export meter. The selected three-phase interfaces are 400 V at each PCS, 33 kV at the collector and 132 kV at the grid, all line-to-line RMS (root-mean-square) reference values, not calculated AC waveforms. Transformer losses are accounted for separately at each level; the 30 kW auxiliary branch is already measured at the PCC. [B6][B7]
2. Bidirectional power conversion
The same power-conversion system (PCS) converts AC to DC when charging and DC to AC when discharging. Transformers change AC voltage; they do not perform this conversion. [B2][B7]
Engineering detail
Each independent battery group supplies a DC-link capacitor and three-leg, two-level voltage-source converter (VSC). Pulse-width modulation (PWM) controls switching, and inductive/capacitive (L/C) filtering reduces switching ripple before transformer coupling. Complementary switch control needs dead time to avoid conducting both switches of a leg together. This is a grid-following architectural explanation, not a switching, harmonics or controller simulation. No extra DC/DC converter is present. Other designs can use one; it adds a conversion boundary and must be modelled explicitly. [B2][B7]
3. Protected DC collection
Both positive and negative conductors connect the batteries to their PCS. Protective devices isolate rack circuits. Structural supports and protective earth are not the DC return. [B1][B6][B7]
Engineering detail
Each of the three groups contains three containers, ten parallel rack equivalents per container and one PCS. The physical groups are not cross-tied on the DC side. Their outputs meet only on the AC collector. The numerical 90-parallel equivalent sums three perfectly matched group currents; each PCS carries one third of the aggregate DC power. Fuses, contactors and isolators are recognizable concepts, not verified protection coordination or fault ratings. [B1][B6][B7]
4. Cell → module → rack → container
Cells in series add voltage. Parallel racks add capacity and current capability. The cutaway shows how a cell fits inside a module, rack and enclosure. [B1]
Engineering detail
The illustrative hierarchy has 16 cells in series per module, 15 modules in series per rack and ten parallel racks per container. Thus 240 × 3.2 V = 768 V open-circuit reference per rack, still 320 Ah per series string. Ninety matched rack equivalents total 28,800 Ah, not 90 times the voltage. The fixed-voltage energy proxy is 768 V × 28,800 Ah = 22.1184 MWh; only 17.69472 MWh lies in the 10–90% teaching window. This is not a manufacturer's nameplate rating. [B1]
5. What happens inside an LFP cell
In this lithium iron phosphate (LFP) / graphite cell, lithium ions travel internally through electrolyte and separator. Electrons move through the external circuit. Energy is stored through reversible electrochemical changes, not as a tank of current. [B1]
Engineering detail
During discharge, lithium leaves the graphite negative electrode and is incorporated at the LFP positive electrode; charging reverses these movements. Cu means copper and Al means aluminium current collectors. The porous separator admits ionic transport while preventing direct electronic contact. This original layered diagram is microscopic and enlarged; it does not represent lithium-metal plating, reaction rates, thermal propagation or a calibrated material model. Electrode polarity labels are clearer here than treating 'anode' and 'cathode' as immutable charging labels. [B1]
6. BMS: monitoring and limits
The battery management system (BMS) monitors cell conditions and communicates permitted operation. It is not the inverter and does not create energy. [B1][B6][B7]
Engineering detail
Real systems monitor individual cell voltages, temperatures and string conditions and may coordinate balancing, contactors and fault responses. The website uses only a charge-based 10–90% operating window with perfect matching. On reaching a bound it stops battery transfer but retains the requested command and continues grid-fed auxiliary energy. This is not a production BMS algorithm, safety certification or operating procedure. [B1][B6][B7]
7. EMS: site dispatch is a separate function
The energy management system (EMS) asks the site to charge, idle or discharge, subject to equipment limits. Your operating buttons make that request; clicking a lesson does not. [B6][B7]
Engineering detail
Site dispatch, BMS limits and PCS switching/control are separate layers. Frequency support, voltage services and grid-forming behaviour require appropriate controls, protection and available energy; they cannot be inferred from cell chemistry. This learning model does not simulate frequency response, islanding, black start, reactive power or network stability. Animation speed changes inspection motion only; simulated-seconds acceleration changes the integration clock, not the MW command. [B6][B7]
8. Cooling, heat and physical access
Battery and power-electronics losses become heat. Cooling equipment and working space matter, even while the site is not exporting electricity. [B1][B6]
Engineering detail
The calculation identifies cell I²R heat separately from PCS and two transformer losses. A lumped 30 kW PCC auxiliary load includes cooling/control demand and its omitted supply losses; no temperature or cooling-capacity equation is solved. Containers sit on pads with maintenance aisles, cabinet supports and cable routes. Open enclosures and enlarged cell/PCS sections are educational inspections, not an operating configuration. Spacing is not a verified fire separation, access-code or structural design. [B1][B6]
9. Discharge, duration and losses
Stored energy runs down during discharge. MW is the transfer rate; MWh is energy. A battery with more MWh can sustain a given MW for longer, subject to its limits and losses. [B1][B2][B5]
Engineering detail
State of charge (SOC) is remaining charge divided by charge capacity, not generally an energy percentage. Only the deliberately fixed open-circuit voltage makes the energy proxy proportional to SOC here. Round-trip efficiency is energy exported / energy imported at the same PCC for a cycle returning to its initial stored charge, with all losses and auxiliaries counted. A partial-cycle ratio or one-way instantaneous efficiency is not round-trip efficiency. [B1][B2][B5]
Check your understanding
Checked passages, further reading & access limits
Passages checked 3 October 2026. Original prose, diagrams and geometry. Only the specified passages were checked, not whole books; software tests are not independent engineering peer review. Numerical assumptions are invented. The required starting titles B3/B4 remain explicit reference-coverage gaps.
- [B1] Gregory L. Plett. Battery Management Systems, Volume I: Battery Modeling. Artech House, 2015. ISBN 9781630810238.
Actual lawful preview passages checked: pp. 2–7, 12–14, 18–19 (units, series/parallel, lithium-ion/LFP and separator). Chapter 2 equations were not accessible here; see B5.
- [B2] Klaus Jäger, Olindo Isabella, Arno H. M. Smets, René A. C. M. M. van Swaaij and Miro Zeman. Solar Energy: Fundamentals, Technology, and Systems. TU Delft course book, 2014.
Actual §§17.3.3 and 17.4.2, pp. 281–284 and 293–295 checked for PWM, filtering and energy units/efficiency. Not the 2016 UIT edition. Use equation (17.25), energy out / energy in; adjacent reversed prose and the simultaneous-open switching typo are not adopted. PV-specific equipment is not transferred to this battery architecture.
- [B3] Kirby W. Beard (editor). Linden's Handbook of Batteries. 5th edition, McGraw-Hill Education, published 2019 (publisher catalog ©2020). ISBN 9781260115925.
Required starting title; further reading only. Metadata/contents checked, relevant technical passages inaccessible. No numerical ratings or safety claims attributed to unread passages.
- [B4] Robert W. Erickson and Dragan Maksimović. Fundamentals of Power Electronics. 3rd edition, Springer, 2020. DOI 10.1007/978-3-030-43881-4.
Required starting title; further reading only. Metadata/contents and introductory abstract checked; relevant detailed inverter/loss passages inaccessible. B2/B7 support the selected architecture instead.
- [B5] Gregory L. Plett. ECE4710/5710, Equivalent-Circuit Cell Models. UCCS author notes, ©2011–2018.
Actual §§2.1–2.2, printed pp. 2–2 through 2–4 checked: charge integration, coulombic versus energy efficiency, OCV − IR. Constant voltage/resistance are our disclosed simplification. Author-hosted HTTP PDF; no executable content is loaded.
- [B6] US DOE / FEMP. Battery Energy Storage Systems: Considerations for Implementation. Summer CAMP T02-S03, 5 August 2025.
Actual PDF p. 8 / printed slide 10 checked for separate BMS, PCS, switchgear, transformer, auxiliary and upper-level control functions. Original diagram and vendor ratings not copied.
- [B7] Jacob Mueller, Michael Ropp and Stan Atcitty. Power Conversion Systems. Chapter 13, DOE Energy Storage Handbook, Sandia National Laboratories, 2021.
Actual single-stage and multi-stage architecture passages, §§3.2.1–3.2.2, pp. 16–18 checked. Original diagram, market-share assertions and historical standards discussion not adopted.
Inside the engineering
Follow the machinery, the working fluids and the electrical connection.
Mechanical
Supported racks and containers sit on foundation pads with maintenance aisles. Separate cooling equipment rejects battery/PCS losses. Enlarged layered cell and PCS sections reveal normally enclosed components; paired DC conductors and three AC phases remain distinct from grounded structural supports. Access dimensions and cable clearances are not compliance claims.
Thermal & chemical
Lithium ions move internally between graphite and the LFP positive electrode; electrons travel externally, in the opposite direction to conventional current. The separator blocks electronic conduction. Charge-based SOC is integrated over simulated time; fixed OCV alone makes energy proportional to SOC here. Resistance, conversion and auxiliaries make energy efficiency less than unity even with unity coulombic efficiency.
Electrical
Each separate 30-rack DC group connects both rails through protection to DC-link capacitance, a three-leg two-level VSC and AC L/C filtering. Three 400 V/33 kV local transformer paths join the AC collector and 33/132 kV main transformer/grid. No extra DC/DC stage or common DC cross-tie. BMS limits, EMS dispatch and PCS control are distinct. One ledger separates cell heat, PCS loss, both transformer levels and grid-fed PCC auxiliaries; selection and visual speed never change charge or dispatch.
Main types and variations
The animation represents a containerised lithium-ion plant. Chemistry, connection topology and inverter controls are separate decisions. A lithium-ion battery can be AC- or DC-coupled, and grid-forming capability must be designed into the complete system.
Lithium iron phosphate (LFP)
LFP is a lithium-ion chemistry with an iron-phosphate positive electrode. Lithium ions move between electrode materials during charging and discharge, while the plant’s external power electronics manage AC exchange. Cells are assembled into modules and racks, with monitoring and protection coordinated at several levels of the system.
What changes in practice
LFP generally has greater thermal stability than NMC, but still needs system-level protection and thermal management. Chemistry alone does not determine the safety of an installation.
Nickel manganese cobalt (NMC)
NMC is another lithium-ion family, using nickel, manganese and cobalt in the positive-electrode material. It stores energy through reversible electrochemical changes like other rechargeable cells. Its battery-management system must respect the permitted cell voltage, temperature and current range while coordinating charging and discharge across the assembled system.
What changes in practice
NMC and LFP involve different trade-offs in density, cost and thermal behaviour. Compare the complete engineered battery system, rather than treating all lithium-ion installations as interchangeable.
Redox flow batteries
Pumps circulate liquid electrolytes between tanks and an electrochemical cell stack. Reactions in the stack absorb or release electrical energy. In a conventional dissolved-species design, tank contents largely determine stored energy while the stack determines power; hybrid flow chemistries with solid active material do not share this independence completely.
What changes in practice
Separating the energy store from the power-producing stack supports different storage durations. Tanks, pumps, membranes and electrolyte management make the plant physically different from a container of lithium-ion racks.
Sodium-ion and sodium-sulphur
These are distinct sodium-based technologies. Sodium-ion cells typically operate near ambient temperature and shuttle sodium ions between electrode materials. Traditional sodium-sulphur batteries use molten active materials at elevated temperature and a ceramic ion-conducting separator. A shared element therefore does not imply the same internal construction or operating conditions.
What changes in practice
Sodium offers a different material supply route. Thermal control, insulation and cell design must match the specific chemistry; sodium-ion should never be confused with high-temperature sodium-sulphur storage.
AC-coupled and DC-coupled storage
With AC coupling, a solar array and battery each use conversion equipment and meet on the AC side. With DC coupling, the battery and PV connect on the DC side of a shared grid-facing conversion arrangement. Both approaches can store solar energy, but the electrical routes and equipment differ.
What changes in practice
Coupling describes the connection architecture, not the cell chemistry. Designers assess conversion paths, shared equipment and the intended operation when combining solar generation with storage.
Grid-following and grid-forming controls
A grid-following inverter synchronises its output with an existing grid voltage. A grid-forming inverter controls a voltage waveform and can provide a reference that helps stabilise the network. This behaviour comes from the converter and its control system; a battery chemistry or enclosure does not supply it automatically.
What changes in practice
Grid-forming storage can contribute stability services. Island operation or network restoration also requires sufficient energy, suitable protection and a coordinated system design; these capabilities cannot be assumed for every battery plant.
Worth knowing
- MW measures power; MWh measures energy. Duration depends on the available starting charge, operating window, losses and auxiliary boundary, not just a nameplate ratio.
- Some energy is lost in each charge–discharge cycle.
- Usable energy depends on temperature, ageing and the permitted state-of-charge window; nameplate energy is not always fully available to dispatch.
See real examples
The atlas includes 320 storage records worldwide.
Show on the mapRelated answers
Textbook references
Suggested technical background for this topic. These books are further reading, not proof that every explanation was checked against every edition. See the full bibliography and review scope.
- Linden's Handbook of BatteriesBattery chemistries, lithium-ion cells, ageing and safety.
- Renewable and Efficient Electric Power SystemsStorage in power systems: energy versus power ratings and round-trip efficiency.
Sources
- EIA: Energy storage for electricity generation ↗
- US DOE / FEMP: Battery system components and electrical single-line diagrams ↗
- Plett: actual battery-modeling book preview ↗
- TU Delft: actual 2014 course-book inverter and energy passages ↗
- Plett: actual charge and resistance model notes ↗
- DOE / Sandia: actual power-conversion architecture chapter ↗
Cutaways reveal enclosed equipment. Layouts, dimensions, materials and animation speeds are illustrative; this is an educational model, not a plant design or a thermodynamic simulation.
