Ask a utility engineer, a fire marshal, and a due-diligence consultant what document they want first, and all three give the same answer: the single-line diagram — ahead of the financial model and the datasheets. It is the one page that says what the plant electrically is.
Yet almost nothing written about single-line diagrams speaks BESS. The textbooks draw substations and motor feeders; a battery plant has things they never mention — bidirectional converters with a DC world behind them, and a revenue meter that spins both ways.
What a single-line diagram is — and why one line
A grid-scale BESS is a three-phase AC system from the PCS outward. Drawing all three phases of every circuit would triple the ink and add nothing at system level, because a healthy power system is electrically balanced — the phases carry the same story. So the drafting convention collapses them into one symbolic conductor: the single-line (or one-line) diagram.
At system level, nothing that matters is lost in the collapse. Every device on the line still carries its rating, its voltage class, and its protective function. What you gain is legibility: topology, fault paths, metering points, and ownership boundaries on a single page. That legibility is why the SLD — not the site plan, not the P&ID — is the drawing that interconnection review, protection studies, permitting, and commissioning all work from.
The grid-scale BESS SLD, from the grid down to the racks
Here is the canonical structure. Real projects vary — more feeders, different voltages, sometimes only one transformation stage on an MV-connected site or an extra one in the substation — but every utility-scale BESS single-line is a variation of this drawing:
Walk it top to bottom, the way power flows on discharge — and remember that on a battery plant every arrow reverses on charge:
- The POI — the dashed line at the top is not decoration. It is the ownership, compliance, and settlement boundary, and the revenue meter beside it spins both ways: a BESS is the rare plant whose SLD must make sense as both generator and load.
- HV disconnect and breaker — the visible isolation point the utility can lock out, then the fault-interrupting breaker (ANSI device 52).
- The main power transformer — one step of the site’s voltage ladder, here 138 kV to 34.5 kV. Its HV neutral is drawn earthed because the utility requires effective grounding at the interconnection; with the common YNd vector group, the delta MV side gets its own ground reference from a grounding transformer on the bus, its neutral taken to earth through a resistor. The scheme is drawn explicitly because it is protection-critical.
- The MV collection bus — the spine of the plant. Every feeder, and the station-service transformer that keeps cooling and controls alive, hangs off this busbar.
- The PCS — the box with ~ on the AC side and = on the DC side. It is rated in MVA, not MW, and it is the boundary between the AC world the utility sees and the DC world behind it.
- The battery — long-short plates, one symbol per DC block, up to 1500 VDC, each rack landing through its own fuse and DC disconnect. The SLD does not draw racks and modules; the one symbol, with its voltage and protective devices, stands for all of them.
One structural habit to notice: the drawing shows one feeder in detail and ghosts the rest (“feeders 2…N”). A 100 MW plant might have thirty-odd identical PCS/transformer skids, daisy-chained a few to a cable onto a handful of MV feeder breakers — drawing every skid adds pages, not information.
The same two products, a different plant
The SLD above is one plant. In practice the same two purchase orders — a DC block and a PCS — get arranged several ways, and the arrangement is the plant. Here are four, drawn from preliminary single-lines for one real pairing: a 2500 kW / 5016 kWh DC block and a 4200 kVA converter.
Two numbers on the converter deserve care, and the datasheet is stricter than the drawing. This converter is rated 4200 kVA/kW at 40 °C, nominal AC voltage and cos φ = 1 — at unity power factor the kVA figure and the kW figure are the same number. The 3780 kW on the drawings is that rating at 0.9 power factor, which is a design assumption about the operating point, not a property of the machine: the same unit is specified 0.5 leading to 0.5 lagging. Ask it for reactive power and the real power available falls, which is why the grid-code conversation and the MW figure are the same conversation — apparent power is the thing that is fixed, and the split between real and reactive is yours to choose.
The rating also carries an ambient with it. The same converter is 3900 kVA/kW at 50 °C. So an MW figure quoted without its temperature, its AC voltage and its power factor is not a rating at all; it is one row of a matrix with the row heading torn off.
The block carries its own duration — 5016 kWh over 2500 kW is 2.0 hours — and that figure is not the plant’s, except in the block-limited row, where the block count cancels and the plant inherits it exactly. The PCS count sets MW, the block count sets MWh, and duration is the quotient. That holds only while the last column stays under 100% — above it the blocks, not the converters, set the megawatts. Four blocks on one converter is a 5.3-hour plant whose blocks are never asked for more than 38% of their rated power. Three blocks on two converters is a 2-hour plant — and the one configuration here where the rule runs out, because two converters would ask for 101% of what three blocks can source. Same hardware, different market, and one of the four is not the plant its converter count suggests.
Three conditions belong on the whole table. The MW column is the 0.9 PF, 40 °C row and nothing more — at unity power factor the same four plants are 4.8, 2.0, 3.6 and 2.4 hours, and at 50 °C the converter-limited plants stretch longer again (the block-limited one holds its 2.0 hours until the derate takes the converters below the block limit). The drawings also state DC energy against AC power, so the quotient inherits that mixing. And 5016 kWh is nameplate DC energy at beginning of life; what a buyer can actually sell is less, after the usable-energy fraction, the round-trip losses — the datasheet rates this converter with its MV transformer at 98.0% peak and 97.53% CEC-weighted, and both are one-way figures, so a full round trip through the same chain is nearer 95% — and the auxiliary load that runs thermal management. The same mixing runs through the converters-ask column, which compares AC converter power against DC block power and so reads slightly low: the converter draws its output plus its own losses. None of this changes the shape of the table. All of it changes the hours you can contract for, which is why the round-trip efficiency and usable-energy definitions belong in the same conversation as the SLD.
That last column is worth a specification review of its own. A block asked for 38% of its rated power is a gentle duty and ages accordingly, but it is also power capability you cannot use — four blocks could deliver 10 MW and the converter will only ever ask for 3.78. The augmentation plan and the capacity warranty are both written as a fraction of nameplate energy, conditioned on the duty the plant actually runs, so this column is an input to both.
The block-limited row is worth dwelling on, because it is the one a summary block will not warn you about. The drawing for that configuration states 7.56 MW — two converters at their 0.9-PF figure — and the plant cannot deliver it, because three blocks can source 7.50 MW and the converters need rather more than that at their input to put 7.56 MW on the line. The shortfall is small, about 3%, and it is exactly the kind of thing that survives into a term sheet. When a sheet gives you an AC rating, multiply the block count by the block rating and check which one is smaller.
One drafting note that trips people reading a set for the first time: two of these drawings were the same rating. Four blocks landing on four DC inputs and four blocks paired onto two DC inputs are the same 3.78 MW / 20.06 MWh system. But the grouping is not a field cabling choice: Power Electronics builds the 4200 kVA frame as a two-input variant at 2295 A per input and a four-input variant at 1148 A per input — same total DC current, different part number, with the input count carried in the model code. What the choice moves is per-input cable size and the DC protection at each input; what it does not move is the plant’s MW or MWh. Read the summary block before concluding that two drawings describe two projects.
The symbols, decoded
The symbol set is standardized (IEC 60617, with the common ANSI/IEEE one-line conventions), which is exactly why an SLD travels between organizations without a translator. These are the ones that do the work on a BESS single-line:
The one that confuses newcomers most is the PCS box. On a solar SLD the inverter flows one way; on a battery SLD the same symbol is a two-way street, and everything upstream — breaker duty, protection settings, the meter — has to be rated and configured for both directions. If you remember one BESS-specific thing about the drawing, make it that.
Reading a preliminary SLD
Most single-lines you will be sent are not finished, and they say so. Learning the annotation grammar saves a round of questions whose answers are already on the page.
TYP — typical. The label applies to every segment in that class, drawn once. 1500 VDC (TYP) on one DC run means all of them; 34.5 kV (TYP) means the whole MV side. It is a
drafting economy, not a hedge, and the reviewer’s job is to check the class is really
homogeneous — one odd-length cable run or one different block, and the TYP is now a defect.
TBD — scoped, not yet filled. A device tagged DS-MV-1 (TBD) exists in the design
intent and has a tag reserved; its rating comes with a later issue. TBD on a tagged device
is normal at this stage. TBD on something untagged is a gap, because nothing is holding
its place.
Dashed outlines are read against the drawing’s own legend — on the canonical sheet higher up this page they mark the ownership boundary, the DC-block enclosure and a repeated feeder, none of which is preliminary. On these sheets they mark equipment whose rating waits on a study: the MV breaker’s interrupting rating is an output of the short-circuit study, and the arrester class an output of insulation coordination. The revenue meter is dashed for a different reason again — its specification comes from the utility’s metering standard, not from a study.
The issue stamp is the first thing to read, not the last. PRELIMINARY — NOT FOR CONSTRUCTION, with the title block’s sign-off boxes empty, tells you the numbers are design
intent rather than commitments. Firms label the stages differently — IFR and IFC, or a
percentage complete, or a house scheme — and the labels do not all mean the same thing: a
percentage measures how finished the sheet is, while IFR-to-IFC measures what it authorises.
The seal is a third question again, and in most US jurisdictions it lands at the permit set,
well before the construction issue. Read the stamp and the seal separately, and do not read
either as the other.
What is deliberately absent is information too. A preliminary sheet that shows no relay functions and no protection scheme is usually not behind schedule; the scheme follows the utility’s interconnection requirements, and the studies that come later fill in what those requirements leave open — so the relay detail is normally left generic until they land. What should worry you is the reverse — a sheet stamped for construction that still carries TBDs.
How to read a BESS SLD like a reviewer
Reading an SLD is not admiring it — it is interrogating it. The working sequence, the same one an interconnection or owner’s engineer runs:
- Start at the POI and trace down. Confirm the metering point, the ownership boundary, and which side of the boundary each device sits on. Misplaced boundaries surface here, at the worst possible time — contract review.
- Check the voltage ladder for consistency. Each transformer’s ratio must agree with the bus voltages above and below it, and with the interconnection study. A transformer nameplated 33 kV hanging off a 34.5 kV bus is exactly the error SLD review exists to catch: a quiet 4.5% over-excitation that eats nearly all of the transformer’s 5% C57 continuous excitation margin at rated load — and the moment the bus runs at its normal +5%, the core sits near 110% and saturates.
- Distinguish breakers from disconnects. Squares interrupt faults; blades provide visible isolation and must never be asked to interrupt a fault — and unless a switch is explicitly load-break rated, it must not be opened under load either. Every maintenance procedure and every lockout plan depends on this distinction being drawn correctly. And while you are at the squares, read their numbers: every breaker’s interrupting duty (kA) must clear the fault study’s worst case, and the transformer impedance (%Z) that sets that fault current belongs on the drawing too.
- Find the protection. CTs and VTs mark where the relays see the system — on the drawing above, the metering CT and VT at the POI feed the revenue meter, and the CT ring over the main breaker feeds the protection relays; alongside the ANSI device numbers (50/51 overcurrent, 87 differential on the main transformer) they tell you whether every fault has a device assigned to clear it — the question a protection-coordination study answers formally.
- Follow the auxiliary power. The station-service path keeps HVAC, controls, and the BMS alive. A reviewer checks what happens to it when the main breaker opens — a plant that loses cooling when it disconnects has a design problem drawn in plain sight.
- Don’t stop at the PCS. The DC side is part of the same drawing: every rack lands through a fused disconnect, and each DC block has an isolation means the commissioning and lockout plans depend on. A BESS SLD that goes quiet below the inverter is incomplete — the fuse and disconnect symbols from the legend reappear here.
- Check the grounding scheme. Neutral earthing at the transformers sets ground-fault current and touch-voltage behaviour; the SLD is where that scheme is declared.
Ten minutes of this discipline on one page catches errors that would cost months found later — which is precisely why everyone asks for the SLD first.
Where the SLD shows up in a project’s life
The same drawing recurs at every gate, at increasing levels of detail: a preliminary single-line in the interconnection application; the POI one-line exhibit attached to the interconnection agreement (the contractual version — the one where the boundary and metering point are frozen); the permit set the AHJ reviews under NFPA 855 alongside the fire-safety documents; the protection-coordination and arc-flash studies built directly on it; and finally commissioning, where the as-built SLD is walked against physical reality, breaker by breaker. It is also the drawing every due-diligence engineer opens first when a project changes hands — and to see one operate, step through the interactive black start of a grid-forming battery, which energizes exactly the chain drawn above, bus by bus.
For the terms behind the symbols, the glossary entries on the one-line diagram, the PCS, transformers, and the MV/LV/HV ladder go deeper on each device — and the switchgear guide covers what actually sits inside those breaker squares.
FAQ
What is a single-line diagram in a BESS project?
A one-page schematic that draws the plant’s balanced three-phase power system as a single symbolic line, so the whole electrical chain — battery racks, PCS, transformers, switchgear, meter, point of interconnection — can be read at a glance with every device’s rating and protective role. It is the first document the interconnecting utility, the AHJ, and every technical reviewer ask for.
What is the battery symbol on a single-line diagram?
Alternating long and short parallel lines — the classic electrochemical cell symbol. On a grid-scale BESS SLD one battery symbol usually stands for a whole DC block or group of racks, labelled with its voltage (commonly up to 1500 VDC) and energy; nobody draws thousands of cells.
What is the difference between a single-line and a three-line diagram?
A single-line diagram collapses all three phases into one conductor to show topology, ratings, and protection — it is the system-level view. A three-line diagram draws each phase (and neutral) explicitly and is used for construction, wiring, phase-specific protection, and anywhere unbalance matters. You design and review on the single-line; you build from the three-line.
The Grid-Scale BESS: Complete Guide walks the full electrical chain this drawing describes — from cell chemistry to the POI — with the same practitioner’s eye.