Design study

79 kWp solar PV and 63 kWh battery system

The mechanical layout, battery control and single line drawings for a 78.88 kWp PV and 63.18 kWh storage design, with an hourly dispatch model.

Drawings and dispatch model

A rooftop PV and battery design study for a commercial unit in Islamabad: 136 Jinko 580 W modules at 78.88 kWp on an 80 kW FoxESS P3-80-Plus hybrid inverter, with nine FoxESS CQ7 modules in series giving 63.18 kWh at 518.4 V.

Three A1 sheets drawn in AutoCAD 2027: the single line diagram, the mechanical layout and the battery control sheet. Each is shown whole, with a closer view of the parts worth reading. Click any sheet to open it full size, or download the PDF and DWG at the end.

All files for this project on GitHub

2. The key diagram. The bottom left corner of PV-04. Array and battery into the hybrid inverter, out to the grid on one side and the essential board on the other, with the device tags the rest of the sheet uses. It shows the shape of the system before the detail

3. The point of connection. Where the system joins the building. A new 160 A way on the existing busbar, opened by the interface relay through a shunt trip, the existing 630 A incomer, the utility transformer outside the board with the point of supply marked, the import export meter, and the current transformer whose signal limits export

4. The mechanical layout. PV-01, the whole sheet. The roof plan with eight rows of seventeen modules set out 3,000 mm clear of the south parapet at a 1,800 mm pitch, the mounting section, the plant room, the notes, the equipment schedule and the design summary

5. The mounting section and row spacing. Detail A at 1:20, south on the left. Ballasted frames at 15 degrees, with the front rail 300 mm and the rear rail 594 mm above the membrane. The dashed line is the shading ray on 21 December at 09:00: at its 24.8 degree profile angle the shadow ends 69 mm short of the next row, which is what sets the 1,800 mm pitch

6. The battery control and performance sheet. PV-03, the whole sheet. The control strategy, a typical April day and the December day that sets the annual peak, the annual performance with and without the battery, and how the overnight charge target trades the battery's jobs against each other

7. A year of dispatch in two days. The same two days from the model, redrawn to read on a screen. In April grid import stays at or below the 55 kW target all day. In December there is no PV left by the afternoon and the state of charge is pinned at the floor, so the import climbs to the 75.2 kW that sets the annual peak

How the battery was sized. The first thing the model said was that the battery was not worth fitting. Charged only from surplus PV it cycled about 13 times a year, because the site load is 3.3 times the generation and there is almost no surplus to store. The battery only earns its place if it is charged off-peak and given the job of carrying the site through the daily outage.

Dispatch was modelled hour by hour for a year in plain Python with no libraries, so every number can be traced by reading the file. Three jobs, in priority order: hold a reserve for the essential board through the evening outage, keep grid import under the maximum demand target, and soak up whatever PV would otherwise be exported.

Period	Action	Limit
00:00 to 05:00	Charge from the grid at the off-peak rate	Stop at 75% SoC
06:00 to 18:00	Charge from PV surplus	Up to 100%
06:00 to 18:00	Discharge only to hold grid import down	Not below 55%
18:00 to 20:00	Grid off, EPS board runs on PV and battery	Down to 10%
20:00 to 06:00	Discharge if import would exceed the target	Down to 10%

Over a year	Without battery	With battery
PV generated	116,742 kWh	116,742 kWh
Site load	383,139 kWh	383,139 kWh
Grid import, total	240,188 kWh	250,546 kWh
of which off-peak charging	none	12,649 kWh
Annual grid peak	75.2 kW	75.2 kW
Hours over the 55 kW target	1,078	695, down 35.5%
Outage demand met	0%	96.8%
Battery throughput	none	12,729 kWh, 223.9 full cycles
Round-trip loss, battery only	none	641 kWh

Total import rises because off-peak charging is itself an import. Of the 12,729 kWh the battery delivers in a year, 10,438 kWh carries the essential board through the evening outage and 2,291 kWh cuts daytime import above the target. The 55% floor holds 1.9 hours of reserve against a 2 hour outage, so 96.8% of outage demand is met. A higher floor would cover more of the outage and leave less for daytime shaving.

The annual peak does not move. It falls on a December afternoon when there is no PV left and the state of charge is pinned at the 55% floor that protects the 18:00 outage. Hours spent over the target do fall, by 35.5%. Lowering the maximum as well needs a second CQ7 stack on port two, or an agreement to accept a shorter backup through the winter. That is a commercial decision, not an electrical one.

How full to charge off-peak is the one setting that trades the three jobs against each other. Charging fuller shaves better and covers the outage better, but leaves no room for the midday PV surplus. At 100% the battery takes no PV at all.

Off-peak charge target	Outage met	Hours over 55 kW	PV stored	Cycles a year
60%	95.5%	965	1,106 kWh	193.5
75%, the one drawn	96.8%	695	720 kWh	223.9
90%	97.4%	503	305 kWh	247.2
100%	97.7%	390	none	260.2

The cable schedule and the interface protection settings as they appear on PV-04. Lengths are design figures for this building. Cable sizes and volt drop are to be confirmed on the actual route, and the relay settings with IESCO, so the sheet stays at revision P3 until they are. Until then the relay values below are working figures, not approved settings.

Tag	Cable	Length	Protection
W1	16 × 1C 6 mm² PV1-F	45 m	None, one string per MPPT
W2	2 × 1C 35 mm² Cu XLPE	8 m	QF1, 125 A DC, 2 pole
W3	4C 95 mm² Cu SWA	32 m	QF2, 160 A MCCB
W4	4C 35 mm² Cu	12 m	QF4, 63 A MCB
W5	1C 50 mm² green/yellow	20 m	Main earthing conductor

Function	Setting	Trip time
27 under voltage	0.80 Un	2.5 s
27 stage 2	0.50 Un	0.5 s
59 over voltage	1.15 Un	1.0 s
81 under frequency	47.5 Hz	0.5 s
81 over frequency	51.5 Hz	0.5 s
ROCOF	1.0 Hz/s	0.5 s
Reconnect delay	After the grid returns	180 s

Equipment figures come from the FoxESS P3 Plus and CQ7 datasheets and the Jinko 72HL4-V datasheet. The load profile, the outage window and the tariff periods are design assumptions, declared at the top of the model, and would be replaced with half-hourly meter data before anything was ordered.

PV-01 mechanical (.pdf) PV-01 mechanical (.dwg) PV-03 battery control (.pdf) PV-03 battery control (.dwg) PV-04 single line diagram (.pdf) PV-04 single line diagram (.dwg)