Roof-Mounted PV to Offset an Irrigation Pump

Hydraulic power.

hp_hyd = (gpm × psi) ÷ 1,714 = (800 × 60) ÷ 1,714 = 28.0 hp

Shaft power. 28.0 ÷ 0.70 = 40.0 hp at the pump shaft.

Electrical input. 40.0 ÷ 0.92 = 43.5 hp = 43.5 × 0.746 = 32.4 kW.

Seasonal energy. 32.4 kW × 1,400 h = ≈45,400 kWh.

Two things worth flagging now, before any array is sized:

• The pump draws 32 kW whenever it runs, but it only runs 1,400 h out of a ~120-day season - about 11.7 h/day average. The load is not a flat annual block; it is a summer-only, partly daytime load.
• A 32-kW motor also carries a utility demand charge that a PV array does not erase; the array reduces kWh and, only at the margin and only if the sun is out, coincident demand.

First-pass sizing. At roughly 5.2 peak-sun hours/day, a performance ratio of 0.80 and a 120-day season, a 1-kW array yields about 5.2 × 0.80 × 120 ≈ 500 kWh/season. To match 45,400 kWh:

45,400 ÷ 500 = ≈91 kW

Area check: at ~20% module efficiency that is roughly 455 m² ≈ 4,900 ft² of module area, which does fit inside the 7,200-ft² south plane.

But the objective is wrong. The match is on annual energy, not on the load's shape:

• At noon the 91-kW array produces about 73 kW while the pump draws 32 kW - most of the output is exported.
• In the evening and at night the pump imports at retail, and pumping is often scheduled at night to reduce wind drift and evaporation loss.
• The exports are credited at retail only because net metering happens to allow it, capped at 110% of annual use. That is a policy exposure, not a physical property of the system. Anything beyond the cap earns 3.5 ¢/kWh.

Better design. Size to the daytime load - roughly 60–75 kW - and shift irrigation toward daylight hours so the array serves the pump directly. The last increment of array capacity is displacing 3.5 ¢ exports in a few years and adds cost for no return. Also verify the interconnect: 91 kW on a rural service may require a study and transformer change, which can dwarf the array cost.

Recommendation to the owner: present the 60–75 kW option with daylight pumping as the base case, and show the 91-kW case only with net metering assumed to persist.

Dead load. Modules plus racking plus ballast or attachments run about 3.5–4 psf over 7,200 ft²: 25,000–29,000 lb added to a roof explicitly documented as having no reserve. Per truss at 8-ft spacing, tributary 8 × 60.5 = 484 ft² × 4 psf ≈ 1,940 lb of new dead load on each truss - small-sounding per square foot, not small summed over the span.

Effects beyond simple dead load:

• Snow behavior changes. Modules shed snow to the eave and create a drift at the lower edge; the array also blocks the uniform redistribution the original design assumed. The controlling load case may be a drift the truss was never checked for.
• Wind uplift on modules is concentrated at corners and edges and is transferred into purlins and trusses as point loads, not uniform load.
• Deflection and ponding - a more flexible roof under added load can pond, which is progressive.

What I would require before signing anything:

1. A sealed structural evaluation by a licensed structural engineer of the actual as-built truss and purlin configuration with the array dead load, drift and uplift applied.
2. If the roof cannot take it, a ground mount - which also gives a better tilt, no penetrations, no leak risk, and easier service. On an ag site with land available this is usually the right answer.
3. No penetrations justified by "it's only 4 psf," and no reliance on a verbal assurance from a truss supplier.

The energy analysis is the easy half. The roof is where this project fails if it is not checked first.