Solar PV Array for Foundry: How Many Panels Does a 1 MW Induction Furnace Need

Solar PV Array for Foundry: How Many Panels Does a 1 MW Induction Furnace Need

A 1 MW induction melting furnace sounds like it needs a 1 MW solar array. It does not, for two reasons. First, the solar array produces energy, not power, and the energy has to match the consumption over time, not instantaneously. Second, the foundry operates for a limited number of hours per day, and the solar produces for a limited number of hours per day. The match between the two determines the right array size. Here is how the sizing actually works.

Start with the load.

A 1 MW induction melting furnace running two shifts per day (16 hours) consumes 8 to 16 MWh of energy per day, depending on the production rate and the furnace efficiency. The peak demand is 1 MW (during the melt), but the average demand is 0.5 to 1.0 MW over the operating day.

For a 24/7 foundry, the energy consumption is 12 to 24 MWh per day. The peak demand is still 1 MW, but the average demand is 0.5 to 1.0 MW around the clock.

The load profile varies - morning startup, first melt, pouring, second melt, afternoon pouring, evening cleanup, night shutdown (for the two-shift case). The instantaneous demand varies from 100 kW (during cleanup) to 1000 kW (during the melt).

The solar resource determines the generation.

The solar resource depends on the location. A 1 MW PV array (using 2026-vintage monocrystalline silicon modules at 22 to 24 percent efficiency) in different locations produces:

- US Southwest (Arizona, Nevada, Southern California): 1800 to 2200 MWh per year

- Southern Europe (Spain, Italy, Greece): 1400 to 1800 MWh per year

- Central Europe (Germany, France, Northern Italy): 1000 to 1300 MWh per year

- North China, Korea, Japan: 1100 to 1500 MWh per year

- Southeast Asia (Vietnam, Thailand, Malaysia): 1300 to 1600 MWh per year

- Middle East, North Africa: 1800 to 2300 MWh per year

- Northern Europe (UK, Netherlands, Scandinavia): 800 to 1100 MWh per year

The specific yield (kWh per kW per year) ranges from 800 to 2300, depending on the location. A sunny, low-latitude, high-altitude, low-cloud location is at the top of the range. A cloudy, high-latitude location is at the bottom.

The match between load and generation.

The array size depends on the desired match between the load and the generation. The three common design points are:

Design point 1: Match the average load. If the foundry's average load is 700 kW, a 1.4 MW array in a sunny location (producing 1800 MWh per year per MW) produces 2520 MWh per year. The foundry consumes 700 kW x 16 hours x 365 days = 4088 MWh per year. The solar fraction is 2520 / 4088 = 62 percent. The foundry needs additional energy from the grid or the BESS for the other 38 percent.

Design point 2: Match the peak demand. A 1 MW array in a sunny location produces 1800 MWh per year. The foundry consumes 4088 MWh per year. The solar fraction is 44 percent. The foundry needs additional energy for the 56 percent not covered by solar.

Design point 3: Match the annual consumption. To cover 100 percent of the foundry's energy with solar, the array has to produce 4088 MWh per year. In a sunny location, that requires a 2.3 MW array. In a moderate location, a 3 MW array. In a cloudy location, a 5 MW array.

The land area for the array.

A 1 MW PV array occupies roughly 4 to 6 hectares (40,000 to 60,000 square meters) of land using 2026-vintage modules at 22 to 24 percent efficiency. The land use is the major constraint for a large array.

A 2 MW array needs 8 to 12 hectares.

A 3 MW array needs 12 to 18 hectares.

A 5 MW array needs 20 to 30 hectares.

For a foundry that does not have 20 to 30 hectares of available land, a ground-mount array is not feasible. Options include:

- Rooftop solar on the foundry building and the adjacent warehouses (typically 0.5 to 2 MW)

- Agrivoltaics (combining solar with agriculture on the same land)

- Solar carport (solar panels over parking lots)

- Off-site solar (utility-scale solar farm with a PPA)

A 1 MW rooftop solar system on a large foundry building is technically feasible if the roof is large enough and structurally rated. The roof area needed is roughly 5,000 to 6,000 square meters (a 70 by 80 meter roof). Many foundries have roofs this large.

The panel technology choice.

In 2026, the dominant PV module is monocrystalline silicon passivated emitter and rear cell (PERC) or tunnel oxide passivated contact (TOPCon). The efficiency is 22 to 24 percent. The cost is $0.20 to $0.30 per watt for the modules.

Bifacial modules (modules that produce power from both sides) are increasingly common. The bifacial gain is 5 to 15 percent depending on the ground reflectivity (albedo). Bifacial modules on a white ground cover or on a reflective rooftop can produce 10 to 20 percent more energy than monofacial modules.

Tracking systems (single-axis trackers that follow the sun during the day) can increase the energy production by 15 to 25 percent compared to fixed-tilt systems. Trackers cost more per watt and require more maintenance, but the energy gain is meaningful.

For a foundry solar array, the choice depends on the site. A ground-mount array in an open field benefits from trackers. A rooftop array uses fixed-tilt (because the roof slope is fixed and the structural loading of a tracker is too high). A carport or agrivoltaic array can use either.

The cost of the array.

The installed cost of a utility-scale solar PV array in 2026 is in the range of $0.50 to $0.80 per watt (DC) for a ground-mount system with trackers, or $0.60 to $1.00 per watt for a rooftop system. The cost includes the modules, the inverters, the mounting structure, the electrical infrastructure, and the installation.

A 2 MW ground-mount array in a sunny location costs roughly $1.0 to $1.6 million. A 3 MW array costs $1.5 to $2.4 million. A 5 MW array costs $2.5 to $4.0 million.

The LCOE (levelized cost of energy) of the solar array is the relevant comparison number. The LCOE is the total life-cycle cost divided by the total life-cycle energy production.

For a 2 MW array in a sunny location:

- Total cost: $1.2 million

- Annual energy production: 3600 MWh

- Annual O&M cost: $12,000 (1 percent of capital)

- Life: 30 years

- Discount rate: 8 percent

- LCOE: roughly $25 to $35 per MWh ($0.025 to $0.035 per kWh)

This is competitive with grid electricity in many regions, and significantly cheaper than grid electricity in regions with high tariffs.

The bottom line on PV sizing. A 1 MW induction furnace does not need a 1 MW solar array for 100 percent solar fraction. It needs a 2 to 5 MW array, depending on the location and the operating hours. The land area is 8 to 30 hectares. The cost is $1 to $4 million. The LCOE is $25 to $35 per MWh in a sunny location, which is competitive with grid electricity. The trend in 2026 is toward larger arrays, higher efficiency modules, and lower LCOE. Solar PV is now a no-brainer for foundries in sunny regions with available land. The economics are tightening in less sunny regions, and the trend will continue as module costs drop.

Author: MONTE INTELLIGENCE solar PV engineering team. For PV array sizing and economic analysis, contact helenxu@cnlymonte.com.