PV Loss Waterfall Analysis: Understanding the 10 Categories of Energy Loss

Every engineer who has modeled a PV system knows the moment of truth: the "waterfall" chart that shows how theoretical irradiance becomes actual delivered energy at the point of interconnection. A 1,800 kWh/kW/yr irradiance resource can shrink to 1,400 kWh/kW/yr of net AC energy — a 22% loss that is the sum of a dozen independent mechanisms.

Understanding each loss category — its typical magnitude, its root causes, and how to reduce it — separates a bankable energy yield estimate from an over-optimistic one. This guide breaks down the ten standard loss categories used in PVsyst and industry-standard energy yield models, with real-world values and mitigation strategies for each.

1. Soiling Loss (0.5-5% annual average)

Soiling — the accumulation of dust, pollen, bird droppings, and other particulates on module surfaces — is the most variable of all loss categories. It depends on rainfall frequency, soil type, wind patterns, agricultural activity, and module tilt angle.

For utility-scale projects in dry climates:

• Southwestern US desert sites: 3-5% annual average without washing, 1-2% with quarterly washing
• California Central Valley: 2-4% annual average, influenced by agricultural dust
• Texas Panhandle / West Texas: 2-5% depending on proximity to farming and wind conditions
• European sites with regular rain: 0.5-2% annual average

Soiling is not uniform across the array. Edges and bottom rows tend to accumulate more dust. Bifacial modules also experience rear-side soiling — typically 1-3% additional loss for the rear side, depending on ground cover and wind patterns. Robotic cleaning systems can reduce soiling loss to under 1% at a cost of roughly $0.15-0.30/module/year.

2. Shading and Near-Scattering Loss (0.5-5%)

Shading from adjacent rows, equipment buildings, fences, terrain features, and the module frame itself reduces the irradiance reaching the active cell area. For well-designed utility-scale projects, row-to-row shading loss is typically 0.5-2% with optimized backtracking. Near-scattering — light that hits the module frame or racking structure near the active cell area — adds another 0.3-0.8%.

Key factors:

• GCR and row spacing — tighter spacing increases shading at low sun angles
• Tracker backtracking algorithm quality — some proprietary algorithms reduce shading by 20-30% compared to simple geometric backtracking
• Terrain slope — uneven terrain can create unexpected shading patterns

3. Reflection (IAM) Loss (2-4%)

The incidence angle modifier (IAM) accounts for the fact that module glass reflects more light at high incidence angles (early morning, late afternoon, winter). At normal incidence (sun directly overhead), reflection loss is ~2-4% for standard AR-coated glass. At 60° incidence angle, it rises to ~10%. At 80°, it exceeds 40%.

IAM loss depends on:

• Module glass AR coating quality — premium coatings reduce reflection by 30-50%
• Module tilt — steeper tilts increase IAM loss in winter, decrease it in summer
• Tracker vs fixed-tilt — trackers keep the module more normal to the sun, reducing IAM loss

Fixed-tilt systems at 20-30° tilt: IAM loss 2.5-3.5%. Single-axis tracking: IAM loss 2-3%.

4. Spectral Mismatch Loss (0.5-2%)

PV modules are rated under AM1.5G standard spectrum, but real sunlight varies with air mass, water vapor, aerosols, and cloud cover. When the actual spectrum shifts relative to the module's quantum efficiency curve, spectral mismatch loss occurs.

Broad-spectrum modules (high-efficiency multi-crystalline or mono-crystalline with good blue response) have lower spectral mismatch than thin-film modules with narrow absorption bands. Bifacial modules also experience spectral mismatch on the rear side, where the reflected spectrum is shifted by the ground surface's reflectance properties.

5. Temperature Loss (3-8%)

Temperature loss is often the single largest loss category. Crystalline silicon modules lose approximately 0.35-0.45% of rated power per degree Celsius above 25°C. At 75°C cell temperature (common on a hot summer day), the module operates at roughly 82% of STC rated power.

Annual temperature loss varies by climate:

• Phoenix, AZ (hot desert): 6-8% annual average, with summer peaks over 15%
• Los Angeles, CA (mild): 3-5% annual average
• Germany (cool climate): 2-3% annual average

Bifacial modules run slightly cooler than monofacial modules (1-3°C lower operating temperature) because the rear side radiates heat to the ground. This temperature benefit is worth roughly 0.3-1.0% additional energy yield — a secondary advantage of bifacial technology beyond the rear-side irradiance gain.

6. Module Quality and Mismatch Loss (1-3%)

Module quality loss covers the difference between nameplate rating and actual measured power (positive tolerance helps — modules rated at 650 W typically test at 645-660 W, with premium manufacturers clustering tightly around the rated value). LID (light-induced degradation) and LeTID (light- and elevated temperature-induced degradation) cause initial power loss of 1-3% in the first weeks of operation, which is separate from long-term degradation.

Mismatch loss between modules in the same string arises from manufacturing tolerances (±3-5% for standard modules, ±1-3% for premium bin-sorted modules). A string of 26 modules with ±3% tolerance can experience 0.5-1.5% mismatch loss. Higher precision binning reduces this to 0.3-0.8%.

7. Wiring (Ohmic) Loss (0.5-2%)

Ohmic losses from DC cables, connectors, fuses, and combiner boxes are a function of cable length, cross-section, and current. In well-designed utility-scale projects, DC wiring loss is typically designed to stay under 1.5% at STC (which translates to ~0.5-1% annual loss because the system operates below STC most of the time).

AC wiring loss on the medium-voltage collection system adds another 0.3-0.8%, depending on the distance from inverters to the step-up transformer and the main transformer to the point of interconnection.

8. Inverter and MPPT Loss (1.5-3%)

Inverter efficiency varies with load level. Modern central inverters achieve peak efficiency of 98.5-99% at 50-80% load, but efficiency drops to 96-97% at light load (10-20%) and 97-98% at full load. The annual weighted average inverter efficiency — accounting for the distribution of inverter loading across the year — is typically 97.5-98.5%.

MPPT tracking efficiency is typically 99.5-99.9% for modern inverters with fast MPPT algorithms. However, multi-peak MPPT scenarios (partial shading) can reduce this to 95-98% if the tracker settles on a local maximum instead of the global maximum.

9. Clipping and Curtailment Loss (0-8%)

Clipping loss — energy lost because DC power exceeds the inverter's AC rating — is a deliberate design trade-off (see our PV system sizing guide for the economics). At the optimal ILR of 1.25-1.45, annual clipping loss is 1-4%.

Curtailment loss — energy that the system could have produced but that was not taken by the grid — varies widely by market. In ERCOT (Texas), curtailment can reach 2-5% for projects with high solar penetration in the same grid node. In CAISO, curtailment is 3-8% for new projects, driven by the duck curve and transmission constraints.

10. Transformer and Auxiliary Loss (1-3%)

Step-up transformers (typically 34.5 kV to 115-230 kV for utility-scale) have efficiency losses of 0.5-1.5%, depending on loading and transformer design (amorphous core vs grain-oriented silicon steel). Fixed losses (iron losses) occur whenever the transformer is energized, regardless of load. Variable losses (copper losses) scale with the square of the current.

Auxiliary power consumption — trackers, inverters (cooling fans and control electronics), SCADA, lighting, security, and module cleaning systems — adds 0.5-1.5% of annual energy. Inverters with active cooling fans consume 200-500 W each during operation, totaling 10-50 kW for a 100 MW plant.

Full Waterfall Example: 100 MW Utility Project

Let's build a complete loss waterfall for a 100 MW_AC single-axis tracking project in the California Central Valley (ILR 1.35, 650 W bifacial modules, central inverters, quarterly washing):

• POA irradiance: 2,200 kWh/m²/yr (100%)
• − Soiling: 2.0% → 2,156 kWh/m²/yr
• − Shading: 1.2% → 2,130 kWh/m²/yr
• − IAM: 2.8% → 2,070 kWh/m²/yr
• − Spectral mismatch: 0.8% → 2,054 kWh/m²/yr
• − Temperature: 5.5% → 1,941 kWh/m²/yr
• − Module quality + mismatch: 2.0% → 1,902 kWh/m²/yr
• − DC wiring: 1.2% → 1,879 kWh/m²/yr
• − Inverter + MPPT: 2.0% → 1,842 kWh/m²/yr
• − Clipping: 3.5% → 1,777 kWh/m²/yr
• − Curtailment: 3.0% → 1,724 kWh/m²/yr
• − Transformer + aux: 2.2% → 1,686 kWh/m²/yr

Net AC yield: 1,686 kWh/kW/yr — a total of 23.4% loss from POA irradiance to the meter. Each percentage point of loss is worth approximately 1,000 MWh/yr for this 100 MW plant — at $40/MWh, that's $40,000/year in lost revenue per percentage point. Optimizing even the small categories (spectral mismatch, wiring, mismatch) can add hundreds of thousands of dollars in project value.

How Energy Optima Builds the Waterfall

Energy Optima generates a complete loss waterfall for every simulation, with hourly computation of each loss category and annual aggregation. The platform provides both a visual waterfall chart and a detailed numeric breakdown, with the ability to drill down into any category to see the hourly drivers.

Users can compare loss waterfalls across multiple design scenarios (different modules, tilts, trackers, inverters, ILR values, and cleaning schedules) to identify which configuration minimizes total losses and maximizes net yield — directly fed into the LCOE and IRR calculations for economic comparison.