The 10 PV Loss Categories and How Energy Optima Models Them

A 100 MW solar PV plant in Arizona with a 2% soiling loss, 3% shading loss, and 1% mismatch loss can lose 16% of its theoretical yield before inverter losses — translating to ~$1.2M/year in lost revenue at $0.04/kWh PPA. Each percentage point of loss compounds through the waterfall, and the difference between a 1,550 kWh/kW/yr and a 1,700 kWh/kW/yr design can swing project IRR by 2-3 points.

Energy Optima's PV designer models ten distinct loss categories, computed hourly from site-specific meteorological data, module parameters, and system topology. The platform's component database — covering 136 PV modules from 30 manufacturers and 200 string inverters from 20 manufacturers — provides the precise electrical and thermal parameters needed to calculate each loss mechanism from first principles rather than default multipliers.

This guide walks through each of the ten loss categories, explains how Energy Optima calculates them, provides typical values for common conditions, and shows how to determine site-specific inputs. We work through a complete 50 MW PV plant in Dubai as a running example and show how the loss waterfall visualization reveals the highest-impact opportunities for optimization.

The 10 Loss Categories at a Glance

Energy Optima organizes PV system losses into ten categories that cascade from plane-of-array (POA) irradiance to net AC energy. Each is computed independently at hourly resolution and then aggregated to the annual waterfall.

#	Loss Category	Typical Annual Range	Primary Inputs
1	Age Loss (Degradation)	0.3-0.7%/yr	Module manufacturer degradation warranty, year of simulation
2	Soiling	0.5-5%	Rainfall data, cleaning schedule, regional dust index
3	Shading	0.5-5%	GCR, row spacing, backtracking algorithm, terrain model
4	Mismatch	0.5-2.5%	Module binning tolerance, string configuration, partial shading
5	DC Wiring	0.5-2%	Cable lengths, cross-sections, string currents, temperature derating
6	Connections	0.3-1%	Number of connectors, fuses, combiner boxes, contact resistance
7	Availability	0.5-3%	Inverter MTBF, scheduled maintenance, grid outage statistics
8	IAM (Reflection)	2-4%	Module glass AR coating, tilt angle, tracker type
9	Spectral Mismatch	0.5-2%	Module QE curve, air mass, aerosol optical depth, water vapor
10	Temperature	3-10%	NOCT, ambient temperature, wind speed, mounting type

1. Age Loss (Degradation)

Age loss accounts for the irreversible decline in module performance over the project life. Crystalline silicon modules typically degrade 0.4-0.7% per year in the first year (due to LID and LeTID stabilization) and 0.3-0.55% per year thereafter. Energy Optima models age loss using the manufacturer's warranted degradation curve — typically a linear or bilinear profile — applied as an annual multiplier to the module's STC rating.

For project-finance grade analysis, Energy Optima allows two approaches:

• Linear degradation — constant annual factor (e.g., 0.45%/yr for a module with a 30-year 80% end-of-life warranty)
• Bilinear degradation — separate first-year rate (typically 1-2% to capture LID) and steady-state rate thereafter

When running a year-1 energy estimate, the age loss contribution is zero. For a 25-year PPA model, the average annual age loss is typically 0.35-0.55% depending on the module's degradation warranty. Energy Optima's component database stores the warranted degradation rate for each of its 136 PV modules and applies it automatically in multi-year simulations.

2. Soiling Loss

Soiling loss is the reduction in transmitted irradiance caused by dust, pollen, bird droppings, and other particulate accumulation on the module surface. It is the most site-dependent of all loss categories and the one where operator decisions (cleaning frequency and method) have the greatest influence.

Energy Optima models soiling as a monthly or seasonal factor applied to POA irradiance, with a soiling rate (percent loss per day) and a cleaning reset model. The default soiling rate resets to zero after each cleaning or significant rainfall event (>5 mm/day).

Soiling by Region

Typical annual soiling losses by climate zone, based on published field studies and Energy Optima's project database:

• Middle East / Arabian Peninsula (Dubai, Riyadh, Doha): 3-6% without cleaning, 0.8-1.5% with monthly robotic cleaning. High dust loading, infrequent rainfall (50-100 mm/yr), frequent sandstorms.
• Southwestern US (Arizona, Nevada, New Mexico): 2-5% without washing, 0.5-1.5% with quarterly washing. Low rainfall but less industrial dust than the Middle East.
• North Africa (Morocco, Egypt): 4-8% in desert fringe areas without cleaning. Proximity to Sahara dust sources increases the baseline.
• India (Rajasthan, Gujarat): 3-7% without cleaning during dry season. Monsoon provides natural cleaning for 3-4 months.
• Northern Europe (Germany, UK): 0.5-2% annual average. Regular rainfall keeps modules relatively clean. Most projects rely on natural cleaning.
• Southeast Asia (Thailand, Vietnam): 0.5-2% during wet season, 2-4% during dry season. Trade-off between heavy rainfall cleaning and higher humidity-related soiling adhesion.

For our Dubai 50 MW example, we use 1.0% soiling loss based on monthly robotic cleaning — a reasonable assumption for a utility-scale plant in the region.

3. Shading Loss

Shading loss captures the reduction in irradiance reaching the active cell area from row-to-row self-shading, horizon obstructions, nearby structures, and terrain features. For single-axis tracker projects, Energy Optima computes shading at sub-hourly time steps using the site geometry: ground cover ratio (GCR), tracker rotation limits, backtracking algorithm, and site topography.

The shading calculation in Energy Optima accounts for:

• Row-to-row shading — driven by GCR and sun position. For a typical tracker with GCR 0.35-0.40 and 60° rotation limits, annual shading loss is 0.5-1.5%.
• Backtracking algorithm — the quality of the backtracking algorithm affects whether shading is eliminated or merely reduced. Energy Optima uses a true geometric backtracking model that computes the minimum backtrack angle for zero row-to-row shadow at each time step.
• Horizon profile — far shading from terrain, buildings, or vegetation is modeled via a horizon line uploaded from site survey data or extracted from a digital elevation model.

For our Dubai example on flat terrain with GCR 0.38 and backtracking, we estimate 0.8% shading loss.

4. Mismatch Loss

Mismatch loss arises when modules connected in the same string operate at different currents or voltages, forcing the string to the lowest common denominator. The two primary sources are manufacturing tolerance variation and partial shading within a string.

String-Level Mismatch Modeling

Energy Optima models mismatch at the string level using a statistical approach. For a string of N modules with power tolerance ±t%, the mismatch loss depends on:

• Binning quality — premium manufacturers bin modules to ±1-3%, while standard modules may have ±3-5% tolerance. Energy Optima's component database stores the actual binning specification for each of its 136 modules.
• String length — longer strings have more statistical cancellation of variation. A 30-module string with ±3% tolerance has lower mismatch loss than a 15-module string, assuming random distribution.
• Partial shading — even a single shaded cell in a module can reduce the entire string current if bypass diodes are not activated. Energy Optima models bypass diode activation at the sub-string level (typically 3 bypass diodes per module).

Typical mismatch loss for a well-designed system using premium bin-sorted modules: 0.3-0.8%. For standard modules: 0.8-1.5%. In our Dubai example with premium bifacial modules (600 W, ±2% tolerance, 28 modules per string), we estimate 0.6% mismatch loss.

5. DC Wiring Loss

DC wiring loss covers ohmic (I²R) losses in the DC cables from module strings to combiner boxes and from combiner boxes to inverter inputs. Energy Optima calculates these losses using the actual cable parameters selected in the design:

• Cable length from string to combiner (based on array layout)
• Cable cross-section (4 mm², 6 mm², 10 mm² standard)
• Number and type of home-run cables from combiner to inverter
• Cable operating temperature (resistance increases ~0.393%/°C)

Energy Optima computes DC wiring loss at each hour using the actual string current, not the STC current. This is important because wiring loss scales with the square of current, and the system operates below STC current for most of the year. A 1.5% wiring loss at STC may translate to only 0.8-1.2% annual loss.

For a well-designed 50 MW plant with string inverters distributed across the array, DC wiring loss is typically 0.8-1.5%. Our Dubai example uses 1.0%.

6. Connections Loss

Connections loss accounts for contact resistance at every electrical junction in the DC circuit: MC4 connectors, fuse holders, disconnect switches, combiner box busbars, and inverter DC terminals. Each connection adds a small resistance — typically 0.5-5 mΩ per connection depending on connector type and installation quality.

Energy Optima calculates connections loss as:

• Number of connections — derived from the string and combiner configuration. Each string has 2 module connectors, 1 fuse holder, 1 combiner box terminal per polarity, plus the DC disconnect and inverter terminals.
• Contact resistance per connection — default values for MC4 (stamped), MC4-Evo2 (machined), Amphenol H4, and other connector types used in the platform's component database.
• Current per connection — hourly string current from the simulation.

Typical annual connections loss: 0.3-1.0%. High-quality machined connectors (MC4-Evo2, Staubli) reduce contact resistance by 40-60% compared to generic stamped connectors. For our Dubai example with 28-module strings and premium connectors, we estimate 0.4%.

7. Availability Loss

Availability loss captures energy not produced because plant components are offline. The primary contributors are inverter downtime (scheduled maintenance, forced outages, firmware updates) and grid curtailment. Energy Optima models availability as a time-based factor applied to the hourly energy output.

Key inputs:

• Inverter MTBF and MTTR — each of the 200 string inverters in Energy Optima's database has manufacturer-specified reliability parameters. For modern string inverters, typical availability is 99.0-99.5%.
• Scheduled maintenance — typically 1-2 days per year per inverter block, totaling 0.3-0.6% downtime.
• Grid curtailment — site-specific, varies by market. In markets with high solar penetration (CAISO, ERCOT), curtailment can add 2-5% loss.

For a 50 MW plant with distributed string inverters (each inverter failure affects only a small portion of the plant), total availability loss is typically 1.0-2.0%. In Dubai with a stable grid and high-quality inverters, we estimate 1.2%.

8. Incidence Angle Modifier (IAM) Loss

IAM (reflection) loss accounts for the fraction of irradiance reflected off the module glass instead of being transmitted to the cells. At normal incidence, reflection is minimal (2-4% for standard AR-coated glass). As the incidence angle increases — early morning, late afternoon, and winter months — reflection increases nonlinearly.

IAM by Module Type

Energy Optima uses the ASHRAE IAM model (a single parameter b0) by default, with the option to use a more accurate physical optics model for premium modules:

• Standard AR-coated glass (b0 = 0.05) — common for most commercial modules. IAM loss: 2.8-3.5% for fixed tilt, 2.3-3.0% for trackers.
• Premium double-layer AR coating (b0 = 0.03) — high-efficiency modules with enhanced anti-reflective coatings. IAM loss: 2.0-2.8% for fixed tilt, 1.8-2.5% for trackers.
• Uncoated glass (b0 = 0.07-0.09) — older or budget modules. IAM loss: 3.5-5.0%.

For single-axis trackers, the annual IAM loss is lower than for fixed-tilt systems because the tracker keeps the module more normal to the sun throughout the day. The reduction is typically 0.5-0.8 percentage points. For our Dubai 50 MW example with single-axis trackers and premium double-layer AR modules, we estimate 2.2% IAM loss.

9. Spectral Mismatch Loss

Spectral mismatch occurs when the actual solar spectrum at the site differs from the standard AM1.5G spectrum used for module STC rating. Energy Optima computes spectral loss from the module's quantum efficiency (QE) curve and site-specific atmospheric conditions: air mass, aerosol optical depth (AOD), precipitable water vapor, and cloud optical thickness.

The calculation uses the Bird clear-sky spectral model to generate the incident spectrum for each hour, then convolves it with the module's QE curve. The ratio of the actual short-circuit current to the AM1.5G short-circuit current gives the spectral correction factor.

Typical values by module type:

• Mono-crystalline PERC: 0.8-1.5% — good blue response, broad absorption
• Heterojunction (HJT): 0.5-1.2% — excellent spectral response across the full range
• Thin-film CdTe: 1.5-4.0% — narrow absorption band, higher sensitivity to spectral shifts
• Bifacial (rear side): adds 0.3-0.8% additional spectral loss on the rear side because the ground-reflected spectrum is shifted by the albedo

For our Dubai example with bifacial PERC modules in a high-AOD desert environment, we estimate 1.0% spectral loss.

10. Temperature Loss

Temperature loss is typically the largest single loss category and the one with the greatest hourly variation. Crystalline silicon modules lose approximately 0.34-0.45% of rated power per degree Celsius above 25°C. In hot climates, cell temperatures routinely reach 65-75°C during peak irradiance hours, producing instantaneous losses of 14-20%.

NOCT-Based Temperature Calculation

Energy Optima calculates cell temperature using the NOCT (Nominal Operating Cell Temperature) method standardized in IEC 61215. The cell temperature T_cell at each hour is:

T_cell = T_ambient + (NOCT - 20) × (G / 800) × (1 - η_mppt / 0.9)

Where:

• T_ambient — hourly ambient dry-bulb temperature from the TMY file
• NOCT — the module's nominal operating cell temperature (typically 42-48°C for glass-backsheet, 40-45°C for glass-glass modules)
• G — hourly POA irradiance in W/m²
• η_mppt — module efficiency at the MPPT operating point (typically 19-22% for modern modules)

Wind speed modifies the NOCT-adjusted temperature. Energy Optima applies a wind speed correction factor based on the module mounting configuration:

• Roof-mounted (close to roof surface): reduced convective cooling, add 2-5°C
• Ground-mounted, open rack (well-ventilated rear): standard NOCT applies
• Ground-mounted, tracker: improved airflow, subtract 1-3°C from NOCT estimate
• Bifacial, open rack: 1-3°C cooler than monofacial due to rear-side radiative cooling to the ground

Typical annual temperature loss:

• Dubai (hot desert): 7-10% annual average, summer peaks over 18%
• Phoenix, AZ: 6-8%
• Madrid, Spain: 4-6%
• Los Angeles, CA: 3-5%
• Berlin, Germany: 2-3%

For our Dubai example with bifacial modules (NOCT 42°C) on single-axis trackers, we calculate approximately 7.5% annual temperature loss.

Worked Example: 50 MW PV Plant in Dubai

Let's assemble a complete loss waterfall for a 50 MW_DC PV plant in Dubai, UAE. Project parameters: bifacial mono-PERC modules (600 W, 42°C NOCT, ±2% binning), single-axis trackers (GCR 0.38, 60° max rotation), distributed string inverters, monthly robotic cleaning, flat terrain with no horizon obstructions.

Step	Component	Loss	Cumulative (kWh/m²/yr)	Cumulative (% of POA)
—	POA irradiance	—	2,320	100.0%
1	− IAM	2.2%	2,269	97.8%
2	− Spectral	1.0%	2,246	96.8%
3	− Soiling	1.0%	2,224	95.8%
4	− Shading	0.8%	2,206	95.1%
5	− Temperature	7.5%	2,041	87.9%
6	− Mismatch	0.6%	2,028	87.4%
7	− DC wiring	1.0%	2,008	86.5%
8	− Connections	0.4%	2,000	86.2%
9	− Availability	1.2%	1,976	85.2%

Net DC yield: 1,976 kWh/kW_DC/yr — a total of 14.8% loss from POA irradiance to the inverter input. After inverter losses (typically 1.5-2.5% for modern string inverters) and clipping (0.5-1.5% at a reasonable ILR of 1.20-1.30), the net AC yield is approximately 1,920-1,940 kWh/kW_DC/yr.

At $0.04/kWh PPA, this 50 MW plant generates approximately 96,000-97,000 MWh/yr worth $3.84-3.88M in annual revenue. Each percentage point of loss reduction adds about $38,000-40,000/year — or nearly $1M over a 25-year project life at a 6% discount rate.

The Cumulative Waterfall Visualization

Energy Optima's loss waterfall visualization presents the ten loss categories as a cascading bar chart, with each bar showing the cumulative remaining energy after that loss is applied. The chart makes the relative impact of each category immediately visible — a user can see at a glance that temperature dominates in Dubai while soiling dominates in a dusty but mild-climate site.

The waterfall in Energy Optima's PV designer includes:

• Stacked bars — each loss category shown as a reduction from the previous bar, with the remaining energy shown at each step
• Color coding — losses grouped by category: optical (IAM, spectral, soiling), thermal (temperature), electrical (mismatch, wiring, connections), and external (shading, availability, age)
• Interactive drill-down — clicking any bar opens a detailed hourly or monthly view of that loss category
• Scenario comparison — multiple waterfalls can be overlaid to compare different module types, tilt angles, cleaning schedules, or inverter configurations
• Benchmark overlay — optional industry benchmark lines (NREL PVWatts default, typical commercial PR) shown on the same chart for reference

The waterfall chart is generated automatically after each simulation run, and the underlying data — hourly loss factors for all ten categories — is available for export as CSV for custom analysis or reporting to lenders and investors.

Sensitivity Analysis

The true power of modeling all ten loss categories individually is sensitivity analysis: understanding which losses have the largest impact on project value and where design optimization or operational expenditure is most effective.

For our Dubai 50 MW example, a sensitivity sweep across key parameters reveals:

• Temperature (7.5%): The single largest lever. Switching from standard modules (NOCT 45°C) to bifacial modules (NOCT 42°C) reduces temperature loss by 0.5-1.0 percentage points, worth ~$38K-75K/yr.
• Soiling (1.0%): Increasing cleaning from monthly to biweekly reduces soiling from 1.0% to ~0.6%, but the incremental cleaning cost ($12-18K/yr for a 50 MW plant) must be weighed against the ~$15K/yr revenue gain.
• Mismatch (0.6%): Upgrading from ±3% binning to ±1% binning (available in the component database for premium modules) cuts mismatch to 0.3%, adding ~$12K/yr.
• IAM (2.2%): Upgrading from standard AR coating to premium double-layer AR coating reduces IAM loss by 0.3-0.5 percentage points, worth ~$12-20K/yr.
• DC wiring (1.0%): Oversizing cables by one gauge (e.g., 6 mm² to 10 mm²) for the longest home runs cuts total DC wiring loss by 0.15-0.3 percentage points.

Energy Optima's scenario comparison feature automates this sensitivity analysis. Users can create a base case and then vary one or more parameters — module type, inverter model, cleaning schedule, GCR, tilt angle — and instantly see the impact on the loss waterfall, net yield, LCOE, and IRR.

Comparison to Industry Benchmarks

How does Energy Optima's loss waterfall compare to standard industry references?

• NREL PVWatts v8 default — uses a simpler loss model with total system losses defaulting to 14.07% (the so-called "NREL default" split across soiling 2%, shading 3%, snow 0%, mismatch 2%, wiring 2%, connections 0.5%, LID 1.5%, nameplate 1%, availability 3%, age 0%). This is a fixed-factor approach that does not capture hourly weather-dependent loss interactions.
• Typical commercial bankable PR (performance ratio) — lenders typically expect 78-83% PR for well-designed utility projects in moderate climates and 75-80% PR for hot desert climates. Our Dubai example yields a PR of approximately 80-82% (net AC vs. STC-rated DC capacity), consistent with expectations for a well-designed project.
• Energy Optima difference — because Energy Optima computes each loss category at hourly resolution using site-specific inputs, the results are typically 1-3% more accurate than PVWatts for hot or dusty sites where the interaction effects are strongest. The platform also allows users to replace default assumptions with project-specific measurements (actual soiling rates from a test array, manufacturer-specific IAM measurements, inverter efficiency curves from factory test reports), increasing confidence in the energy yield estimate for project finance.

For a deeper dive into how loss categories interact with string sizing and inverter matching, see our guides on PV system sizing guide and string sizing and inverter matching. For a general introduction to loss waterfall methodology, see PV loss waterfall analysis.