A flat solar peak: clipping or something else?

Confirm which curve is flat

Start with the label and source. Is the graph DC solar input, AC inverter output, battery charging power or a software total? A limit on one quantity is not automatically a limit on another. Record the unit, reporting device and whether the samples are fresh.

Also inspect the chart scale. A display capped at a chosen axis maximum can make data look clipped even when the underlying readings continue changing. A rounded number may hide small variation. Compare the numeric values and the equipment’s own display at approximately the same time.

Understand the actual clipping hypothesis

The PVWatts system-design documentation explains a model in which inverter output is limited by its rated AC size when available array power is higher. That is the basic clipping idea. It does not establish that every flat segment in an off-grid or battery system has this cause.

Check the installed inverter’s documented ratings and operating conditions. Do not equate the array’s DC nameplate rating with an AC output ceiling. The array and inverter are different parts of the conversion path.

Keep competing explanations on the page

A battery-oriented system may not be trying to harvest all available solar when the destination cannot accept it. Export restrictions, operating modes and thermal limits may also be relevant for particular equipment. These are questions for the model’s documentation, not diagnoses to assign from a generic list.

Compare several ordinary occurrences

Note whether the ceiling repeats at the same value on clear days, whether it moves with the load or battery state, and whether it appears only after a status change. A reproducible pattern gives the installer something concrete to investigate.

For example, a plateau that occurs only while the displayed battery is full raises a different question from a fixed AC-output ceiling reached at several battery states. Neither observation alone proves the answer, but it helps choose which evidence to collect next. Do not change limits to force a test.

Do not turn peak shape into an annual loss claim

You cannot calculate annual lost energy by multiplying one flat afternoon by every day of the year. Weather, season, temperature, array orientation and operating demand change. Even a model-based estimate depends on correct inputs and assumptions.

If the question is whether the installation is behaving as designed, provide the model/firmware, observation sheet and comparable graph windows to the manufacturer or installer. If the question is future system design, use a qualified design process rather than treating a monitoring screenshot as a sizing tool.

The fair day-comparison guide helps you prepare those graph windows without mixing partial days or missing data into the evidence.