PV Engineer
Hourly PV + storage simulation
byEco Green Energy
Starting the simulation engine

1Site

Standard time, used to place working hours and tariffs. Set automatically from the place.

2Weather data

    When the two disagree more than their measured accuracy, the P90 widens. The simulation itself runs on the source chosen above.

      3Module

      Datasheet values

      4Inverter

      Maker and model; a number is read as the AC power in kW. Available once the engine is ready.

      Datasheet values

      5Array

      Direction the modules face, clockwise from north: 180 faces south, 0 faces north. Trackers: axis orientation.

      6Plant condition

      Which plant is this yield for? On the validation bench the physics held within a few percent at every step. The rest of the gap between model and meter was how the plant is kept.

      The gap is what real plants lost against the as-designed model. A plant whose inverter clips loses less than the gap.

      Losses

      What you type wins over the plant condition. Near shading is a flat declared loss: the engine does not model the geometry of obstacles. DC and AC ohmic losses are entered at full power, as on a single-line diagram; the engine applies them hour by hour as I²R, so their annual effect is lower.

      7Consumption

      Monthly consumption, from the billsoptional

      All twelve months, in kWh. They replace the annual figure: each month is scaled to its own bill, and the peak follows from them. A base load in kW is held in every month.

      8Storage

      What the battery must do. Each service is sized on its own, then they add up: a backup reserve cannot also be cycled.

      9Project

      The engine is loading

      An engineer's yield study, in the browser.

      Hour by hour over a typical year, then over the whole life of the plant: every loss from sky to meter, the P50 and the P90, and what a battery really changes once it has aged.

      1. Physics from pvlib, the open library maintained by Sandia and NREL, the same models engineering offices use.
      2. The module is your datasheet. A single-diode model is fitted to it and checked against its own curves.
      3. Weather from several databases, compared, so the uncertainty is measured rather than assumed.
      4. Brand-blind results. Products are suggested only afterwards, from the results, and never change them.

      Runs on this device. Nothing you enter is sent anywhere except the weather request (coordinates only).