PV & Hybrid System Design: Complete Guide

PV modules and strings

Connecting identical modules in series increases voltage while current stays approximately that of one module. Parallel strings increase available current. An MPPT tracker has its own voltage/current limits; multiple connectors on one tracker are not independent MPPTs. When parallel strings share one input, a fault in one string can be fed back by the others. The designer screens the reverse current as (p − 1) × I_sc, the contribution of the other strings at the temperature-corrected short-circuit current under the stated assumptions, and compares it with the module's maximum series-fuse rating. It is not a complete worst-case bound: irradiance enhancement, bifacial gain and possible inverter backfeed can add to it. Above the rating, string overcurrent protection is required; if the rating was not entered, the need is reported as NOT ASSESSED. The overall design status is shown as a badge: Provisional means every checkable limit was met but the result still depends on your datasheet values; Partial design means not all the entered modules could be connected; Not fully assessed means a check (for example the full-load MPPT range) could not be made because data was not entered. None of them is an approval. Permanent limits of the monthly model sit in the collapsed “Scope and assumptions” panels (how to read the results).

Fields are checked as you type. A value that is impossible on its own (a zero Voc, a tilt above 90°, a loss of 100 %) turns its field red and shows a ⚠ icon and a message beside it. Relations between fields, such as Vmp below Voc, the MPPT voltage window or the battery voltage window, are reported in the result instead.

Entering the MPPTs. Most inverters have identical MPPTs, so you enter their data once together with their number (1 to 12). If your inverter has MPPTs with different limits, tick “The MPPTs are not identical” and enter each one; unticking copies the first MPPT to all the others.

Cold Voc and hot Vmp

Voc increases at low cell temperatures for common silicon modules (a negative coefficient): check the highest Voc, normally at the cold end, against all absolute voltage ratings. Vmp usually decreases with cell temperature: check the lowest Vmp, normally at the hot end, against the MPPT minimum and the highest Vmp, normally at the cold end, against the operating maximum. The designer evaluates BOTH design temperatures for every limit, so an unusual or mistyped positive Voc coefficient does not hide the true highest Voc (it moves it to the hot end); such signs are accepted with a warning, because they are far more often a typing error than a real module. The inverter startup voltage is screened the same way: the lowest string Voc over the design range (normally at the hot end) must stay above the MPPT startup voltage, not only the STC Voc. A hot-end value below it is shown as a warning, "startup at the maximum design cell temperature is NOT PROVEN", and the layout is kept, because many inverters keep operating below their startup voltage and the operating-voltage range is checked separately. Tick the conservative startup screen to exclude such layouts instead. An STC Voc below the startup voltage always excludes the layout. A voltage pass still does not guarantee startup, which also depends on the irradiance, the inverter's behavior and the manufacturer's startup specification. The optional full-load MPPT voltage range needs a separate result.

Orientation and trackers

East, south and west arrays generate different daily profiles. Mixing different orientation strings on one tracker can degrade yield. The initial optimizer keeps separate orientation/module groups on independent MPPTs and respects physical PV input count and Isc ratings. The monthly energy estimate converts horizontal irradiation to the tilted plane. The share of diffuse light matters: it is higher in winter, when a tilted array gains less over a horizontal one. By default KiloVolt takes each month's diffuse share from the monthly table of the site data (for the built-in default, its illustrative values); a constant fraction, or the Page correlation, which derives it from how clear the month is, can be selected instead. For the default project, switching to Page lowers the January yield by about 1 to 2 % and changes the annual total by less than 1 %. Both options are statistical approximations, not measurements.

DC/AC ratio and clipping

A 25.38 kWp PV array with a 20 kW inverter has DC/AC = 1.269. This alone is not a safety failure, nor proof of suitability. The preliminary allocator ranks by pre-clipping energy potential, not proven AC-energy optimum. Follow the manufacturer DC limits, AC ratings and operating conditions. Monthly irradiation totals cannot quantify actual inverter clipping because instantaneous DC power peaks are unknown.

Battery coupling and selection

A DC-coupled battery shares a DC/hybrid conversion path; an AC-coupled battery uses its own battery inverter. Matching voltage and power figures is insufficient: the manufacturer must support the battery model, control communications and permitted module/tower arrangement. In series-connected modules the voltage window scales with modules per tower, whereas integrated power-module towers may have a manufacturer-defined fixed external voltage range. Parallel towers generally need consistent, manufacturer-approved layouts; the tool defaults to equal module counts per tower. Usable energy depends on SOC limits and efficiencies.

Batteries are optional. The battery fields and results appear only after you tick “Include batteries”. Without it the design has no battery category at all: only the strings, the energy and its value estimate.

Monthly energy estimate limits

Without a weather time series, use documented monthly horizontal irradiation and an explicit approximate tilt/azimuth transposition. This can estimate annual and monthly PV yield, not the time-dependent self-consumption, charge cycles, grid curtailment or precise battery savings. Any battery-savings scenario must state its assumed direct-use fraction, cycles/day, available charge/discharge hours per day, and battery power limits. A no-benefit case must not automatically recommend storage.

Solar data is optional. The electrical limit checks of every string use only the design cell temperatures, never solar data. For the energy estimate, KiloVolt uses a built-in generic default (an illustrative climate computed by KiloVolt from simple assumptions, not measured data and not a specific place) and says so next to the result; its figures are estimates only; enter your own monthly irradiation and temperature for another site or a better estimate. In automatic mode the solar data only helps rank layouts that are already electrically feasible.

How the energy is estimated. Monthly horizontal irradiation is converted to the tilted plane with an anisotropic sky model (Hay-Davies with the Reindl horizon term). The module temperature is estimated from the daytime ambient temperature and the irradiance; the temperature coefficient of the module you entered then sets the power lost at that temperature. The result lists the losses: temperature and low irradiance, optical (angle of incidence, net of spectral), shading, soiling, DC wiring, other system losses (mismatch, AC wiring, availability) and the inverter; the defaults total about 14 % system loss plus about 2 % optical. It also shows the in-plane irradiation and the performance ratio. The model was compared with an independent solar-resource simulation tool for sites in Greece (generic crystalline silicon, −0.40 %/°C): at the site used to calibrate it, annual energy agrees within about 1 % and single months within about 6 %; at a second site, not used to build the model, the annual differences were +0.1 % (30° south) and +2.5 % (20° east) with single months within about 4 %, the east-facing case being the weaker one. A modern module with −0.26 %/°C yields about 3 % more, which is physical. Two sites in Greece are not every climate: colder, cloudier or non-Greek sites have not been checked, and the reference tool is itself a model, not a measurement.

Zero-export and economics

Zero export means surplus PV may be curtailed if loads and batteries cannot accept it. Export restriction is not inverter clipping. A legitimate financial comparison considers avoided imports and lost export compensation. Real dispatch/NPV optimization needs sufficiently granular consumption and production profiles as well as actual component costs.

The value estimate. Enter an electricity price in €/kWh and the result shows the annual value of the estimated production: the annual energy multiplied by that price, and the same per kWp. It is a rough estimate. It does not model the tariff structure, how much is self-consumed or exported, degradation, price changes, costs or payback. The battery savings scenario is separate and appears only with “Include batteries”; it rests on your own assumptions and is not an hourly simulation.

Safety and commissioning

Before applying the result, verify string connector compatibility, cable protection, DC isolators, SPD requirements, earthing and bonding, applicable national interconnection rules, inverter grid-code compliance and documented battery/inverter certification.

Key engineering equations

Temperature-corrected open-circuit voltage

Apply only within the documented assumptions above.

Array rating

Apply only within the documented assumptions above.

Monthly preliminary production estimate

Here P_STC is in kWp, monthly plane-of-array irradiation H_POA is in kWh/m², and G_STC = 1 kW/m² is the reference irradiance. Their ratio is equivalent sun-hours (h), giving kWh after multiplication by rated kW and the dimensionless monthly performance ratio. This is a statistical estimate only.